13th Annual Conference of the International Society for Quality of Life Research

...

3 downloads 234 Views 5MB Size

Qual Life Res (2006) DOI 10.1007/s11136-007-9211-7

© Springer Science+Business Media B.V. 2007

International Society for Quality of Life Research

Abstracts presented at the

13th Annual Conference of the International Society for Quality of Life Research “HRQOL Research: Making an Impact in the Real World”

October 10 - 14, 2006 Lisbon Portugal

an electronic supplemement to Quality of Life Research journal

How to Cite Abstracts:

Please use the following format to cite abstracts included in this document:

Mitchell, J., Bradley, C., (2006) Quality of Life in Macular Degeneration: Age-Related Macular Degeneration Alliance International (AMDAI) White Paper. 2006 International Society for Quality of Life Research meeting abstracts [www.isoqol.org/2006mtgabstracts]. The QLR Journal, A-68, Abstract #1728

International Society for Quality of Life Research, 6728 Old McLean Village Drive, McLean, VA 221010 USA; telephone: 703-556-9222; fax: 703-556-8729; email: [email protected]; web: www.isoqol.org Quality of Life Research is published monthly, except January and July. Periodicals postage paid at Rahway, NJ USPS No. 0.13-175. U.S. Mailing Agent: Mercury Air Freight International Ltd., 365 Blair Road, Avenel, NJ 07001. Published by Springer. P.O. Box 17, 3300 AA Dordrecht, The Netherlands, and 101 Philip Drive, Norwell, MA 02061, U.S.A. Postmaster: Please send all address corrections to Quality of Life Research, c/o MKercury Airfreight International Ltd., 365 Blair Road, Avenel, NJ 07001, U.S.A

International Society for Quality of Life Research Lisbon, Portugal, October 10 –14, 2006 ~ Schedule-at-a-Glance The abstracts are grouped by symposia, followed by all of the oral presentations, and conclude with all of the poster presentations. Thursday, October 12, 2006 8:00 – 9:00 am Special Interest Group meetings 8:00 am – 12:30 pm Poster Session 1 on display (p. A-70 – A-83) 9:00 –10:30 am Plenary 1: Welcome and Keynote Address 10:30 –11:15 am Break and meet the authors poster session 11:15 am – 12:45 pm Concurrent Sessions Symposium: The Value of Patient Reported Outcomes in Clinical Practice (p. A-2 – A-3) Oral Sessions: Elderly 1, Caregivers 1, IRT 1, Endocrinology (p. A-17 – A-23) 12:50 – 2:00 pm Lunch panel session: “What the Editors Say!” 2:00 – 3:30 pm Concurrent sessions Symposium: Can we use patient reported outcomes to compare the quality of care achieved by different healthcare providers? (p. A-3 – A-4) Oral Sessions: Children 1, Minimally Important Differences, Oncology 1, Methodology 1 (p. A-23 – A-29) 2:00 – 6:15 pm Poster Session 2 on display (p. A-83 – A-97) 3:30 – 4:00 pm Break 4:00 – 5:30 pm Concurrent sessions Symposium: Quality of life in public health policy (p. A-4 – A-5) Oral Sessions: Mental Health, Respiratory, GI, Psychometrics (p. A-29 – A-36) 5:30 – 6:15 pm Meet the authors poster session 6:00 – 7:00 pm Mentor/Mentee session Friday, October 13, 2006 8:00 – 9:00 am Special Interest Group meetings 8:00 am – 12:30 pm Poster Session 3 on display (p. A-97 – A-109) 9:00 – 10:30 am Plenary 2 “QOL and the policy makers” 10:30 –11:15 am Break and meet the authors poster session 11:15 am – 12 :45 pm Concurrent sessions Symposium: Bridging the Gaps in Understanding Symptom Burden and QoL Impairment in Cancer Patients (p. A5 – A-7) Oral Sessions: AIDS, Preferences 1, Children 2, Pain (p. A-36 – A-42) 12:45 – 2:00 pm ISOQOL Business Meeting Lunch session 2:00 – 3:30 pm Concurrent sessions Symposium: Integrating values and preferences in clinical practice guidelines: how can we make it happen? (p. A7) Oral Sessions: Oncology 2, Cardiovascular, Caregivers 2, Methodology 2 (p. A-42 – A-48) 2:00 – 6:15 pm Poster Session 4 on display (p. A-109 – A-123) 3:30 – 4:00 pm Break 4:00 – 5:30 pm Concurrent sessions Symposium: Enhancing the Quality of Cancer Care Through Research, Practice, and Policy (p. A-8 – A-9) Oral Sessions: Rheumatology, Qualitative, Urology 1, IRT 2 (p. A-48 – A-54) 5:30 – 6:15 pm Meet the authors poster session 6:00 – 7:00 pm Special Interest Group meetings Saturday, October 14, 2006 8:00 am – 12:30 pm Poster Session 5 on display (p. A-123 – A-136) 9:00 – 10:30 am Plenary 3 10:30 – 11:15 am Break and meet the authors poster session 11:15 am – 12 :45 pm Concurrent Sessions Symposia: Patient-Reported Outcomes: Update on the Mayo/FDA Joint Meeting Publications; The Value of Health Related Quality of Life Parameters in Predicting Clinical Outcomes in Cardiology and Oncology (p. A-9 – A-11) Oral Sessions: Neurology, Preferences 2, Surgery (p. A-54 – A-59) 12:45 – 2:00 pm Lunch on your own 2:00 – 3:30 pm Concurrent sessions Symposium: The Patient Reported Outcomes Management Information System (PROMIS): A Cooperative Effort to Advance and Enable Clinical Research (p. A-11 – A-13) Oral Sessions: Urology/Kidney Diseases, Psychometrics 2, IRT 3, Cancer 3 (p. A-59 – A-65) 2:00 – 6:15 pm Poster Session 6 on display (p. A-136 – A-147) 3:30 – 4:00 pm Break 4:00 – 5:30 pm Concurrent sessions Symposia: Translation and Cultural Adaptation of PRO Questionnaires: Review of the activities of the ISOQOL TCA SIG; The NEURO-QoL Project: Using Multiple Methods to Develop A HRQOL Measurement Platform to be Used in Clinical Research Across Neurological Conditions (p. A-13 – A-16) Oral Sessions: Cancer 4, Methodology 3, Elderly 2 (p. A-65 – A-69) 5:30 – 6:15 pm Meet the authors poster session 7:30 pm Gala Dinner

A-1

respectively. The intervention took place during 3 consecutive visits. Patients' completed the EORTC QLQ-C30 and, where available, a condition-specific EORTC module via a touch screen computer. At each visit, both patients and nurses were provided with a graphic summary of current and previous results. Outcomes were assessed via self-report questionnaires and medical record reviews. RESULTS: A total of 361 patients under the care of 10 oncology nurses were approached to participate in the study, of whom 298 agreed. Two hundred nineteen patients completed the study; 79 were lost to follow-up. HRQL-related topics were discussed significantly more frequently in the EC than the UCC (mean = 4.8 vs. 3.8 topics; p < 0.05). Nurses' awareness was significantly higher in the EC , with a mean ICC across HRQL topics of 0.75 vs. 0.52 for the UCC. Awareness of levels of daily activities, pain and overall quality of life was particularly enhanced. Significantly more HRQL-related medical record notations were made in the EC vs. UCC (mean = 24 vs. 20; p < 0.05), and significantly more EC patients received HRQL-related advice (p < 0.05). No statistically significant group differences were observed in patient satisfaction or changes in self-reported HRQL over time. CONCLUSIONS: Incorporating standardized HRQL assessments in daily clinical oncology nursing practice facilitates the discussion of HRQL issues, increases nurse' awareness and stimulates HRQL-related medical record notations and advice.

SYMPOSIA Symposium 1360 The Value of Patient Reported Outcomes in Clinical Practice David Osoba, QOL Consulting, West Vancouver, BC, Canada, Neil K. Aaronson, Psychosocial Research & Epidemiology, The Netherlands Cancer Institute, Amsterdam, North Holland, The Netherlands, Michael D. Brundage, Oncology, Queen's University, Kingston, Ontario, Canada, Galina I. Velikova, Cancer Research UK Centre Leeds, St James's University Hospital, Leeds, UK Patient reported outcomes (PRO) include all aspects of a patient's health status that are reported directly by the patient. They inclue such diverse aspects as reports of needs, symptoms and health-related quality of life (HRQL) As recently as ten years ago, most researchers working in HRQL would have thought that the barriers to measuring HRQL in clinical practice were too great to be overcome. However, today it is a practical possibility. One of the main barriers, that of providing HRQL or other PRO information to health-care personnel in real time, has been solved with the use of touch-sensitive computer screens. Thus, a patient can complete a questionnaire in the clinic and the physician or nurse can have the analyzed information available for viewing almost immediately. This symposium presents the work of three independent groups who have been studying various aspects of HRQL assessment in hospital-based clinical practice settings. Each group has used a different approach to investigate the questions of whether measuring HRQL in the clinic makes a difference to patientphysician communication, clinical decision-making, and to the patients’ welfare as expressed by their HRQL. The results are strikingly successful. They show that when physicians or nurses are given patients HRQL information, communication is more relevant to the patient’s concerns and that both physicians and nurses try to provide assistance in the domains of concern. They also show, in one study, that patients whose physicians were given HRQL data experienced an improvement in some HRQL domains. Future progress will certainly involve the use of PRO information to assist patients in determining their preferences for therapeutic clinical decision-making. The use of IRT approaches with computer adapted testing (CAT) are likely to make the process more adaptable to individual patient’s needs. However, the methods for assessing clinically relevant change over time when using CAT still need to be developed. Nevertheless, assessment of PRO in clinical practice looks like it’s here to stay.

Individual Abstract Number:1810 Patients' Judgments about the Value Quality of Life Information when Considering Lung Cancer (NSCLC) Treatment Options Michael D. Brundage, Oncology, Queen's University, Kingston, Ontario, Canada, Deb Feldman-Stewart, Oncology, Queen's University, Anne Leis, Epidemiology, University of Saskachewan, Andrea Bezjak, Radiation Oncology, Princess Margaret Hospital, Univeristy of Toronto, Joseph L. Pater, Oncology, Queen's University, Kingston, Ontario, Canada AIMS: Clinical trials have shown that chemotherapy (CTX) after thoracotomy for NSCLC improves survival rates but temporarily impairs HRQL scores. We sought to systematically explore how NSCLC patients value HRQL information (when added to treatmentrelated toxicity and survival information) in a treatment decision context. METHODS: Patients from two Cancer Centers who had surgery but no CTX for NSCLC participated in a structured,1:1 interview, acting as surrogate decision-makers. The interviewer first provided the decision context (choice: CTX or not). Information regarding toxicity and survival outcomes was then provided. Participants rated their preference for CTX (0-10; 10 most preferred) and specified the minimum survival gain (SAT) each required for accepting CTX. HRQL (EORTC QLQ-C30) information was then shown (randomized to either small-magnitude or large-magnitude HRQL effects of CTX). We assessed the impact of HRQL information on each participant’s preference ratings and SAT. Further, the participant rated the usefulness of each type of information (global HRQL, Functional Domains, and treatment sideeffects, 0-10; 10 most useful). RESULTS: 71 patients participated (56% female; mean age 68y; 20% had post-secondary education). Mean baseline SAT for accepting CTX was 10.5% (absolute increase in cure rate). On average, participants shown small-magnitude HRQL-effect decreased their SATs for CTX (mean = 9.5%) while those seeing large HRQL effects increased their SATs (mean = 15.9%) [p=0.04]; similarly, patients’ mean preference scores for CTX increased or decreased when shown small- or large-HRQL effect, respectively. 48 patients (67%) rated the usefulness HRQL information as high as, or higher than, toxicity information for making the treatment decision. 16 (23%) patients rated the functional HRQL information as more useful than the global scores, whereas 8 (11%) rated the global information as more useful. CONCLUSIONS:

Individual Abstract Number:1639 The Use of Health-Related Quality of Life (HRQL) Assessments in Daily Clinical Oncology Nursing Practice: A Community Hospital-Based Intervention Study Neil K. Aaronson, Psychosocial Research & Epidemiology, The Netherlands Cancer Institute, Amsterdam, North Holland, The Netherlands, Doranne L. Hilarius, Paul Kloeg, Pharmacy, Medical Center Alkmaar, Symone B. Detmar, Quality of Life, TNO Prevention and Health, Chad M. Gundy, Psychosocial Research & Epidemiology, The Netherlands Cancer Institute, Amsterdam, North Holland, The Netherlands AIMS: This study evaluated the efficacy of incorporating standardized HRQL assessments as a routine part of outpatient chemotherapy treatment in terms of 5 outcomes: 1) nurse-patient communication, 2) nurses' awareness of patients' HRQL, 3) patient management, 4) patient satisfaction and 5) patients' HRQL. METHODS: The study used a sequential cohort design with repeated measures. Two consecutive cohorts of patients were included in an usual care (UCC) condition or an experimental condition (EC),

A-2

HRQL information affected participants' decisions and was rated as highly useful for a variety of purposes.

remiss as the primary objective of most elective care is improvement in HRQL, and we will learn relatively little about the performance of a healthcare provider by focusing on rare outcomes experienced by the minority of patients. For example, if we wish to compare the performance of different hospitals on the outcomes of hip replacement surgery, it would seem vital that we know how they compare on measures of pain reduction and functional improvement, as well as their respective post-operative complication rates. This symposium will examine the possibility of using patient-reported outcome measures (PROMS) to compare the quality of care achieved by different healthcare providers. James Lewsey will first outline the main statistical issues that must be considered when using PROMS in quality of care studies. Mary Anne Hope will then describe the Medicare Health Outcomes Survey, which uses measures such as the SF-36 to compare the performance of managed care plan providers in America. Next, Arlene Bierman will describe how we can use PROMS to improve quality of care. Finally, John Browne will chair a discussion of how we can move this field forward without making the mistakes experienced in other fields of clinical governance.

Individual Abstract Number:1812 Informing Physicians of Cancer Patients Quality of Life (QOL) Physicians' Attitude, Effects on Communication and DecisionMaking Galina I. Velikova, Cancer Research UK Centre Leeds, Leeds University, Leeds, UK, P Selby, J Brown, CTRU, Leeds University, Leeds, UK Randomized trials show that routine use of QOL data in oncology practice facilitates patient-physician communication and may improve patient well-being. AIM: To explore the effects of feeding back QOL information to oncologists on communication and medical decisions. To study the influence of physicians'attitude and severity of QOL impairment. METHODS: Exploratory content analysis of 3 consecutive encounters from a randomized trial-286 patients and 28 physicians comparing 3 arms: Intervention:completing QOL questionnaires with feedback to clinicians,Attentioncontrol:completing QOL questionnaires without feedback and Control:no QOL data. Communication in the 3 groups was compared using mixed effects models. In the intervention group, a detailed analysis was performed of associations between communication and QOL impairment, physicians' ratings of clinical usefulness and actual use of QOL data. RESULTS: 215 patients and 860 consultations were analyzed. Providing QOL information to oncologists led to more consistent discussion of insomnia(p=0.004), dyspnea (p=0.003) and physical functioning (p=0.027) than the control groups. In the intervention group, QOL scores influenced discussion of symptoms, but not functions. Oncologists explicitly used QOL scores in general terms (52%), followed by emotions (28%), fatigue (29%), insomnia (27%), dyspnea (25%) and pain (23%). Physicians perceiving QOL data as useful, more often referred to the QOL information. Discussion of emotional issues was associated with positive physicians' ratings of QOL clinical usefulness. Actions in response to emotional issues were recorded in 25% of intervention consultations vs 19% of controls, due to enquiry about reasons for distress. No differences in the number of supportive care medications were observed. CONCLUSIONS: Informing oncologists of patient QOL leads to more consistent discussion of non-specific symptoms and functions. Individual physicians' perception of clinical usefulness of QOL data is related to actual use of the data and to discussion of emotional issues.

Individual Abstract Number:1242 Statistical Issues to be Considered When Comparing the Performance of Healthcare Providers Using Patient-Reported Outcomes James Lewsey, Public Health and Policy, London School of Hygiene and Tropical Medicine, London, UK A number of statistical issues must be addressed if we are to move forward with the use of patient-reported outcome measures (PROMS) in clinical governance. Many methods for dealing with these issues have been proposed for other clinical areas (e.g. the use of survival rates to identify 'deviant' performance in cardiac surgery) but these issues have yet to be discussed in detail with respect to PROMS. The main issues are as follows. 1. What constitutes a 'clinically important difference' when discussing the performance of a healthcare provider. Should we use standards based upon the distribution (e.g. +/- two standard errors from a national average) or should we use absolute benchmarks. 2. How should we 'risk-adjust' to allow for fair comparison between providers? What role should a pre-treatment score on the measure in question have in risk adjustment? For riskadjustment purposes, a regression approach is often used. Normal regression models assume that the outcome measure is normally distributed. This may be an inappropriate assumption for PROMS as they are often discretely measured, bounded and have skewed distributions. The constant variance assumption may be questionable too. Can we use data transformation, Bayesian modelling or non-parametric approaches to solve this problem? 3. How do we deal with the non-independent nature of data collected at the healthcare provider level? Assuming that a parametric approach is suitable, multilevel models would assume that outcomes from patients within the same provider are not completely independent. A feature of such modelling is shrinkage where the average outcome for providers with few (observed) patients is shrunk towards the overall average outcome. This obviously has implications for making comparisons. 4. How can we perform 'timely' assessments of performance? The objective of continuous monitoring is to identify divergent performance as quickly as possible. Methods and applications to do this can be increasingly found in the medical literature (e.g. CUSUM, control charts). Are these appropriate for PROMS?

Symposium 1241 Can We Use Patient Reported Outcomes to Compare the Quality of Care Achieved by Different Healthcare Providers? John P. Browne, Public Health and Policy, London School of Hygiene and Tropical Medicine, London, London, England, James Lewsey, Public Health and Policy, London School of Hygiene and Tropical Medicine, Mary Anne D. Hope, Survey Analysis, Health Services Advisory Group, Canada, Arlene S. Bierman, Faculties of Medicine and Nursing, University of Toronto, Canada Clinical governance plays an increasingly important role in modern healthcare. It is now widely accepted that the performance of healthcare providers should be assessed to see if it meets certain standards. These standards are often focused on the structure and process of care, but increasingly there is an emphasis on real patient outcomes. Many developed world healthcare systems now require clinicians and hospitals to provide some data on how they perform with respect to outcomes such as surgical complications and death. However, there has been relatively little focus on patient-reported outcomes such as health-related quality of life (HRQL). This is

Individual Abstract Number:1680 The Medicare Health Outcomes Survey Mary Anne D. Hope, Survey Analysis, Health Services Advisory Group, Canada The Medicare Health Outcomes Survey (HOS) assesses the physical and mental health outcomes of the elderly enrolled in Medicare

A-3

managed care in the United States and is the first health outcomes assessment for the Medicare population. The Medicare HOS assesses a managed care plans ability to maintain or improve the physical and mental health of its members over time. The Medicare HOS V1 instrument incorporates the SF-36, demographics, chronic medical conditions, activities of daily living (ADLs), smoking, healthy days, urinary incontinence, depression, and physical symptoms as well as other items. The HOS measure is administered at baseline to a randomly selected sample of beneficiaries from each managed care plan that hold risk and cost contracts as well as Social HMOs. The two-year follow up sample includes beneficiaries that are identified from the baseline sample who are alive, enrolled in their original managed care plan, and had sufficient SF-36 data to derive a physical health summary score and a mental health summary score. The goal of the performance measurement analysis is to compare physical and mental health outcomes in managed care plans, in terms of the percentages of beneficiaries who were better, the same, or worse than expected at the two-year follow up. Multivariate statistical methods are used for case mix adjustment, so all plans will be as equal as possible in terms of demographic and socioeconomic characteristics, chronic medical conditions, initial health status, and other design variables. An overall F-test is used to test whether plans differ significantly on the outcome measures. If the overall F-test is significant, then t-statistics are used to express the significance of each plans difference from the overall national results. Beneficiary level results are aggregated to derive the plan, state, and HOS national percent better, same, and worse than expected values. Outliers are those plans that performed significantly better (i.e., better than expected) or significantly worse (i.e., worse than expected) when compared to the national average.

provides a clinically meaningful metric for their interpretation. The HOS measure may provide a useful tool for assessing and improving functional health outcomes related to the management of common chronic conditions and geriatric syndromes in the elderly.

Symposium 1910 Quality of Life in Public Health Policy Serge Briancon, Serge Briancon, Chu Nancy Clinical Epidemiology and Evaluation, Nancy, France, Carole Cretin, Francis Guillemin, School of Public Health, Nancy, Pascal Auqier, Laboratoire de Sante Publique, France Few countries have put the focus of national health policy on preventing and moderating quality of life impairment in major chronic diseases. Among difficulties to implement such policy planning are the many different diseases concerned, the variety of actors in society, the difficulty in choosing a universal system for monitoring health plan and assessing its benefits. Recent changes in disease approach (ICF), acceptance of indicators (by law 2004) and federation of partners (e.g. patients groups) have made this possible in France. As a part of the 2004 Public Health Law, The French Ministry of health has recently prepared a plan for improving quality of life of chronic diseases patients. This symposium will present the rationale for the development of such a policy, discuss the choices that have been made in French plan improving quality of life in public health policy between improvement of care management patient education, professional and social insertion and finally the place of research in such a plan. It will be open to share experience with other countries health policy planning.

Individual Abstract Number:1681 Using the Medicare Health Outcomes Survey to Improve Quality of Care Arlene S. Bierman, Faculties of Medicine and Nursing, University of Toronto, Canada

Individual Abstract Number:1911 National Health Policy Planning in France: Implementing Programme Towards Improving Quality of Life of Persons with Chronic Disease Serge Briançon, Ecole de santé publique (EA 4003), NancyUniversité, France

Improving HRQOL in elderly individuals with chronic illness is dependent upon understanding modifiable factors associated with poor health outcomes. The Medicare Health Outcomes Survey (HOS), is a longitudinal, self-administered survey, which utilizes the SF-36® (an internationally recognized, scientifically validated measure of physical and mental health status), and additional case mix/risk adjustment variables to measure population-based HRQOL and health outcomes among elderly and disabled enrollees in Medicare managed care health plans in the US. The HOS performance measure is an assessment of a health plan's ability to maintain or improve the physical and emotional health of enrollees over a two-year period of time. The study population included managed care enrollees age 65 and older (n=360,711) who responded to the Cohort 1 (1998-2000) or Cohort II (1999-2001) Baseline surveys. Adapting a previously validated severity metric (TIBI) used in the HOS instrument, we examined the association of symptom severity with SF-36 PCS, MCS, and 8 subscales, and two year mortality for five common chronic conditions: congestive heart failure, ischemic heart disease, chronic obstructive pulmonary disease, diabetes, and arthritis. We also examined the association of the presence of urinary incontinence, visual impairment, and hearing impairment with these outcomes. Effective interventions are available for each of these conditions that can improve HRQOL, and there is considerable evidence that these conditions and their associated symptoms are often suboptimally managed in practice. Multivariate techniques were used to control for various patient characteristics. Symptom severity was strongly associated with worse physical and mental health measured by the SF-36 for all conditions. The association of these scores with symptoms as they present in practice

In the context of recent legislation in France, the law on the rights of patients (march 2002) has put the emphasis on quality of life (QoL) as a major outcome to consider in several chronic conditions out of 100 objectives. Further, the law for public health (august 2004) has set priority to develop 6 national plans for health, in the following topics: cancer, environment, handicap, violence, rare diseases and improving quality of life for patient with chronic disease. This has led to elaborate proposals for various implementations of actions toward improving QoL at the population level. The recent development of the International Classification of Functioning and health (ICF) framework has offered the opportunity to develop a unifying concept of the chronic disease (CD). This allowed framing actions applicable across an unlimited list of chronic conditions, and at various levels of determinants. Many actors have been involved in the elaboration process, including Ministry of health, national social security, and several health governmental agencies with broad missions: promotion and health education, health statistics and health monitoring, … etc. Partnership with patients associations and self-help groups has been strongly valued during the elaboration process, resulting in the promotion of specific actions like patient partner programme toward training of health professionals by patients, and promoting the use of patient-reported outcomes overall. Several structured proposals have been made, so as to improve epidemiological knowledge about CD, with national coordination

A-4

-

to improve measurement and research on QoL in CD, in clinical practice, national monitoring surveys, and implementing resource centres for QoL, to warrant patient education of good quality, with specific resources to develop caregiver and proxy training, in initial and continuing education of health professional and self-help groups to ensure coordination and quality of care, in medical and social areas to facilitate social, school and professional integration, in particular for children, unemployed persons, and against social exclusion A medico-economic evaluation of some actions has been proposed to accompany these proposals. Focusing on QoL outcome as a strategy revealed to be an opportunity to facilitate implementation of many kinds of actors and of interventions in prevention, treatment, rehabilitation using drug treatment, non-drug treatment and complex strategies of care. It is planned to conduct evaluation of actions implemented.

instruments across different conditions, researches investigating determinants of quality of life through longitudinal approaches and finally researches evaluating the impact of interventions aiming at improving HRQOL. Collaborative research at national and international level will be encouraged especially by creating funding for fellowship. Interdisciplinary collaboration will be privileged especially between epidemiologists and human and social scientists, and between epidemiologists and health economists. Individual Abstract Number:1914 National Quality of Life Observation, Monitoring and Evaluation Pascal Auquier, Laboratoire de santé publique, Université AixMarseille, France The chronic diseases due to their number their variety and their societal weight require taking into account the repercussions which they lead in quality of life. The consideration of the quality of life in the field of the health is a recent step and of American north impulse. The works led in the field of the chronic diseases consisted of researches for settling of evaluation of tools. Few applied researches were developed, but the quality of life appears as a criterion of decision more and more frequently reported as a supplement to criteria of judgment more traditional. The application of a research policy in the specific field of the chronic diseases bases on 5 measures: (1) Coordinate the research works on the quality of life and the chronic diseases; (2) strengthen the priority ' quality of life and chronic diseases ' in the institutional grants; (3) Encourage national and European partnerships; (4) encourage the development of interdisciplinary collaborations; (5) value the research activities on the quality of life. Reference centres for the coordination of the quality studies of life in the chronic diseases are a set of multi-field competence organized around thematic specialized: rheumatology, psychiatry, neurology, cardiovascular, cancer research. Each centre assumes different missions: (1) expertise for the apprehension of the quality of life of a disease or a group of diseases, having developed the competence specific and recognized in this domain; (2) because of the rarity of the teams specialists to exercise, the centre will have a national or international attraction. Labellisation of the reference centres will be conduct by an independent expertise. A condition of contract clarifying the missions and the activities expected from the reference centres for the coverage of the diseases must be defined. Among these missions will have to represent to give to the reference centres a mission of information of the medical teams, the patients and the families. The absence of organized links between the reference centres and institutions are harmful. It is thus advisable to coordinate them in national network under the authority of the group schedule.

Individual Abstract Number:1912 Implementing National Policy Towards Chronic Disease: Health Policy and Partnerships Carole Cretin, Alexandra Fourcade, Ministère de la santé, France Having quality of life as a health policy object is not usual, representing a true change in our models. Public health practices are generally constructed around punctual responses to health problems, focusing on short interventions by health professionals, and the considered outcomes are death or relief. Including quality of life in health policy program implies to go beyond the separation cure/prevention, individual responses/organised responses. It also implies to develop extended partnerships, quality of life being not the private ground of professionals. The French State is the pilot, initiating, driving, presenting and coordinating the activities through the common frame offered by the 2004 public health law. Implication of patients is the first and main step in such a process involving quality of life and patient reported outcomes. Thus a national concertation of all patient associations has been conducted, in 2004 using a Delphi method, leading to 15 proposals which has been publicly presented and discussed and then included in the multidisciplinary and multiprofessionnal work groups built around 1. Epidemiology, measure and research on HRQoL; 2. Health Education; 3.Health professional training; 4. Coordination of patient management; 5. Social, professional and scholar insertion; Defining health policy programs aiming at improving quality of life in patients with chronic disease require collaboration and discussion with stakeholders working in fields beyond the narrow definition of health problems.

Symposium 1475

Individual Abstract Number:1913 Health Plan Policy and Population Research Development Programs Francis Guillemin, Ecole de santé publique (EA 4003), NancyUniversité, France

Bridging the Gaps in Understanding Symptom Burden and Quality of Life Impairment in Cancer Patients Henning H. Flechtner, Dept. of Child and Adolescent Psychiatry, Henning H. Flechtner, Dept. of Child and Adolescent Psychiatry, University of Cologne, Cologne, NRW, Germany, Tatyana I. Ionova, Quality of Life, Multinational Center of QoL Research, St. Petersburg, Russia, Andrei A. Novik, Hematology Department, National Pirogov Medical Surgical Center, Moscow, Ru, Moscow, Russia, Samuel M. Salek, WSP Centre for Socioeconomic Research, Cardiff University, Cardiff, UK

A health plan policy on quality of life cannot be imagined without development of research programs. Proposals have been brought in this field by multidisciplinary work groups involving patients, professionals, researchers and decision makers. These proposals remain to be eventually adopted in the plan. Conceptual, metrological, etiological, prognostic and evaluative researches are under consideration. Researches on quality of life should be driven as a priority in institutional funding proposals, including researches on the chronic disease rather than chronic diseases and adaptation and utilisation of ICF, researches on psychometric properties of HRQOL

At present much is done in patient-reported outcomes research in oncology. Although QoL is recognized as an important outcome and symptom control is common in cancer, clinicians rarely provide comprehensive QoL and symptom assessment. Translation of the

A-5

results of patient-outcomes research in clinical practice is worthwhile. Better symptom control can produce significant improvement in patient's QoL. Implementation of symptom assessment in routine cancer practice is a first step toward improving quality of cancer care. Understanding of patient's QoL at different time-points of disease course provides the basis for better treatment outcomes. Thus, the symposium aims to bridge the gaps in understanding symptom burden and quality of life impairment in cancer patients at different stages of the disease and its treatment. In this symposium, we will: -Present the state-of-the art of fatigue measurement and management in cancer patients. The focus will be given to the long-term cancer survivors and importance to control fatigue in patients at long-term follow-up will analysed. -Review the current progress in symptom research and present the results of national epidemiological study on prevalence and treatment of cancer-related symptoms. Approaches to identify the populations who experience poor symptom assessment and management will be presented. -Present the results of patient-reported outcomes studies in palliative care patients. Data on the evidence for the efficacy of a palliative medicine outpatient clinic will be reported. The approaches to improve quality of care for palliative patients will be discussed. Discuss the approaches to minimize the gap between researchers and physicians in the field of quality of life and symptom assessment in cancer patients. Emphasis will be given to application of the system of grading QoL impairment to clinical practice. The ultimate goal of the symposium is to make links between patient-reported outcomes research and clinical practice in oncology, and in doing to improve quality of cancer care.

treatment of this symptom to less than 30% of patients. In addition, two thirds of patients reported at least one moderate-to-severe symptom. The results of this symptom outcome study will provide the initial data on symptom control in Russia and will establish the framework for identifying the populations who experience poor symptom assessment and management. Individual Abstract Number:1497 Symptom Interference with Quality of Life: Application to Clinical Practice Andrei A. Novik, Hematology Department, National Medical Surgical Center, Moscow, Ru, Moscow, Russia For the patient suffering from advanced cancer the symptoms are the focus of concern. Unrelieved symptoms lead to significant distress and impaired quality of life (QoL). Although QoL is recognized as an important outcome in advanced cancer, clinicians rarely use QoL data in their everyday decision-making. Grading of QoL impairment and analysis of symptom burden is worthwhile to provide an adequate management of advanced cancer. We have introduced a novel system of grading QoL impairment. This system provides physicians with quantitative assessment of QoL worsening in cancer patients as compared to a population norm (PN). Cancer patients may experience 5 grades of QoL impairment: none, mild (25% decrease from a PN), moderate(25-50% decrease), severe(50-75% decrease) or critical(>75% decrease). The practical application of this system has been tested on a sample of 500 advanced cancer patients. The majority of patients exhibited a critical or severe QoL impairment (at least a 50% QoL decrease as compared to PN). Symptom profiles differed remarkably between the groups of patients with different grades of QoL impairment, i.e. QoL impairment was directly correlated with the broadness of the spectrum and the severity of the symptoms. It is quite obvious that a patient with advanced cancer rarely experiences a single symptom. Symptom burden is the result of co-occurrence of a number of symptoms. In conclusion, improving a patient's QoL is a key issue in treating advanced cancer patients. Target treatment is impossible without accurate information about QoL impairment. Identification of the grade of QoL impairment is the starting point in the management of patients with advanced cancer. Understanding of symptom interference with a patient's QoL provides the basis for better cancer outcome measurement.

Individual Abstract Number:1491 National Multi-Center Epidemiological Study on Prevalence and Treatment of Cancer-Related Symptoms Tatyana I. Ionova, Quality of Life Department, National Cancer Center, St.Petersburg, Ru, St. Petersburg, Russia, Andrei A. Novik, Hematology, National Pirogov Medical Surgical Center, Moscow, Russia, Svetlana A. Kalyadina, Quality of Life, Anton V. Kishtovich, Quality of Life Department, National Cancer Research and Treatment Center, Shelly Wang, Charles S. Cleeland, Symptom Research Department, The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA To have a better understanding of the current status of symptom management, the national multi-center epidemiological study among patients with advanced cancer was initiated in Russia in 2005. Specific goals of this study were to describe the prevalence of pain, fatigue, and other cancer-related symptoms among cancer patients being treated at multiple sites across Russia and to assess current symptom treatment and response to it among different centers. The total sample size is 500 patients with advanced cancer from 10 cancer centers across Russia. Symptom status was assessed using the M.D. Anderson Symptom Inventory and Brief Pain Inventory at two time points. The analysis of data from 352 patients, male/female ratio 151/201, mean age 44.3 (SD 30.1), included in the study from all participating centers is presented. As a result, the most severe symptom was fatigue (4.3) with its prevalence of 83.7%. Other pronounced symptoms were pain (3.4), distress (3.4), and sadness (3.1) with their prevalence of 66.7%, 67.7%, and 68.7% respectively. Out of the total sample, 70 patients (27.7%) had only mild symptoms and 179 patients (71%) reported at least one symptom as moderateto-severe. Among them there were 72 patients (40.2%) who had 5 or more moderate-to-severe symptoms and 10 patients (5.6%) who had 10 or more such symptoms. Fatigue was treated only in 66 (26.9%) patients; the decrease in fatigue level was achieved in 15 (22.7%) patients. Inadequate pain management (as confirmed by the negative value of the Pain Management Index) was provided for 38.5% of patients. In conclusion, fatigue was prevalent in the vast majority of advanced cancer patients, but the attending oncologists offered the

Individual Abstract Number:1479 The Impact of Fatigue on the Quality of Life of Cancer Patients Henning H. Flechtner, Dept. of Child and Adolescent Psychiatry, University of Cologne, Cologne, NRW, Germany The impact of fatigue on the quality of life of oncology patients is substantial and under recognized. Fatigue in these patients may begin with a simple decrease in physical activities, but can progress to include a wide range of negative effects that often culminate in patients feeling out of control, lonely, and isolated. According to the general paradigm of cancer survivors and understanding, in general, surviving cancer patients experience some limitations after the end of treatment but ultimately attain a reasonably good level of functioning. An examination of subpopulations and further analyses of data suggest, however, different recovery patterns which may not be in accordance with the above mentioned principle but reveal various patterns with changing levels of functioning. Disturbingly, 60% of the survivors in our population of patients with Hodgkin's disease, who were treated in recent trials of the German Hodgkin Study Group and the European Organization for Research and Treatment of Cancer Lymphoma Group, have medium to high levels of fatigue after 5 cancer-free years and in conjunction with these findings show severe limitations in various areas of subjective quality of life. Fatigue levels are related to a number of variables including medical

A-6

and psychosocial factors. From cross-sectional together with longitudinal data it can be concluded that at least for a subgroup of survivors (and a much smaller group of control persons) a high risk for developing fatigue levels above a relevant threshold can be identified. From the available data it can be hypothesised that the "high level fatigue" subgroup is the group of survivors with the greatest risk for negative psychosocial consequences and maladaptation sequelae in various aspects of quality of life. Investigations are essential to determine the current status of longterm survivors in more detail and to link that status to conditions observed during the treatment of acutely ill patients. Data from different tumor entities, including testicular cancer and breast cancer, suggest furthermore that these findings may be different and varying according to tumor type.

Individual Abstract Number:1918 Making Progress: Values and Preferences in American College of Chest Physicians Guidelines Holger Schünemann, Associate Professor, Italian National Cancer Institute Regina Elena & McMaster University, Hamilton, Canada This session will review the methods, challenges and solutions to integrating patients values and applied to real life examples of clinical practice guidelines based on more than 20 years of clinical practice guidelines development by the American College of Chest Physicians (ACCP). The examples will be based on solving problems for clinical practice for clinicians and provide in introduction to the problem. Individual Abstract Number:1919 The Inclusion of Quality of LIfe and Health-Realted Preference Data in the Development of Nice Guidance (formerly The Patient Perspective in Nice Guidelines) Louise Longworth; Guidelines Technical Advisor, National Institute for Health and Clinical Excellence UK, London, UK

Individual Abstract Number:1654 Monitoring Self-Reported Quality of Life Amongst Patients Attending a Palliative Medicine Outpatient Clinic to Aid Treatment Decision Taking Samuel M. Salek, WSP Centre for Socioeconomic Research, Cardiff University, Cardiff, UK

The National Institute for Health and Clinical Excellence (NICE) is the independent organisation responsible for providing national guidance to the National Health Service on the promotion of good health and the prevention and treatment of ill health. The NICE Technical Appraisals programme issues guidance on the use of new and existing technologies based on an appraisal of their clinical and cost effectiveness. Consideration of the impact of a technology on quality of life is an important component of the assessment and appraisal of the evidence for each Technology Appraisal and it is recommended that health benefits should be measured in quality adjusted life years for cost effectiveness analyses. The inclusion of quality of life and health-related preference data raises a number of challenges in the development of NICE guidance.

Evidence for the efficacy of a palliative medicine outpatient clinic has not been reported. Studies in palliative medicine are difficult because the cases are clinically heterogeneous and attrition rates are high through disease progression and patient death. Evaluation of the efficacy of any clinic services is notoriously difficult and it would be ethically problematic to randomly select patients for a control group. The Revised McGill Quality of Life Questionnaire (R-MQOL), designed to assess all five domains (physical well-being; physical symptoms; psychological; existential; support) in this population, was used to monitor patients attending outpatients. Over a two-year period, 219 patients attended the weekly palliative medicine outpatient clinic at the regional oncology centre. All were asked to complete R-MQOL in the waiting room, prior to consultation. There were 906 palliative care consultations during two years, for which 64% had R-MQOL completed. 138 patients attended the clinic for the first time, of whom 77 completed R-MQOL on their first and followup visits. Mean time from first visit to follow-up was 5 weeks (range 1-59). Mean time from first visit to death of others was 9.5 months (range 1-30). Patients' mean age was 59 years (range 26-84). 71% were female and 29% required help to complete R-MQOL. Patients reported improved global QoL at first follow-up self-assessment (p=0.002), improved physical well-being (p=0.004), and an improved psychological domain (p=0.01) (sadness, depressed, worried, and afraid of the future). The trends toward improvement in the existential (p=0.25) and support (p=0.70) domains may have reflected the ceiling effects from the items as the initial scores were high. Patients attending palliative medicine outpatient clinic showed a surprising improvement in their reported QoL from first to second visits, despite their progressing disease and impending death. This efficacy suggests that clinic attendance may be a cost-effective contribution in improving patients physical and psychological well-being.

Individual Abstract Number:1920 Overcoming Challenges with Integrating Values and Preferences in Sign Guidelines Robin Harbour, Quality and Information Director, Scottish Intercollegiate Guideline Network, Edinburgh, Scotland This presentation will start by looking at the influence of patient concerns on stable angina guidelines produced by SIGN in 1998 and 2006. It will go on to look at what has changed in the interim to enhance patient input to the guideline process, and the influence this has had on the published guidelines. The author will review some of the recent developments and highlight those that have worked well, and those that have been less effective. Remaining issues and barriers to the effective use of patient experience to produce patient focused guidelines will be considered.

Symposium 1917

Individual Abstract Number:1922 Values, Preferences and Guidelines: The Long Road to Helping Clinical Decision Makers Holger Schünemann, Associate Professor, Italian National Cancer Institute Regina Elena & McMaster University, Hamilton, Canada

Integrating Values and Preferences in Clinical Practice Guidelines: How Can We Make It Happen? Holger Schunemann, Istituto Regina Elena, Italian Nat'l Cancer Inst Rom, Rome, Italy, Louise Longworth, PhD; Guidelines Technical Advisor, National Institute for Health and Clinical Excellence UK, London, UK; Robin Harbour, Quality and Information Director, Scottish Intercollegiate Guideline Network, Edinburgh, Scotland

This presentation will review and summarize methods used and ways forward for values and preference assessment for guideline efforts. It will describe the results of a review conducted for the World Health Organization and suggest ways forward how health related quality of life researchers and those assessing values and preferences can contribute to improving the quality of guidelines. Sponsored by AstraZeneca, Sweden

A-7

Individual Abstract Number:1302 Use of Population-Based Cancer Registries to Assess and Improve Health-Related Quality of Life: The SEER-MHOS Project Steven B. Clauser, Outcomes Research Branch, National Cancer Institute, Bethesda, MD, USA

Symposium 1133 Enhancing the Quality of Cancer Care Through Research, Practice, and Policy Ron D. Hays, General Internal Medicine/Health Services Research, UCLA, Los Angeles, California,, Bryce B. Reeve, Outcomes Research Branch, National Cancer Institute, NIH, Steven B. Clauser, National Cancer Institute, Neeraj K. Arora, Outcomes Research Branch/ARP/DCCPS, National Cancer Institute, USA, Anne Bredart, Supportive Care Department, Institut Curie, Jordi Alonso, Health Services Research Unit, Institut Municipal d'Investgacio Medica

AIMS: Demonstrate the utility of linking population-based cancer registry data with HRQOL survey data to evaluate health care performance and identify quality improvement opportunities for treatment of cancer patients and survivors. Implications of linking Surveillance Epidemiology and End Results (SEER) cancer registries with Medicare health outcome surveys for Medicare quality improvement studies will be discussed. METHODS: U.S. National Cancer Institute SEER cancer registry data were linked with the Medicare Health Outcomes Survey (MHOS) that includes the SF-36 health survey, self-reported health conditions, and sociodemographic information collected from 1998-2002. The MHOS is administered annually to a random sample of 1,000 Medicare enrollees in each Medicare participating managed care plan. MHOS respondents were successfully linked to eleven SEER cancer registries in the MCO plan market areas across the U.S. SEER contains detailed data on cancer diagnosis, stage, and initial cancer treatment. Some 35,000 MHOS respondents identified in SEER as being diagnosed with cancer were aged matched to an equal number of non-cancer plan enrollees who also took the MHOS survey. RESULTS: Data comparing the SF-36 scale and summary scores for those with and witout cancer will be presented. Negative effects of cancer on SF-36 scores not fully explainable by age, time since cancer diagnosis, and other comorbidities are hypothesized. We will also examine the associations of type of cancer and stage of disease with SF-36 scores. Comparisons will outline which cancers and conditions have the lowest HRQOL and where opportunities for improvements in managing HRQOL exist. CONCLUSIONS: Population-based linkages of cancer registry and survey data allow opportunities for evaluating clinically relevant differences in HRQOL in managed care programs. These assessments can be useful in planning and improving the quality of care cancer patients receive.

This panel session will highlight key international initiatives to improve the quality of cancer care. The presentations will examine the use and value of incorporating patient reported outcomes (i.e., health-related quality of life and evaluations of health care) in multiple arenas of application such as cancer clinical trials, evaluation of the quality of provider-patient communication, patient satisfaction with care, and population-based monitoring of cancer patients and survivors to inform research and policy. Presentations will describe projects conducted in Europe, Taiwan, and the United States. The presenters and the discussant will offer insights about the implications of the body of work for international quality of cancer care research, practice, and policy. Individual Abstract Number:1276 Patient-Reported Outcomes Assessment in Cancer Clinical Trials Bryce B. Reeve, Steven B. Clauser, Outcomes Research Branch, National Cancer Institute, Joseph Lipscomb, Department of Health Policy and Management, Emory University, Atlanta, GA, USA AIMS: There is not yet a broad-based, empirically grounded understanding about ways that health-related quality of life (HRQOL) findings from clinical trials inform cancer decision making. This presentation will summarize findings of a National Cancer Institute (NCI) conference that will: examine the use and value of HRQOL in clinical trials; identify challenges to more frequent or extensive application; and identify research opportunities to enhance the decision relevance of HRQOL measurement in clinical trials. METHODS: In September 2006, the NCI will host the conference, "Patient-Reported Outcomes Assessment in Cancer Trials: Evaluating and Enhancing the Payoff to Decision Making." As background, an international group of clinical researchers will prepare case studies using data from previous clinical trials to examine issues for the planning, application, and analysis of HRQOL endpoints in treatment and symptom management cancer trials. At the conference, researchers will report on HRQOL "success stories" as well as on trials that afford learning opportunities on how to improve the application of HRQOL in randomized cancer investigations. Throughout, there will be an emphasis on approaches to improve the value-added of HRQOL assessment. RESULTS: For each type of clinical trial, the case studies will examine: the decision whether to include a HRQOL measure in a trial; planning the data collection and analysis; challenges encountered during conduct of the trial; and the analysis, interpretation, and reporting of HRQOL data following the trial. Further, the "culture" that exists within the group sponsoring the trial, in terms of resources, perspectives, and policies, will be reviewed for its impact on application of HRQOL data. CONCLUSIONS: Findings will examine when, where, and how the measurement of HRQOL in trials can yield valuable information for decisions about cancer care. The conference will identify "best practices" for the application of HRQOL measures in a range of cancer clinical trials. Results will also inform the research agenda on improving the application of HRQOL measures in clinical trials generally.

Individual Abstract Number:1319 Linking Patient-Centered Communication in Cancer Care with Patient Health Outcomes Neeraj K. Arora, Outcomes Research Branch/ARP/DCCPS, National Cancer Institute, USA AIMS: Patient-centered communication (PCC) between health care providers and patients is a key indicator of cancer care quality. To understand the full impact of PCC, it is important to examine the relationship between PCC and patient outcomes such as healthrelated quality of life (HRQOL). However, psychometrically sound measures of PCC are lacking and only few studies have examined the relationship between PCC and patient health outcomes. In this presentation, we will discuss the U.S. National Cancer Institute's (NCI) efforts to facilitate 1) advances in measurement of patient reports of communication with their providers, and 2) examination of specific mechanisms by which PCC might impact patients' HRQOL. METHODS: We will describe a conceptual framework for studying the relationship between PCC and patient outcomes at every phase of the cancer continuum. We will highlight conceptual and methodological challenges in linking patients' assessment of PCC with their health outcomes. Finally, we will present data from an ongoing study of 774 cancer survivors, diagnosed with leukemia, colorectal, and bladder cancers, 2-5 years prior to the study; we tested measures of patient reports of their communication with their physicians in this study. RESULTS: Evidence for the psychometric adequacy of our measures of PCC was obtained (e.g., Cronbach's

A-8

alpha for the information exchange scale was 0.81 and 0.90 for the decision-making scale). Survivors whose physician had a greater participatory decision-making style also reported significantly more positive health appraisal in terms of improved self-efficacy, perceived control, and decreased uncertainty (p<0.01 for all). Mediation models examining the relationship between PCC, health appraisal, and HRQOL are being tested and will be presented at the conference. CONCLUSIONS: The NCI has identified the assessment, monitoring, and improvement of the delivery and impact of PCC across the cancer continuum as a key research priority. This presentation highlights the NCI's current efforts for advancing the state-of-the-science in this very important research area.

Symposium 1792 Patient-Reported Outcomes: Update on the Mayo/FDA Joint Meeting Publications Jeff A. Sloan, Health Sciences Research, Mayo Clinic, Rochester, MN, Marlene H. Frost, Women's Cancer Program, Amylou Dueck, Health Sciences Research, Mayo Clinic, Jeff A. Sloan, Health Sciences Research, Mayo Clinic, Rochester, MN, USA This symposium will provide an overview of the five manuscripts produced as part of a joint Mayo Clinic/FDA project to disseminate and operationalize information relating to issues surrounding the recent FDA draft guidance on patient-reported outcomes (PROs). The manuscripts were produced by a team of 25 experts in PRO outcomes drawn from academia, industry, regulatory agencies, and other stakeholders. Each of the five papers will be summarized in fifteenminute presentations, followed by fifteen-minute discussion period. The five manuscripts, which will be published in a peer-reviewed journal along with a manuscript authored by FDA representatives, and other manuscripts describing PRO assessment in Canada, Europe, and from the patient perspective. The five presentations on PROs will summarize: 1)Conceptual Issues 2)Designing a Measurement Strategy 3)Instrument Selection Issues 4)Sufficient Evidence for Reliability and Validity 5)Interpretation and Presention Presentations will include material from the draft manuscripts as well as feedback from A review process incorporated into the manuscript development and responses from the FDA. CO-Chairs for the Mayo/FDA meeting, Jeff A. Sloan, Ph.D. and Marlene H. Frost, Ph.D. will represent the group of authors on this state of the science project.

Individual Abstract Number:1344 Discussion of Presentations on Enhancing the Quality of Cancer Care through Research, Practice, and Policy Jordi Alonso, Health Services Research Unit, Institut Municipal d'Investgacio Medica The implications of the presentations for cancer research, clinical practice, and policy will be discussed from an international perspective. Individual Abstract Number:1206 Cross-Cultural Assessment of Hospital Patient Satisfaction and Applications in Oncology Anne Bredart, Laure Copel, Supportive Care Department, Institut Curie, Florence Joly, Medical Oncology, Centre François Baclesse, Sylvie Dolbeault, Supportive Care Department, Institut Curie, Andrew Bottomley, Quality of Life Unit, EORTC, Brussels, Belgium AIMS: The assessment of patient satisfaction with cancer care is of particular importance because cancer and its treatment affect patients' functioning and well-being. Cancer care involves alternative treatment options with differing effects, requiring attention to patients' values and preferences within the context of patient-centered care. This presentation will describe an instrument to assess patient satisfaction with hospital care in oncology, the EORTC INPATSAT32, and summarize its use in a pilot study aimed at evaluating supportive care for cancer patients. METHODS: The EORTC IN-PATSAT32 is a 32-item satisfaction with care questionnaire measuring patients' appraisal of hospital doctors and nurses, as well as hospital organization and services. Psychometric properties of the IN-PATSAT32 were evaluated in a sample of 762 patients in oncology settings in France, Southern and Northern Europe, and Taiwan. The IN-PATSAT32 was also administered to evaluate a supportive care department (DISSPO) of a major cancer center in France. The survey was completed at baseline and 2 months by a sample of cancer patients referred to the DISSPO in comparison to a matched control group (n = 100 pairs). RESULTS: Support for the reliability and validity of the EORTC IN-PATSAT32 was obtained in a sample of 647 patients (85% response rate) with a mean age of 57 years (range 19-91) and including 59% female and 52% with more than a compulsory educational level. Institutional factors and geographical/cultural origin were strong predictors of patient satisfaction ratings. In the pilot study evaluating the DISSPO with the IN-PATSAT32 as an indicator, preliminary analyses highlight significantly lower satisfaction with nurses and the interpersonal skills of other personnel in DISSP0 than control patients at baseline. CONCLUSIONS: The EORTC IN-PATSAT32 appears to be a psychometrically robust questionnaire that should be useful in the quality of oncology service evaluation and research.

Individual Abstract Number:1795 Instrument Selection: What Is Sufficient Evidence for the Reliability and Validity of Patient-Reported Outcome Measures? Marlene H. Frost, Women's Cancer Program, Mayo Clinic, Rochester, MN, USA This is part of a series of manuscripts dealing with patient-reported outcomes (PROs) in clinical trials. At its most elemental, patient reported outcomes (PRO) assessment involves asking the patients questions and evaluating their answers. Instrument developers need to be clear about what they want to know, from whom they want to know it and why, if what they learned is credible, and whether they can interpret what they learned in the context of the research objectives. Because credible instrument development is neither inexpensive nor technically trivial, researchers must first determine that none of the available measures meet their research objectives. We suggest that the tasks of either reviewing current instruments or developing new ones originate from the same basic premise: PRO assessment requires a well-articulated conceptual framework. Once defined in the context of the research objectives, the conceptual framework needs to be adapted to the population of interest. We discuss how qualitative methods enrich the conceptual framework and facilitate the technical measurement tasks of item development, testing, and reduction. We discuss how innovations such as item banks and computer adapted testing will influence PRO instrument development. While items are the essential building blocks for instruments, scales are the primary unit of analysis for PRO assessment, and methods for scoring and combining them are discussed. Finally, PRO assessment is meaningless if the key figure chooses not to cooperate. We consider how respondent burden influences the quality of PRO assessment This paper also focuses on the necessary psychometric properties to support use of PROs in drug evaluation. We discuss the kinds of evidence needed to indicate the PRO measures has a sufficient level of reliability and validity, when existing evidence is sufficient , psychometric evaluation approaches

A-9

when a measure is revised, and the applicability to different phases of clinical trials.

The second presentation will give the results of a systematic review of studies that examined the relationship between cancer patients' self-reported HRQOL parameters and length of survival. This talk will basically give the audience a detailed view of the amount of work that has been done in the oncology field. The third talk will provide the results of a large scale project recently conducted by the European Organization for Research and Treatment (EORTC) Quality of Life Unit, which has investigated the prognostic value of HRQOL data in a number of advanced disease cancer populations (including breast, colorectal and non-small cell lung cancer). The last presentation will provide results of a large study undertaken by the University of Bristol and will discuss its findings in light of other evidence for the prognostic value of HRQOL scores for survival and other clinical outcomes in gastrointestinal cancer patients. Although still much has to be done in this area, for example specific prognostic HRQOL parameters have to be identified in specific sub-groups of patients, there are a number of potential clinical implications. For the time being, one of the main strengths of these results is that of providing a strong evidence-based rationale for including HRQOL assessment into routine clinical practice.

Individual Abstract Number:1796 Patient Reported Outcomes: Conceptual Issues and Measurement Strategies Amylou Dueck, Health Sciences Research, Mayo Clinic, Rochester, MN, USA This paper is part of a series of manuscripts dealing with the incorporation of patient-reported outcomes into clinical trials. Specifically, this manuscript deals with conceptual issues related to assessment of PROs. The four questions addressed included: 1. What is a PRO concept? 2. What is an adequate conceptual framework 3. What are the consequences of proceeding with instrument development without a well-established conceptual framework? 4. Do specific conceptual issues vary or evolve by phase of medical product development (i.e. phase 1 to IV trials)? Another focus of this paper is on the design of a measurement strategy related to the utilization of patient reported outcomes (PROs) in support of a labeling claim. PROs may be useful in adding critical information necessary to support a labeling claim that cannot be obtained through collection of other types of data. We will discuss the process involved in arriving at the measurement strategy, selection of appropriate instrument or instruments, and issues regarding revalidation of instruments.

Individual Abstract Number:1345 Health Related Quality of Life as Prognostic Factor for Cardiovascular Events Marco Bobbio, Department of Cardiology, "Arturo Pinna Pintor" Foundation, Turin, Plinio Pinna-Pintor, Department of Cardiology, "Arturo Pinna Pintor" Foundation

Individual Abstract Number:1798 Analysis, Interpretation, and Reporting Results Based on Patient-Reported Outcomes Jeff A. Sloan, Health Sciences Research, Mayo Clinic, Rochester, MN, USA

Since the sixties physchological distress (ie Type A personality) was studied as risk factor of coronary artery disease (CAD). Most of recent investigations are addressed to analyse the bi-directional relationship between phychological distress and CAD: each one has been observed to cause the other. In this presentation we will focus our attention to the role of depression and quality of life as risk factor of cardiac events in patients with and without CAD. In two prospective cohort studies of 4,493 and 11,216 subjects free of cardiovascular disease at baseline, after 6 and 18 years respectively, a strong association was found between depressive symptoms and the long term incidence of CAD and total mortality. In one study it was observed that depression was associated more strictly with stroke than with myocardial infarction. Besides depression is a common comorbid condition in CAD patients. In two recent meta-analysis post myocardial infarction depression was demonstrated to be strongly associated with patient-reported health status affecting symptom burden and quality of life, with a 2- to 2.5-fold increased risk of impaired cardiovascular outcome. Furthermore, post myocardial infarction depression affects return to work and the utilization of medical resources. Subjective general health and self esteem were significantly related to death in patients who experienced a cardiac event in one study and in another study hospital readmissions were higher in patients with higher scores of SF-36 physical functioning, general health and mental health. The role of quality of life as predictor of congestive heart failure or of depression as predictor of events in patients with implantable cardioverter-defibrillator has been already studied. So far several investigations correlated quality of life and depression with cardiac events in patients with different cardiac disease. However few data are available on the efficacy of interventions focused to modify life style behaviors or to treat depression.

This paper is part of a series of manuscripts dealing with the incorporation of patient-reported outcomes (PROs) into clinical trials. The issues dealt with in this manuscript concern the common pitfalls to avoid in statistical analysis and interpretation of patient-reported outcomes. Specifically, the questions addressed by this manuscript involve: What are the analysis landmines with PRO data in clinical trials and how can they be avoided (e.g., missing data, multiplicity, etc)? What are the best (alternative) methods to assess clinical significance of PROs? How should null results be interpreted? How do you present PRO data most effectively in an FDA application? In labeling? The manuscript provides key literature resources for existing resources and proposes new guidelines where appropriate.

Symposium 1147 The Value of Health Related Quality of Life (HRQOL) Parameters in Predicting Clinical Outcomes in Cardiology and Oncology Fabio Efficace, Quality of Life Unit, EORTC Data Center, Brussels, Belgium, Marco Bobbio, Department of Cardiology, "Arturo Pinna Pintor" Foundation, Jane M. Blazeby, Department of Social Medicine and Clinical Sciences, University of Bristol, UK, Bristol, UK, Carolyn Gotay, Cancer Research Center of Hawaii, University of Hawaii, Honolulu, Hawaii, USA The aim of the symposium is to provide a state of the art perspective of studies investigating the prognostic value for clinical endpoints of patient reported outcomes (including HRQOL) in two major areas of medicine: cardiology and oncology. The first presentation will address the role of depression and HRQOL parameters as risk factor of cardiac events in patients with and without coronary artery disease. In addition, major studies in the area of cardiology will be discussed.

Individual Abstract Number:1472 The Prognostic Value of QOL Ratings in Cancer Patients: A Systematic Review Carolyn Gotay, Cancer Research Center of Hawaii, University of Hawaii, Honolulu, Honolulu, Hawaii, USA

A-10

Aims: This presentation provides a systematic review of studies that examined the relationship between cancer patient-reported quality of life (QOL) and length of survival. The primary question underlying this analysis was whether QOL ratings provide additional information beyond that supplied by traditional biomedical parameters in predicting how long patients live following a diagnosis of cancer. Methods: All published research reports were examined that assessed QOL ratings in relation to patient survival and controlled for biomedical variables (e.g., stage of disease, tumor size). Data were abstracted and reviewed independently by four raters. 66 papers were found, published 1986 through 2005. The most common cancer sites were breast (n=13) and lung (n=14), but many other sites were studied. The most frequent QOL measures were QLQ-C30 (n=32) and FACT (n=8). Most reports included both univariate and multivariate analyses. Results: Despite considerable variability in samples and methods, 60 of 66 studies found that QOL ratings were related to survival time in multivariate analyses that controlled for biological variables. QOL often had a stronger relationship to prognosis than standard clinical predictors. Multiple explanations are possible: patients may perceive factors that cannot be detected by clinical indicators, QOL may be linked with health behaviors that affect survival, and/or positive attitudes affect survival through biological factors. Conclusions: Patients' own perspectives may portend their clinical course. QOL ratings deserve consideration in determining eligibility and stratification for clinical trials, and in clinical decision-making.

Individual Abstract Number:1398 Quality of Life Parameters as Prognostic Factors in GastroIntestinal Cancer Jane M. Blazeby, Department of Social Medicine and Clinical Sciences, University of Bristol, UK, Bristol, UK There is growing evidence to support the value of baseline quality of life (QOL) scores as independent prognostic factors for survival in patients with gastrointestinal cancer. The role of self reported QOL data in predicting other clinical outcomes is less clear. This paper presents the results of a study from Bristol and will discuss the findings in light of other evidence for the prognostic value of QOL scores for survival and other clinical outcomes in gastrointestinal cancer. Consecutive series of patients selected for upper gastrointestinal surgery completed the EORTC QLQ-C30 questionnaire. Multivariable regression models, adjusting for known clinical risk factors, investigated relationships between QOL scores, major morbidity, hospital stay and survival status at six months. 121 of 130 patients completed questionnaires (response rate 93%). There were 29 (24%) major complications and 22 (18%) patients died within six months of surgery. QOL scores were not associated with major morbidity but were significantly related to survival status at six months after adjusting for known clinical risk factors. A worse fatigue score of 10 points (scale 0 to 100), corresponded to a 37% increase in the odds of death within six months after surgery (95 confidence interval 12% to 68%, p = 0.002). Pre-treatment social function scores were moderately associated with hospital stay, (p= 0.021), a reduction in social function by 10 points corresponded to an increase in hospital stay of 0.93 days (95% confidence intervals 0.12 to 1.74). This study confirms the findings of others that QOL scores are prognostic for survival especially for patients with advanced gastrointestinal cancer. The role of QOL in predicting morbidity and mortality after cancer surgery remains uncertain and larger studies are needed. Although QOL scores have important prognostic value, the place of using questionnaires in clinical decision-making remains uncertain. Further work is needed to understand and evaluate the role of using QOL assessments in clinical decision-making

Individual Abstract Number:1342 Investigating the Value of Patient's Self Reported Quality of Life Data in Predicting Survival in Advanced Cancer: Results of a Large Scale EORTC Project Fabio Efficace, Quality of Life Unit, EORTC Data Center, Brussels, Brussels, Belgium Patient's self-reported quality of life (QOL) parameters have various uses in medical research, including supporting clinical decisionmaking by providing the patient's perspective or, as recently highlighted, providing prognostic information for medical endpoints. Several recent studies have shown that QOL parameters can be independent prognostic factors for survival in advanced cancer patients. The studies in this area have used different QOL questionnaires: the Functional Living Index-Cancer (FLIC), the Rotterdam Symptom Checklist (RSCL) and several have also used the European Organisation for Research and Treatment of Cancer, Quality of Life Questionnaire-C30 (EORTC QLQ-C30). Whilst the use of different measures has hindered outcomes comparisons, they do provide complementary evidence of the strong and important association between QOL and survival. The EORTC recently conducted a large scale project to investigate this issue in patients with Breast, Brain, Colorectal, Prostate and Lung cancer. All the studies used the EORTC QLQ-C30 to examine the prognostic value of baseline QOL data. Interestingly, they all identified at least one QOL parameter being prognostic in the final multivariate model in patients with advanced disease. As an example, 'appetite loss' and 'pain' were shown to independently predict survival in a population of metastatic breast and non-small cell lung cancer respectively. Whilst the reason for this association is not yet entirely clear and deserves further investigation, it is likely that patient's judgment on their own underlying health condition provides more accurate prognostic information than those provided by some traditional medical data. A future challenge is using this information with individual patients in daily clinical practice to better help the clinician monitor disease progression and improve accuracy of prognosis. This presentation will provide an overview of recent findings of the EORTC experience in this area of research by also focusing on methodological challenges when interpreting results.

Symposium 1127 The Patient Reported Outcomes Management Information System (PROMIS): A Cooperative Effort to Advance and Enable Clinical Research David Cella, Psychiatry, Northwestern University Medical School, Evanston, IL, Bryce B. Reeve, Outcomes Research Branch, National Cancer Institute, NIH, Bethesda, MD, Dagmar Amtmann, Rehabilitation Medicine, University of Washington, Seattle, WA, Ron D. Hays, Medicine, UCLA, Los Angeles, CA, Richard C. Gershon, Psychometrics and Informatics, Center for Outcomes, Research and Education, Evanston, IL, USA The PROMIS is a five-year cooperative group funded under the United States NIH Roadmap Initiative to re-engineer clinical research. This effort combines the expertise of over 60 investigators and 40 consultants in classical test theory, item response theory, computerized assessment including computer adaptive testing, qualitative methods, clinical trial research methods and clinical practice across a wide range of conditions that affect both physical and mental health. Six primary research institutions, a coordinating center, and many affiliates have been working collaboratively with the NIH to create first-rate pools of psychometrically calibrated questions ("item banks") to measure key generic concepts such as pain, fatigue, physical functioning, social role participation and emotional distress. This symposium summarizes progress after nearly two years of activity. The first presentation by Reeve and Cella

A-11

provides an overview of the PROMIS project including expected goals, the PRO domain framework, the research network development process, and the impact PROMIS may have on future clinical research. The second presentation by Amtmann describes the rigorous qualitative and quantitative procedures that led to the development and finalization of PROMIS items, including considerations on item wording, time frame of response and response options. Then Hays will outline and detail the plan for item calibration and validation of the PROMIS item banks, now that network testing of large item pools is underway. The success of PROMIS, while grounded in solid up-front development and appropriate item/data analysis, ultimately depends upon its value to those in the clinical research community and in clinical practice who wish to study patient reported outcomes. Therefore, Gershon and Bass will conclude by discussing the bulding of PROMIS instruments to serve its primary end-users: investigators conducting clinical research. The versatility and flexibility of the developing information system for other applications will also be highlighted.

To describe the methodology for and preliminary results from building PROMIS item banks that will be used as the bases of Computer Adaptive Tests (CAT) of patient reported outcomes (PROs). We employed an iterative process of item identification, domain assignment, item selection, qualitative item review, and item re-writing. An extensive network of experts from outside PROMIS participated in all stages of the development. First, existing instruments were identified and items were catalogued and reviewed. Second, items were subjected to a ‘binning and winnowing’ process to ensure relevance and pare down the number of items. Items were changed to improve readability, to ensure that items were optimized for CAT, and to make them consistent with preferred response sets. Third, focus groups of individuals with chronic conditions were convened to identify areas in each domain most relevant to participants. New items were written to ensure that all areas important to the research participants were measured. Fourth, at least 5 participants (including persons with low literacy and mild cognitive impairment) reviewed each item and gave feedback regarding clarity and meaningfulness of item time frame, stem, and response options. Items that did not perform as intended in cognitive interviews were dropped or rewritten. The final item pool consisted of 56 items that best represented each of 14 identified subdomains. Because the item pool was too large to be administered to any one sample, different survey forms were developed that contained subsets of items from each subdomain. The rigorous process PROMIS employed in building the item banks resulted in banks of items that (a) meet the needs of clinicians and researchers and (b) address aspects of each measured domain most important to individuals living with chronic conditions. We gained a rich understanding of how items are perceived by a broad range of potential respondents and of the processes that respondents use to select their answers. Findings related to the time frame, response options, and clarity of item stems will be summarized in the presentation.

Individual Abstract Number:1262 The Patient-Reported Outcomes Measurement Information System (PROMIS) Project Overview Bryce B. Reeve, Outcomes Research Branch, National Cancer Institute, NIH, Bethesda, MD, David Cella, Psychiatry, Northwestern University Medical Center, IL, USA As part of the U.S. National Institutes of Health (NIH) Roadmap efforts to accelerate and strengthen the nation's clinical research enterprise, the Patient-Reported Outcomes Measurement Information System (PROMIS) initiative establishes a collaborative relationship between the NIH and seven research sites to build a public domain Internet-based resource - the PROMIS - that will measure key healthrelated quality of life (HRQOL) domains that are common to a variety of chronic diseases. Specifically, the PROMIS project will build item banks and both short form and computerized-adaptive tests for reliable, valid, and efficient HRQOL assessments. This presentation will provide an overview of the PROMIS project including expected goals, domain framework, development process, time line, and potential impact on the research field. Started in September, 2004, the PROMIS Steering Committee meets monthly by phone or face-to-face meetings to develop protocols, review progress, and discuss scientific issues related to the building of the PROMIS. The PROMIS network includes clinicians and researchers both with extensive training in psychometrics, survey research, qualitative methods, and information technologies and with experience in health outcomes measurement. After establishing a HRQOL domain framework, item bank development began for the domains of physical functioning, pain, fatigue, emotional distress, and social role participation. Protocols for qualitative item review (e.g., cognitive testing, focus groups), quantitative item review (using item response theory methods), and sampling plan were formalized and currently under implementation. Having a standardized, validated, and dynamic system to measure patient-reported outcomes efficiently in study participants with a wide range of chronic diseases and demographic characteristics would greatly enhance the clinical research enterprise and facilitate comparisons between research studies. Ultimately, this type of system will be useful in medical practice, for the purpose of measuring treatment response and guiding therapy. Individual Abstract Number:1274 Building the PROMIS Item Banks Dagmar Amtmann, Rehabilitation Washington, Seattle, WA, USA

Medicine,

University

Individual Abstract Number:1292 The Psychometric Evaluation and Calibration Plan Ron D. Hays, Medicine, UCLA, Los Angeles, CA, USA AIMS: The construction and evaluation of item banks to measure health-related quality of life is a fundamental objective of the PatientReported Outcome Measurement Information System (PROMIS) project. This presentation provides an overview of the methods used in the PROMIS item analyses and proposed calibration of item banks for physical functioning, fatigue, pain, emotional distress, and social/role participation. METHODS: Analyses include evaluation of data quality (e.g., logic and range checking, spread of response distribution within an item), descriptive statistics (e.g., frequencies, means), item response theory model assumptions (unidimensionality, local independence, monotonicity), model fit, differential item functioning, and item calibration for banking. RESULTS: In analyses of existing data sets, the PROMIS network evaluated both a general IRT model, Samejima's Graded Response Model (GRM), and two models based on the Rasch model framework, the Partial Credit Model and the Rating Scale Model. Based on these analyses, PROMIS network decided that the GRM offers a flexible framework for modeling the participant responses to examine item and scale properties, to calibrate the items of the item bank, and to score individual response patterns in the PRO assessment. After a review of the item properties including evaluation of differential item functioning across both demographic and clinically different groups, the final selected item set will be calibrated using the GRM and CAT algorithms will be developed. The item pools for each unidimensional PROMIS domain will include item pools with more than 50 items. CONCLUSIONS: PROMIS is faced with the challenge of creating psychometrically sound and valid banks in a short amount of time. Up to 15 item banks will be developed and

of

A-12

over ten disease populations will be represented in the initial calibration sample. The enormity of the project requires flexibility in the methods used. A large scale evaluation phase will follow the initial wave of testing in order to examine alternative methods that may yield maximally interpretable and efficient results.

The group has established a list of 9 Methodology Research Questions, and chosen the first 2 to start: 1. When should an original questionnaire and its translations be revised and updated? 2. Which aspects of the linguistic validation methodologies make a difference? Specific results will be presented. The fourth presentation will be an illustration of how different studies might be designed to evaluate translations. The Obesity and Weight Loss Specific Quality of Life Questionnaire (OWLQOL) is chosen as an example. The cost of documentation (minimal evidence) and for fielding different translations for comparison (maximum evidence) will be presented.

Individual Abstract Number:1753 Building The PROMIS Instruments Richard C. Gershon, Center on Outcomes, Research and Education, Evanston Northwestern Healthcare, Michael Bass, Informatics, Center for Outcomes, Research and Education, Evanston, IL, USA

Individual Abstract Number:1906 Translating PRO Measures - Focus on Translation Difficulties Catherine Acquadro, Mapi Research Trust, Lyon, France

AIMS To describe the properties of the PROMIS Item Banks and how they will be used to enable the creation of short forms and Computerized Adaptive Tests (CAT). METHODS Calibrated item banks enable simple construction of several instrument types. Calibrated item banks can be used to construct long static forms, short forms, or to enable computerized adaptive testing. Researchers and clinicians may also utilize the banks to select items based on unique content interests and formulate custom short-form or fulllength instruments. In every case, using the precalibrated item bank allows the instruments to be prevalidated and produce standardized scores on the same scale. The PROMIS Software system will facilitate researcher access to each of these instrument types. In order to insure easy use of the PROMIS item banks, short forms and CATs we interviewed investigators and study coordinators at each of the primary research sites as to how they develop instruments and implement data collection. The results were compiled into an extensive software specification which includes an online item library, computer based test delivery for fixed and CAT instruments (online, standalone, PDA and integrated voice response), centralized data storage, scoring and reporting.

Background The special interest group (SIG) on Translation and Cultural Adaptation (TCA) had its first meeting in Hong Kong at the Annual ISOQOL meeting in 2004. The main objective of the TCA SIG is to identify and advance research in different areas of interest to all parties involved in the field of Translation and Cultural Adaptation. Three priority areas of research were identified and organised in three subgroups: 1. Translation Difficulties, 2. Equivalences and 3. Methodologies. Objectives The objectives of the subgroup 1 are: 1.To collect, review and analyse examples of translations difficulties in the adaptation of Patient-Reported Outcomes (PRO) measures; and 2. To create a list of ‘good practices’ for developers of questionnaires and translators, based on the review of common translation difficulties and proposed solutions on how to prevent them. Methods The group has organised its activities in 4 steps: 1. Categorisation of difficulties; 2. Collection of examples of translation difficulties; 3. Review of examples; and 4. Recommendations. The collection of examples was organized with the collaboration of five teams of volunteer researchers from China, Germany, India, Italy, and Russia. The review of examples and the categorization of translation difficulties was performed by the chair of subgroup 1, and submitted to the approval of the members of Subgroup 1. Results Problems arising from the translation process fall into four categories: Cultural difficulties, Conceptual / Semantic Difficulties, Idiomatic Difficulties, Grammatical difficulties. Examples from the 5 languages will be presented. Conclusion Four categories of translations difficulties are proposed. To prevent them, we recommend the improvement of the quality of the source versions by using Brislin’s guidelines, and by performing a Translatability Assessment.

Symposium 1905 Translation and Cultural Adaptation of PRO Questionnaires: Review of the Activities of the ISOQOL Translation and Cultural Adaptation Special Interest Group (TCA SIG) Katrin Conway, Mapi Research Institute, Lyon, France, Catherine Acquadro, Mapi Research Trust, Sonya Eremenco, Center on Outcomes, Research, and Education (CORE), Evanston Northwestern Healthcare, Evanston, IL, Mona Martin, Health Research Associates, Seattle, WA, Donald Patrick, University of Washington, Seattle, WA, USA

Individual Abstract Number:1907 Comparison Between Translations of the Same Rating Scale Sonya Eremenco, Center on Outcomes, Research, and Education (CORE), Evanston Northwestern Healthcare, Evanston, IL, USA

The TCA-SIG had its first meeting on October 2004 in Hong Kong. The main objective of the group is to identify and advance research in areas of interest to all parties involved in the field of Translation and Cultural Adaptation. The group is divided in 3 subgroups: 1. Translation Difficulties, 2. Equivalences, 3. Methodologies To illustrate the work performed by the TCA SIG, four presentations are proposed: the first three will review the activities of the 3 subgroups, the fourth presentation will be an illustration of how different studies might be designed to evaluate translations. Subgroup 1 will present examples of translation difficulties and propose a formal categorization. Recommendations to improve the quality of the source questionnaires and to prevent translation difficulties will be made. Subgroup 2 has focused its activities on the creation of a database of response choices from prominent QOL instruments across languages. Comparisons of similar rating scales in the same language for different instruments will be made to assess the differences in their respective translations. The main recommendation of the subgroup is to propose standardizing the translation of the most commonly used ratings scales in PRO research. The objectives of subgroup 3 are to identify and prioritize a research agenda for the development and comparison of methods in cross-cultural research.

Background The special interest group (SIG) on Translation and Cultural Adaptation (TCA) had its first meeting in Hong Kong in 2004. The main objective of the TCA SIG is to identify and advance research in different areas of interest to all parties involved in the field of Translation and Cultural Adaptation. Three priority areas of research were identified and organised in three subgroups: 1. Translation Difficulties, 2. Equivalences and 3. Methodologies. Objectives The objectives of subgroup 2 are: 1. The creation of a database of response choices across languages; 2. The collection and review of existing classification of equivalences; 3. The analysis of equivalences of different items; and 4. The review of cognitive debriefing methodology to achieve conceptual equivalence. It was decided to start in priority with the creation of a database of response choices. Methods The project was planned in a 3 step process: 1. Identify prominent instruments and important rating scales, 2. Collect respective translations for rating scales; 3. Make this database available to ISOQOL members via the website. Results An initial list

A-13

of instruments was proposed: SF-36, FACIT, EORTC, IQOL (urinary incontinence) and OWLQOL (obesity). Permission was granted for the FACIT, IQOL and OWLQOL scales, and is still being sought for SF-36 and EORTC scales. The FACIT rating scales have been compiled for over 50 languages. After ratings scales from other instruments are obtained, comparisons of similar rating scales in the same language for different instruments will be made to assess the differences in their respective translations. Conclusion The main recommendation of the subgroup is to propose standardizing the translation of the most commonly used ratings scales. Next steps would involve an assessment process of different translations of the same source rating scale to determine which is the most accurate translation and requests to developers to provide rating scale translations from their instruments. The goal is to reduce variation in rating scale translations of the same source so that the resulting data can better be compared across studies.

methodological expertise, practical challenges, judgment, and cost are involved. The relative merit of different practices can be evaluated to determine the extent to which translations meet the following criteria: are faithful to the intent of the author(s) of the original instrument, use forms of expression acceptable to respondents, show evidence of linguistic and psychometric validation and are documented. Methods Both randomised and quasiexperimental studies are needed to evaluate linguistic and psychometric validations of translated instruments. The Obesity and Weight Loss Specific Quality of Life Questionnaire (OWLQOL) was chosen to illustrate how different studies might be designed to evaluate translations according to the criteria listed above. The following steps were used for creating over 20 translations of the OWLQOL: initial cross-cultural item development, harmonization, cognitive debriefing, and field-testing. Costs and outcomes from each step were collected and study designs identified for evaluating the merit of the approach taken. Results Conceptual and semantic difficulties were identified in cross-cultural elicitation of items. Adoption of a previously-studied response scale lowered cost. Field testing provided psychometric evidence at considerable expense.The key experimental design recommended includes a small study of forward-only translation against the complete culturally adapted version. Small quasi-experimental designs are suggested for evaluating acceptability to respondents. The cost of documentation and for fielding different translations for comparison will be presented. Conclusion The collection of evidence for different linguistic and psychometric validation practices for PRO instruments is an on-going process. Some cost-cutting strategies can be identified.The relative merit of simultaneous cross-cultural item elicitation versus translation needs increased attention.

Individual Abstract Number:1908 Translating PRO Questionnaires: Review of Methodology Research Questions Mona Martin, Health Research Associates, Seattle, WA, USA Background The SIG on Translation and Cultural Adaptation had its first meeting at the Annual ISOQOL meeting in 2004. Its main objective is to identify and advance research in areas of interest to all parties involved in the field of Cultural Adaptation. Three priority areas of research were identified and organised in 3 subgroups: 1. Translation difficulties, 2. Equivalences and 3. Methodologies. Objectives The objectives of subgroup 3 are to identify and prioritize a research agenda for the development and comparison of cultural adaptation methodologies. Methods The subgroup has organised its work in 3 steps: 1. Establishment of list of Methodology Research Questions; 2. Prioritization of questions; 3. Selection of questions to form work groups. Results A list of 9 Methodology Research Questions was established: 1.When should an original questionnaire and its translations be revised and updated? 2.Which aspects of the linguistic validation methodologies make a difference? 3.In which context should multiple language versions be developed for one country, or across more than one country where the primary language spoken is the same? 4.What should be reported on a given linguistic validation and how should this be presented. 5.What is the relationship between linguistic, conceptual and psychometric equivalences? 6.How many patients are necessary for cognitive debriefing? 7.Should there be grading systems for translations? 8.How should non-English questionnaires be translated? 9.Should there be a different process for cultural adaptation and item development in developing countries? The first 2 questions were chosen as a start: 1. Revisions should be considered carefully in terms of (a)current information;(b)documentation available verifying problem areas in existing translations;(c)methodological improvements warranted; 2.The first steps in evaluating when linguistic methodologies make a difference are compiled in terms of the errors that they allow to be removed and an assessment of the consequences had these errors been allowed to remain. Conclusion Only 2 questions are answered and a lot remain to be explored.

Symposium 1353 The Neuro-QOL Project: Using Multiple Methods to Develop a HRQOL Measurement Platform to be Used in Clinical Research Across Neurological Conditions David Cella, Psychiatry, Northwestern University Medical School, Claudia Scala Moy, National Institutes of Health, National Institute of Neurological Disorders and Stroke, David Victorson, Northwestern University Feinberg School of Medicine, Cindy Nowinski, Center on Outcomes, Research and Education, Evanston Northwestern Healthcare, A Peterman, Psychology, University of North Carolina at Charlotte, Deborah M. Miller, Neurology, Cleveland Clinic, Cleveland, OH, USA Neuro-QOL is a 5-year, multi-site NINDS funded project to develop bilingual (English/Spanish), clinically relevant and psychometrically robust HRQL item banks responsive to the needs of researchers in a variety of neurological disorders and facilitate comparisons of data across clinical trials in different diseases. Traditional measures of disease status do not represent the full impact of these disorders and treatments. Neuro-QOL seeks to develop IRT-based patient reported outcomes of functioning across social, mental and physical wellbeing, which will lead to more efficient, flexible and responsive assessment. HRQL measurement has been incorporated into many recent neurology clinical trials as secondary outcomes. Measurement scales have been developed in various disease settings, but some existing scales have questionable validity. There is little consensus on what methods should be used within or across studies or disease areas. Lack of consensus about best tools and approaches makes it is difficult to compare the relative burden of various neurological conditions. It also makes it difficult to compare the relative benefits of different treatments. Investigators are reluctant to design trials with the primary objective of comparing HRQL, presumably because HRQL outcomes are too subjective, hard to define, complex to administer and too difficult to interpret. The Neuro-QOL project has

Individual Abstract Number:1909 Approaches for Evaluating Practices in Translation and Cultural Adaptation of Pro Instruments Donald Patrick, University of Washington, Seattle, Washington, USA, Seattle, Washington, Catherine Acquadro, Mapi Research Trust, Lyon, France, Katrin Conway, Mapi Research Institute, Lyon, France Objectives Cultural adaptation of a PRO instrument includes linguistic and psychometric validation of translations of original questionnaires, instructions, and manuals. Considerable skill,

A-14

completed several tasks including: identifying criteria for the acceptance of HRQL measures in the neurology community, selecting target neurological conditions (stroke, ALS, epilepsy [adult/pediatric], multiple sclerosis, Parkinson's disease, and muscular dystrophy), identifying relevant items and domains, developing an extensive item library and beginning the iterative process of item bank development. In this symposium, we will provide an overview of Neuro-QOL and its progress, including a description of steps we have taken to collect and synthesize important data from experts and patients which was gathered through surveys, in depth interviews and focus groups.

QOL domains for each disease. Concurrently, we conducted 11 focus groups with patients (or caregivers) to understand what QOL areas are affected by their diseases. We will soon begin individual patient/caregiver interviews to expand upon these findings. We also reviewed existing measures for relevant items to enter into our Neuro-QOL Item Library. A first wave of items (~2700) have been entered and gone through an extensive multi-step review process for preliminary domain assignment. A second wave of items is currently being entered and will undergo the same process. Finally, we performed a keyword search to find the number of articles published in each of the potential domains. Results. Our preliminary domains are Emotional Distress; Perceived Cognitive Function; Physical Function; Social Function; Sexual Function; Fatigue; Pain; Communication and Language Difficulty; Sleep Disturbance; Positive Psychological Function; Personality and Behavioral Change; Bowel and Bladder Function. Conclusion: Next we will synthesize and integrate this disparate data to finalize domain selection.

Individual Abstract Number:1418 Neurology Professionals' Attitudes Toward Patient Reported Outcomes (PROs) for Neurological Conditions David Victorson, Northwestern University Feinberg School of Medicine, Evanston Northwestern Healthcare, Evanston, IL, USA Aims: The identification of important criteria for the acceptance of patient reported outcome (PRO) measures by the neurology clinical research community. Methods: Using a survey we adapted from oncology, we polled the opinions of neurology professionals regarding their knowledge, attitudes, and use of PROs and PRO data in clinical and research settings. Psychometric properties of the modified survey were evaluated with multi-trait scaling and exploratory factor analysis (EFA). Using linear regression analyses, we identified two primary factors that resembled positive or negative viewpoints on PROs. All qualitative responses were examined for recurring patterns and themes. Results: 89 respondents (43% neurologists, 18% physiatrists, 39% other) completed the survey. The majority (91%) had experience as an investigator in a clinical trial and reported having used PROs (56%). Multi-trait scaling revealed different response patterns between neurologists and oncologists. EFA identified two clear factors: Enthusiasm for PROs and Reluctance to adopt PROs. High Enthusiasm was related with the belief that PROs are clinically helpful (p<.00) and being willing to use PRO results to augment practice habits (p<.00). High Reluctance was predicted by a preference for clinical activities over PROs (p<.00), lack of knowledge of PROs (p=.01), and the belief that ambivalent construct and domain definitions limit the use of PROs (p=.01). Qualitative analysis revealed that adequate psychometrics, ease of administration/use, relevant content and considering the patients perspective are important criteria. Conclusions/Relevance: Enthusiasm for and reluctance to embrace PROs is associated with attitudes and beliefs that can be identified and changed in order to promote acceptance of PROs.

Individual Abstract Number:1748 Domain Selection In Neuro-QOL: Synthesizing Data from Patients, Experts and the QOL Literature A Peterman, Psychology, University North Carolina at Charlotte, NC, USA Aims: Neuro-QOL has, as its goal, the development of item banks to be used to compare QOL outcomes across clinical trials for 7 neurological conditions. Choosing domains for bank creation was a primary task of the early phase of the project. It required a balance between two key requirements:1) providing comprehensive QOL coverage across disorders with highly variable impact within conditions (e.g.,stroke) and across conditions (e.g., multiple sclerosis and adult epilepsy) and 2) creating testable item banks. Requirements for the latter include: a) construct unidimensionality; b) construct coverage across the full range of difficulty; and c) a limited number of total items to minimize burden. Method: A 'chair' was designated for each of the 7 conditions: s/he received lists of domains/subdomains identified in data collection with neurology experts, patient focus groups and an exhaustive literature search. These chairs carefully reviewed these multiple data sources to determine the domains/sub-domains mentioned with the greatest frequency and/or assigned the highest importance. Results: Table 1 contains sample information gleaned from four data sources on 2 domains important for the condition ALS. This presentation will illustrate the process of domain selection, with examples of discrepancy and consensus across data sources and conditions. Domains and subdomains with the greatest overlap in recommendations will be chosen for creation of item banks to be tested in patients with one of the 7 identified conditions and in a 'healthy normal' sample. Conclusion: Neuro-QOL is one of several large, cross-disease, item bank projects currently underway. Appropriate domain selection is critical to the project's success and to future use of QOL item banks in clinical trial research.

Individual Abstract Number:1746 Identifying Important QOL Domains in Neurological Disorders Cindy Nowinski, Center on Outcomes, Research and Education, Evanston Northwestern Healthcare, Evanston, IL, USA Aims. An essential step in the development of the Neuro-QOL measurement system is identifying specific QOL domains to include in it. In order to accomplish this, we use data from multiple sources. Methods. We gathered data from:1) the literature, 2) clinical and QOL experts 3) patients and 4) existing measures. Beginning with a comprehensive literature review of the QOL effects of neurological diseases, we then obtained expert input via 2 waves of telephone interviews plus a web-based survey, and patient perspectives via focus groups and interviews. Our first wave of experts (n=44) identified QOL domains they thought important to assess for people with all neurological disorders or only for specific disorders. We also surveyed 89 professionals (43% neurologists, 18% physiatrists, 39% other) on the most important domains to assess for diseases they selfselected. After the diseases to target for Neuro-QOL were chosen, we conducted another 63 interviews (9 per disease) to review symptoms and issues related to previously generated (from the literature review)

Results Domain

Expert Interviews #1

Physical function

(3)

Independ --------ence

A-15

Expert Interviews #2 Eating, drinking, move Functional independ.

Patient MD-QOLfocus expert survey group Loss of (7) mobility fine & daily a motor; gr (8) (3) dressing & driving

Individual Abstract Number:1751 Next Steps For The Neuro-Qol Project: Testing and Refinement of the Final HRQL Instrument Deborah M. Miller, Neurology, Cleveland Clinic, Cleveland, OH, USA AIM: Describe the work that will lead to the completion of NeuroQoL. Methods: To complete Phase I we will analyze several large neurological datasets containing HRQL items to develop an item library of over 3,000 items. Items will be 'binned' or grouped into supra-ordinate domain categories. Next, items will be 'winnowed', to pare down duplicate or redundant items. Next, items will undergo qualitative item review (cognitive interviewing) with English and Spanish speaking patients and experts. Then Phase II will commence. We will test and refine the final instrument with clinical patient samples. This includes two waves of testing item banks and scales from Phase I. Our goal is to produce a generalizable measurement platform so subject inclusion criteria is broad. Adult patients must be 18 years or older, able to read English (or Spanish for those versions), use a touch screen computer and have one of the target conditions. In the 1st wave, all items in all banks and scales will be field tested in five sites per disease. On average, adult patients will answer items from 3 of the 9 generic banks and 1 to 2 of the disease-specific scales. Pediatric patients will complete 3 of 6 banks and 1 to 2 of 3 targeted scales. Subjects will be re-tested six months later to assess sensitivity to change. Next, item bank analysis will begin and short forms of each item bank developed. During the 2nd wave of this phase, a second administration of the instrument will assess test-retest reliability. The core instrument will be constructed by combining each short form into a single questionnaire. The item banks will contain questions representative of a common theme or trait. After the items in that bank have been calibrated using item response theory, each item will be representative of the entire trait. Thus, the 'final instrument' can take several forms: a full-length instrument, short forms for each scale and computerized adaptive testing. A final version of the full-length instrument used in wave 2 testing constitutes the instrument most likely to be used for future clinical trials.

A-16

living with a chronic metabolic bone disorder METHODS: Twentyone participants (11 male, 10 female) with ages ranging from 59-91 years, were interviewed retrospectively about their responses to supplementary HRQL questions, using cognitive interviewing techniques and semi-structured topic guides. RESULTS: Participants discussed undergoing a process of goal setting in achieving a desired level of quality of life. Participants did this by developing reduced goals using their self in the past as a reference point from which to adjust their current expectations for their quality of life. Participants also used other unhealthy people as a reference frame from which upward comparisons were made, thus reinforcing the hypothesis that comparison with others is often used in shaping one's own expectations. Participants tended, however, to avoid making selfcomparisons in the short term or making downward comparisons against other healthy people. CONCLUSIONS: These findings suggest the inclusion of supplementary questions in HRQL questionnaires could be useful to identify biased or obscured results in clinical trials. However, improvements to the design of these supplementary questions are warranted.

ORAL SESSIONS Quality of Life in the Elderly 1 1666 /Self-Reported Cognitive Functioning and Quality of Life in Elderly Patients with Mild Cognitive Impairment: Results of a Surveillance Study Corinna Petersen, Monika Bullinger, Medical Psychology, University Hospital, Hamburg, Germany, Volker von den Driesch, MedicalScientific Productmanagement, Dr. Wilmar Schwabe Arzneimittel, Karlsruhe, Germany, Lutz Frölich, Mental Health, University of Heidelberg, Mannheim, Germany AIMS: Mild cognitive impairment (MCI) is one of the conditions threatening the aging population. Since the detection of impairments in cognition is one of the main symptoms of MCI, which is usually reported by the patients themselves, the subjective representation of mental function is an important diagnostic task. The symptoms of MCI may have a serious impact on quality of life. METHODS: In a surveillance study, n= 2001 patients with mild cognitive impairments were examined before and into 12 weeks of treatment with two herbal preparations, Tebonin® and Neuroplant®. Measurements included standardised quality of life questionnaires. RESULTS: Results showed marked impairment of quality of life compared to the reference population in MCI patients before treatment and higher impairments in quality of life and cognitive functions in patients with low Mini Mental Status Test (MMST) and high Geriatric Depression Scale (GDS) scores. Tebonin® and Neuroplant® treatment improved quality of life significantly especially in patients with more severe clinical depression and cognitive functioning deficits. Self-reported quality of life and self-reported cognitive competence correlated significantly and highly and changes in clinical symptomatology as assessed by the GDS and the MMST correlated with changes in quality of life and cognitive functioning. After matching patients with Neuroplant® vs. Tebonin® treatment, both groups showed comparable improvements of quality of life and cognitive function. CONCLUSIONS: Treatment with Tebonin® and Neuroplant® shows a change of quality of life in self-reported function which, however, can not be causally related to medication because of the nonrandomised design of the study. Nevertheless, results indicate that over 3 months of Neuroplant® as well as Tebonin® treatment, quality of life and self-reported cognitive function may be positively influenced.

1772 /Elderly Disability, Living Arrangements and Quality of Life Yeda A. Duarte, Nursing, University Of Sao Paolo, Sao Paolo, Sao Paolo, Brazil, Maria Lucia Lebrão, Epidemiology, University Of Sao Paulo, Sao Paulo, Sao Paulo, Brazil AIMS: Population aging has happened rapidly in Latin America and the Caribbean.The high prevalence of multiple chronic diseases, its consequences and resulting disabilities among the elderly, challenges the health care system in countries around the world.These chronic conditions are very frequently, the cause of disabilities.This might make them dependent and, in these situations, they have an important impact in the elderly quality of life.The aim of this research is to describe family support for the elderly according to their functional status. METHODS: It was used SABE study (Health, Well-being and Ageing) database, which comprised seven countries of Latin American.A standardized questionnaire was used.Probabilistic sample(n=10906)was used.Functional status of the elderly was measured by the ability to perform activities of daily living and functional limitation was defined as the need for help to perform at least one of these activities.Family support corresponded to the help the elderly received from six categories of living arrangements(LA).Each category was transformed in a binomial variable and considered independent for the analyses. The dependent variables,help received or not,were transformed in dichotomic variable.It was applied step regression logistic model and 5% significant level.It was used odds ratio to measure how improve or decrease the probability to receive help in each LA. RESULTS: Except Argentina, no relatives presence in the LA improve possibility of help, changing between 1.9 in Cuba to 7.9 in Uruguai.In the AIDL, the LA that had older people only showed the most important odds(0.4 in Brazil to 3.4 in Argentina).Children presence in LA showed significant difference too(0.4 in Brazil to 2.3 in México. Nevertheless, in all countries not more than 55% of their needs were met. CONCLUSIONS: These results portray a vulnerable population and a scenario that tends to get worse in the forthcoming years, as the health care system is continually challenged by population ageing.The families do what they can do but this is not enought. The elderly quality of life will be at risk.

1433 /Measuring Expectations of Quality Of Life: How to Improve Health Related Quality of Life Assessment Clare Robertson, Anne L. Langston, Health Services Research Unit, University of Aberdeen, Aberdeen, Scotland, UK, Sally Stapley, Health Sciences, University of York, York, Yorkshire, UK, Elaine McColl, Centre for Health Services Research, University of Newcastle upon Tyne, Newcastle upon Tyne, UK, Marion K. Campbell, Health Services Research Unit, Peter Fayers, Public Health, University of Aberdeen, Aberdeen, Scotland, UK AIMS: Health-related quality of life (HRQL) and other self-reported outcomes are frequently used in clinical trials. Phrasing of questions are, however, often non-specific and the extent to which patients' responses are affected by personal reference frames is unknown. We carried out a quantitative study that showed that many participants contrasted their current status against themselves at an earlier time, or against some self-defined reference group. The reference frame used appeared to impact on HRQL in a magnitude similar to differences regarded as clinically important. A qualitative investigation to explore this in detail was carried out and the results are reported here. The aim of this qualitative study was to investigate the reference frames used when evaluating quality of life in an elderly population

1253 /The Pursuit of True Patient Outcomes from Chronic Disease Self-Management Programs (CDSMP). The Flaws of the Pre-Post Assessment Joanne E. Jordan, Richard H. Osborne, Medicine, University of Melbourne, Parkville, Melbourne, Victoria, Australia

A-17

AIMS: The accurate measurement of self-reported health outcomes using standardised questionnaires presents challenges to both evaluators and providers of programs. This study examined the classical preintervention-postintervention questionnaire assessment and the single post-intervention transition question assessment (participant's overall change i.e. improvement/decline) to determine whether one method more accurately reflected participant reported outcomes from CDSMP. METHODS: A combination of qualitative and quantitative methods in a format described as 'questerviews' was applied in 25 interviews with people who had recently participated in a CDSMP. Five questionnaire items covering skill and technique acquisition were used. For each question participants were asked 'how did you come to that answer?' This qualitative discourse was used as a relative 'gold standard' and compared with questionnairebased preintervention-postintervention and transition question assessments. RESULTS: From 121 question triangulations (i.e. 25 interviewees x 5 questions with 4 missing responses) derived from preintervention-postintervention assessment and the transition question assessment; the minority of participant quantitative outcomes (40%) were found to be internally consistent (i.e.the two assessment methods agreed). Based on participants qualitative discourse the transition question was more reflective of participant reported outcomes from the CDSMP. When considering qualitative discourse as a true indicator of CDSMP benefits, 54 of 121 (45%) responses suggested that the preintervention-postintervention assessment was a flawed indicator of program outcome and the flaw originated from the preintervention assessment. This often related to a respondent's change in perspective. CONCLUSIONS: This study suggests the pre-post assessment has poor validity for the assessment of health education program outcomes. The use of the transition question may result in a more accurate reflection of the impact of the program on individuals.

Local health interventions and policy initiatives will be suggested to ensure that better financial resources and living conditions are available for older adults.

Quality of Life from the Caregivers Perspective 1 1392 /An Empirical Study of Proxy Viewpoint in the Evaluation of Health-Related Quality of Life (HRQL) Chad M. Gundy, Psychosocial Research and Epidemiology, The Netherlands Cancer Institute, Amsterdam, The Netherlands, Neil K. Aaronson, Psychosocial Research and Epidemiology, The Netherlands Cancer Institute, Amsterdam, North Holland, The Netherlands AIMS: There are situations in which patient self-reported HRQL must be substituted by proxy assessments. Pickard & Knight (2005) recently delineated two proxy perspectives: the so-called "proxypatient" (proxy assessment of how a patient would assess himself) and the "proxy-proxy" perspective (the proxy's own assessment of the patient). The purpose of this study was to evaluate these two alternative approaches in terms of bias and agreement. METHODS: The EORTC-QLQ-C30 was administered to a sample of 224 cancer patients and their proxies (i.e., significant others). The proxies were randomly assigned to either the "proxy-patient" or the "proxy-proxy" perspective. Differences between the conditions in patient-proxy bias (i.e., differences in mean scores between patients and proxies) and agreement were evaluated using MANOVA, percentage agreement, and intra-class correlation coefficients. RESULTS: Small yet significant amounts of bias were found in both conditions. However, no significant differences in bias were observed between the two conditions for the 15 subscales or for a summated scale. Contrary to expectations, few differences between conditions were noted in percentage agreement. The only significant difference was a higher level of agreement for the "proxy-proxy" condition on the role functioning scale: exact (43% vs. 27%, p=.01) and global (+/- 20%) agreement (71% vs. 60%, p=.07). The cognitive functioning scale showed a significantly better correlation in the "proxy-proxy" condition (ICC = .56 vs .34; p=.04), with a similar tendency for the physical functioning (ICC = .80 vs .69, p= .08) and role functioning scales (ICC = .73 vs .59, p=.06). Conversely, the diarrhea item performed better in the "proxy-patient" group (ICC = .50 vs .67, p= .05). CONCLUSIONS: Functional EORTC scales may show an improvement in proxy-patient agreement if the proxy is instructed to report his own opinion for those scales. These results deserve further study, as they run contrary to the pattern postulated by Pickard & Knight.

1250 /Factors Associated with QOL of Older Adults in Brazil Lisiane G. Paskulin, Lucila C. Vianna, Nursing, Federal University of Sao Paulo, Sao Paulo, Sao Paulo, Brazil AIMS: Quality of life(QOL)of older adults is particularly important in emerging countries given aging populations and significant poverty.The aim of this study was to explore factors associated with QOL of older people in a Health District in Porto Alegre,Brazil. METHODS: The WHOQOL-BREF, OARS ADL scale and a health and sociodemographic data sheet were administered to a random sample 288 elders through a household survey. Analysis included Pearson correlations and multiple regression analysis. RESULTS: The mean age was 71.3 years (SD=7.5) and 67.4% were female.Twelve variables explained 34% of the variance in QOL.Significant factors included: health, education, comorbidities and opportunities for physical activities.The physical domain explained 40% of the variance of QOL. Factors explaining physical QOL included: comorbidities, followed by education, health, opportunities for physical activities and type of housing.Thirteen variables (including education, medications,and opportunities for leisure activities) explained 29% of the variance on the psychological domain and 10 variables (including education, smoking and comorbidities) explained 21% of the variance on the environment domain.Only four variables were significant predictors of the social domain, explaining just 5% of the variance (including type of housing and opportunities for leisure activities). CONCLUSIONS: While there were different factors associated with QOL of this sample in each domain and overall QOL, there were similarities: variables related to higher perceived health, higher socioeconomic status (e.g., education and living in condos) and healthier habits (e.g., opportunities for leisure activities, not smoking and physical activity).This study has enabled us to identify factors that could be used to plan strategies to improve QOL of older adults in this region.

1375 /An Application of Full-Information Item Bi-Factor Model to Item Reduction: A Brief Depression Measure for Alzheimer's Caregivers Chih-Hung Chang, Buehler Center on Aging, Northwestern University Feinberg School of Medicine, Chicago, Illinois, Sarah Gehlert, School of Social Service Administration, University of Chicago, Chicago, Illinois AIMS: Caring for individuals suffering from Alzheimer's disease and related disorders has profound consequences for family caregivers. Brief, validated, and reliable depression measures are needed to screen for Alzheimer's caregivers at risk. This study aimed to derive a brief, psychometrically-sound and clinically-meaningful depression measure using a multidimensional item response theory (IRT)-based item bi-factor model for graded response data. METHODS: Data were obtained from 1,229 caregivers recruited in the Resources for Enhancing Alzheimer's Caregiver Health (REACH) Project. The REACH Project, established in 1995, is a unique, multi-

A-18

site research program sponsored by the National Institute on Aging (NIA) and the National Institute on Nursing Research (NINR). The Center for Epidemiological Studies - Depression (CES-D) scale was analyzed. The CES-D is a 20-item scale used to assess general psychological impairment and primary depression in four subdomains: 1) depressive affect (7 items); 2) somatic symptoms (7 items); 3) positive affect (4 items); 4) interpersonal relations (2 items). Respondents rate the frequency of 20 symptoms over the past week by choosing one of the four response categories ranging from "0"= rarely, or none of the time" to "3"="most or all of the time". The item bi-factor model for graded response data, implemented in PolyBIF, was used. Each CESD-D item was modeled to load on the overall depression score and its respective subdomain. RESULTS: A total of 8 items were retained, 2 from each subdomain, primarily based on their factor loadings and content coverage. The brief CES-D demonstrated good internal consistency (Cronbach's coefficient alpha=.796) and correlated well with its original scale (r =.935, p<.001). CONCLUSIONS: The 8-item CES-D scale is easy to administer and its brevity can potentially reduce respondent burden in research and clinical settings. This study also demonstrates the value of full-information item bi-factor model for improving the measurement of multidimensional depression.

1374 /Quality of Life of Parents of Children with Cancer Anne Klassen, David Dix, Sheila Pritchard, Maureen O'Donnell, Lori Tucker, Pediatrics, University of British Columbia, Vancouver, BC, Canada, Parminder Raina, Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada AIMS: To describe QOL of parents of children with cancer by comparing their SF-36 scores with population norms. To explore relationship between parental QOL and child characteristics, i.e., age, health status, severity of side effects, prognosis, relapse, and time since diagnosis. METHODS: A convenience sample of 96 parents of children in active treatment at BC Children's Hospital (Vancouver, Canada) from March to August 2004 were asked to complete a questionnaire that included the Short-Form 36, the Functional Status II(R) and various clinical and sociodemographic items. RESULTS: The response rate was 78%. Most questionnaires (n=66)were completed by the mother. Compared with SF-36 Canadian norms [1], parents reported poorer QOL (i.e., mental health, energy, role emotional, social), with effect sizes >.70. Poorer psychosocial QOL was related to poorer child health status (r=.530; p<.001), more severe side effects from cancer treatment (r=-.416; p<.001), and a shorter time since diagnosis (r=.413; p<.001). QOL was not related to the child age, prognosis or whether the child had relapsed. CONCLUSIONS: The outcome in most published studies of parents is anxiety or depression. Few studies have considered the broader concept of QOL. Our findings are in agreement with 2 recent studies using the SF-36. Eiser et al. [2]showed that UK mothers report poorer QOL compared with population norms, and that parent and child's QOL were correlated. Yamazaki et al. [3] showed poorer QOL in Japanese parents compared with parents of healthy children. References 1. Hopman WM et al. Canadian normative data for the SF-36 health survey. CMAJ,2000;163:265-71 2. Eiser C et al. Quality of life in children newly diagnosed with cancer and their mothers. HQLO,2005;3:29. 3. Yamazaki S et al. Health-related quality of life of mothers of children with leukemia in Japan. Quality Life Res,2005;14:1079-85.

1532 /Health and QOL of Hospital Workers: Development of a Staff Questionnaire to Measure Health and Health Determinants in a Setting Approach Ursula Karl-Trummer, Sociology of health and medicine, LBISHM, Vienna, Austria, Ulrike Ravens-Sieberer, Psychology of health, Robert Koch Institute, Berlin, Germany, Christina Dietscher, Robert Griebler, Sociology of health and medicine, LBISHM, Vienna, Austria, Engelbert Hartter, University clinic, Physical medicine and rehabilitation, Vienna, Austria, Sonja Novak-Zezula, Sociology of health and medicine, LBISHM, Vienna, Austria AIMS: The European Agency for Safety and Health at Work states that health of hospital staff is increasingly at risk and asks for attention and action. The basis for the sustainable improvement of working conditions in hospitals is a systematic monitoring of health and well being as defined by the WHO. Led by the Ludwig Boltzmann Institute for Sociology of Health and Medicine, an interdisciplinary team of scientists and a group of practitioners from six Austrian hospitals are working on the development and implementation of a scientifically tested staff questionnaire for measuring the health of hospital staff that can be applied in routine assessments. The presentation will discuss the basic model of the interrelationship of health and the workplace, the developed staff questionnaire and present data from the first survey. METHODS: On the basis of literature review, Delphi and focus groups with relevant professional groups, a questionnaire was developed and pretested in two hospitals. The pretest dataset includes information of 250 employees of different professions and wards. The questionnaire measures self rated health using the SF12 and indicators from the COPSOQ. It measures health determinants in a setting approach (structures, processes, reward systems), including social relations (cooperation, conflicts) and personal resources and contains questions on the acceptance and suitability of the instrument. The associations between well-being, different setting aspects and personal resources was tested by fitting a structural equation model. RESULTS: The pretest results show that the developed questionnaire is highly accepted and suitable. The structural equation model shows that setting determinants have a strong impact on the working demands and consequently on the health of hospital employees, with a small moderating impact of personal resources (self-efficacy). CONCLUSIONS: The developed questionnaire is an appropriate instrument for routine assesments of health and health determinants of hospital staff.

1419 /Colon Rectal Cancer Patients Quality of Life, Measured by Themselves and Proxys Alexandra Oliveira, Medical Oncology, Hospital de São Sebastião, Santa Maria da Feira, Portugal, Francisco L. Pimentel, Secção Autónoma Ciências da Saúde, Universidade de Aveiro, Aveiro, Portugal AIMS: Patient related outcomes are nowadays a major concern in health care providers. Health Related Quality of Life (HRQoL) are one of these outcomes, so it is very important to know that in each particularly cancer diagnosis. Colon rectal cancer is one of the most frequent cancers in Portugal. The aim of our study was to analyse if proxys can provide useful information on HRQoL of colorectal cancer patients. METHODS: We used the EORTC QLQ-C30 in a sample of 79 colon rectal cancer patients and 79 proxys from oncology department of Hospital de São Sebastião, Portugal. The inclusion criteria was having a colon rectal cancer diagnosis and be with a significant other (proxy). The exclusion criteria was having impaired neurological or cognitive state which didn t allow answering the questionnaire. RESULTS: Our sample was composed by 35 female (44,3%), and 44 male (55,7%) patients, with an age average of 64,5 (SD=11,1) and by 60 female (75,9%) and 19 male (24,1%), proxys with an age average of 51,3 (SD=13,8). We found moderate to good agreement (ICC ranged from 0.45 to 0.81 and Pearson correlation ranged from 0.47 to 0.84) between responses from patients and those from their proxies for the QLQ-C30 scales. The results were moderate for role, emotional, cognitive and social functioning, nausea and vomiting, pain, dyspnoea, insomnia, constipation, diarrhoea, financial difficulties and global QoL, and

A-19

were good for physical functioning, fatigue, appetite loss and for total QoL score. Statistically significant differences were found for physical functioning, fatigue, global QoL and total QoL score between patients and their proxies. CONCLUSIONS: In general, proxies rated the patients as having a more impaired QoL than the patients themselves. Significant other rated patients as having higher levels of disability and symptoms. Proxies only provide valid and useful information about cancer patients QoL for QLQ-C30 domains where the correlations were good, however, should be used with caution particularly for physical functioning, fatigue, global QoL and total QoL score.

primarily by lack of translation equivalence or by cultural differences. METHODS: A large international database was assembled of EORTC QLQ-C30 data from over 38,000 respondents from 124 separate studies. Differential item functioning (DIF) analyses using ordinal logistic regression were used to compare responses to the 24 items included in the multi-item scales, first when grouped by the translation used and then by cultural/geographical clusters. Interpretation was aided by comparing the pattern of results when analyses were grouped by translation used and by geo-cultural clusters. The statistical results were also compared to results from structured interviews with bilingual individuals. Additionally, 'factorial' DIF analyses were conducted for bilingual countries to examine simultaneously the effect of translation and cultural grouping. RESULTS: At least one instance of statistically significant DIF was found for each of the 12 translations, relative to English. In a few cases the bilingual interviews suggested a specific translation problem, but in many cases the cause of the DIF was unclear. European results were more clearly explained when grouped by translation than by geographic group, while the results for East Asia suggested that cultural factors may be more important. CONCLUSIONS: DIF analyses can aid in identifying international differences in scale scores as a function of translation inequivalence and/or cultural factors. Such results have implications for international studies using the EORTC QLQ-C30 and other healthrelated quality of life instruments.

Item Response Theory 1 1720 /Evaluation of a Preliminary Physical Function Item Bank Supports the Expected Advantages of the Patient-Reported Outcomes Measurement Information System (PROMIS) Matthias Rose, Janine Becker, Jakob B. Bjorner, QualityMetric, and Health Assessment Lab, Waltham, MA, Barbara Gandek, ., Health Assessment Lab, Waltham, MA, Jim Fries, Stanford University School of Medicine, Palo Alto, CA, John E. Ware, Jr., QualityMetric, and Health Assessment Lab, Waltham, MA AIMS: The Patient-Reported Outcomes Measurement System (PROMIS) was initiated to improve precision, reduce respondent burden, and enhance the comparability of health outcomes measures. We used Item Response Theory (IRT) to construct and evaluate a preliminary item bank for physical function assuming four subdomains. METHODS: Data from seven samples (N=17,726) using 136 items from nine questionnaires were evaluated. A generalized partial credit model was used to estimate item parameters, which were normed to a mean of 50 (SD=10) in the US population. Item bank properties were evaluated through Computerized Adaptive Test (CAT) simulations. RESULTS: IRT requirements were fulfilled by 70 items covering activities of daily living, lower extremity, and central body functions. Items on upper body function, and need for aid or devices did not fit the IRT model. The original item context partly affected item parameter stability, and inflated slope parameters. In simulations, a 10-item CAT eliminated floor and decreased ceiling effects, achieving a small standard error (< 2.2) across scores from 20-50 (reliability >.95 for a representative US sample). This precision was not achieved by any comparable fixed length item sets. CONCLUSIONS: The methods of the PROMIS project are likely to substantially improve measures of physical function and to increase the efficiency of their administration using CAT.

1547 /Correcting for Differential Item Functioning (DIF) Using Two Different Techniques: What Does Enhance the Validity and Diagnostic Quality of Adolescents HRQoL Scores? Michael Erhart, Ulrike Ravens-Sieberer, Child and Adolescent Health, Robert Koch-Institut, Berlin, Germany, Curt Hagquist, Social Sciences, Karlstad University, Karlstad, Sweden, Stephane Robitail, Public Health, University Hospital of Marseille, Marseille, France, Stef van Buuren, Quality of Life, TNO, Leiden, Netherlands AIMS: Comparable HRQoL assessment requires that persons with the same trait parameter value respond to an item with similar probability regardless of age, gender or culture. This absence of DIF is a prerequisite for reasonable conclusions regarding individual HRQoL as well as the circumstances within and between populations. If DIF is rooted in aspects relevant for the trait to be measured however the particular item might be retained. We investigated the impact of DIF correction techniques on the validity and diagnostic quality of the Rasch-scaled physical well-being scores of the KIDSCREEN-52 HRQoL instrument. METHODS: The data of 22000 children and adolescents (8-18 y.) from 13 European countries of the KIDSCREEN study were analyzed. Two approaches for DIFcorrection were compared: In (1) the item parameters of items with cultural DIF were estimated separately per country (Partial Credit Model (PCM)) in (2) the DIF items were excluded. For (1) and (2) the construct validity of the PCM scores were assessed within multilevel regression analysis on health- (SDQ, HBSC-SCL) and social (OSSS) HRQoL determinants. The F-value ratio assessed relative validity (RV). Diagnostic quality was estimated by the area under the ROC curve (AUC) when screening for individuals at risk. RESULTS: Individual changes due to correcting for DIF ranged from -.22 to +.30 SDs for (1) and from -1.78 to +1.35 SDs for (2). Approach (1) leads to small but stat. significant increase of the scores validity to R2=.25 (RV=1.02) and the diagnostic quality to AUC=.77. Approach (2) yields lower validity (R2=.23; RV=.91) and diagnostic quality (AUC=.75). The measured cross-cultural variation (variance partitioning coefficient) increased from .14 to .18 for (1) and .20 for (2). CONCLUSIONS: The results indicated not to exclude HRQoL items solely because of DIF. After examining the item content, an IRT based correction might be employed. Correcting for cultural DIF enhances the sensibility of HRQoL scores for cultural differences.

1332 /Distinguishing Between Cultural and Linguistic Response Patterns in the EORTC QLQ-C30 Using Differential Item Functioning (DIF) Analyses Neil W. Scott, Peter M. Fayers, Public Health, University of Aberdeen, Aberdeen, UK, Neil K. Aaronson, Psychosocial Research & Epidemiology, Netherlands Cancer Institute, Amsterdam, North Holland, The Netherlands, Andrew Bottomley, Quality of Life Unit, EORTC, Brussels, Belgium, Alexander de Graeff, Division of Medical Oncology, Internal Medicine, Academic Medical Centre, Utrecht, Netherlands, Mogens Groenvold, Palliative Medicine, Bispebjerg Hospital, Copenhagen, Denmark, Michael Koller, Center for Clinical Studies, University Hospital Regensburg, Regensburg, Germany, Morten A. Petersen, Bispebjerg Hospital, Copenhagen, Denmark, Mirjam A. Sprangers, Medical Psychology, University of Amsterdam, Amsterdam, Netherlands AIMS: To determine whether there are international response differences within each subscale of the EORTC QLQ-C30 quality of life questionnaire and to investigate whether these are caused

A-20

scores ranged from -6.3 to 3.6 logits (range=9.9 logits) and had an average standard error of 0.69. Correlations between 2-item subscales and the 20-item scale ranged from 0.74 to 0.88 (mean=0.83). For the 4-item subscales, correlations ranged from 0.76 to 0.94 (mean=0.90). The highest correlations were obtained in the content- and difficultybalanced, 4-item subscales. Sample Results Based on the 4-Item Content-Balanced Subscales (a total of 5 subscales): Mean residuals were low (0.04, -0.01, 0.09, 0.12, -0.06). Approximately 60% of residuals were < 0.5 (scale range=10 units). Approximately 5% of the sample had residuals >1.5. CONCLUSIONS: IRT shows promise for the development of multiple brief fatigue subscales equated to the same metric. Four-item subscales based on content- or difficultybalancing performed well in a significant portion of the sample, though fatigue estimates for some individuals varied substantially from those based on the longer scale. Future research should distinguish the impact of measurement error and equating error on subscale scores.

1213 /A Comparison of Adaptive and Standard Questionnaires in Patients with Back Pain Jacek A. Kopec, Health Care and Epidemiology, Maziar Badii, Medicine, University of British Columbia, Vancouver, BC, Canada, Mario McKenna, Eric C. Sayre, Viviane D. Lima, Pamela Rogers, Arthritis Research Centre of Canada, Vancouver, BC, Canada, Marcel F. Dvorak, Department of Orthopaedics, University of British Columbia, Vancouver, BC, Canada AIMS: We developed IRT-based item banks for 5 domains of health-related quality of life (HRQL), namely Walking (W), Handling Objects (HO), Daily Activities (DA), Pain/Discomfort (P/D), and Feelings (F) (ISOQOL 2004, 2005). The objectives of the current study were a) to develop a computerized adaptive test (HRQL-CAT-5) using these 5 item banks; b) to assess feasibility, reliability, validity and efficiency of HRQL-CAT-5 in patients with back pain; and c) to compare HRQL-CAT-5 with standard back pain questionnaires and a generic health instrument. METHODS: An adaptive algorithm was implemented in a web-based questionnaire administration system. The questionnaire included HRQL-CAT-5, modified Oswestry Disability Questionnaire (MODQ), RolandMorris Questionnaire (RMQ), SF-36 Health Survey, and standard clinical and demographic information. Patients diagnosed with mechanical back pain who were 18-74 years of age and able to communicate in English were eligible to participate. RESULTS: A total of 215 patients completed the questionnaire and 84 completed a re-test. We set the number of items per domain in HRQL-CAT-5 to 3-7. The average number of items answered was 5.2. Mean standardized scores (Av=50, SD=10) ranged from 41 (P/D) to 45 (F). Test-retest correlations (ICC) ranged from 0.83 (F) to 0.92 (P/D). The ICC for the MODQ, RMQ, and Physical Component Summary (PCS36) were 0.92. Among the HRQL-CAT-5 domains, DA had the highest ceiling effect (6%). The ceiling effect was 0.5% for P/D, compared with 2% for MODQ and 5% for RMQ. HRQL-CAT-5 discriminated well according to satisfaction with symptoms, duration of current episode, presence of sciatica, and receiving disability compensation. Average relative efficiency (RE) for between-group discrimination was 0.87 for P/D, 0.67 for DA, and 0.62 for W, compared with 0.89 for MODQ, 0.80 for RMQ, and 0.59 for PCS-36. CONCLUSIONS: HRQL-CAT-5 is feasible, reliable, valid, and efficient in patients with back pain. The P/D domain is as reliable and discriminating as the back pain specific measures, and better discriminating than the SF-36.

Endocrinology and QOL 1327 /Long-Term Response of Specific Quality Of Life (QOL) Dimensions to Growth Hormone (GH) Replacement in Elderly Patients With GH Deficiency (GHD) Maria Koltowska-Haggstrom, Anders F. Mattsson, KIGS/KIMS/ACROSTUDY Medical Outcomes, Pfizer Health AB, Sollentuna, Sweden, Stephen M. Shalet, Department of Endocrinology, Christie Hospital, Manchester, United Kingdom AIMS: To determine specificity in QoL in adults with GHD (>60 and <60 years of age) before GH therapy and their response to longterm replacement compared with the normal population. METHODS: QoL was measured by QoL-AGHDA (Quality of Life - Assessment for Growth Hormone Deficiency in Adults) in normal population in England & Wales, the Netherlands, and Sweden and compared with corresponding data from KIMS (Pfizer International Metabolic Database) for up to 6 years' GH therapy. Both pooled population samples and patients were divided into 2 age subgroups >60 (n=1135; 375 at baseline respectively) and <60 (n= 2630; 1336 at baseline respectively) years of age. QoL-AGHDA items were divided into 5 dimensions and responses within dimensions were evaluated in comparison with the population data. RESULTS: Before treatment younger patients reported more problems within all QoL dimensions. Tiredness was the most impaired dimension in both groups (34% more problems in the older group vs 45% in the younger), followed by memory (24% vs 36%), tenseness (21% vs 31%), self confidence (14% vs 23%) and problems with socializing (12% vs 20%). Overall the most dramatic improvement occurred during the first year of GH therapy for all dimensions and in both patient groups. The level of reported problems with socializing reached population values in both patient groups by 2 years of treatment and remained there for 6 years of follow up. After 3 years of GH therapy patients >60 reached population values for all other dimensions. Patients <60 years, despite substantial improvement initially, reached population levels only after 6 years (self-confidence and tenseness) but did not do so for memory and tiredness. CONCLUSIONS: Adults with GHD benefit from GH in all dimensions, measured by QoL-AGHDA, showing a similar pattern of response to treatment irrespective of age; however elderly patients reached "normal values" much earlier than the younger.

1377 /Development of Very Brief Fatigue Subscales Using Item Response Theory Karon F. Cook, Allison Fleming, Kurt L. Johnson, Karynsa Hinton, Dagmar Amtmann, Rehabilitation Medicine, University of Washington, Seattle, WA AIMS: Fatigue is a common, often disabling symptom in multiple sclerosis (MS) and is amenable to intervention. Having brief, accurate tools to assess fatigue is critical for clinical care and research. This study examined the psychometric characteristics of the Modified Fatigue Impact Scale (MFIS) to evaluate whether item subsets could approximate total scale score. METHODS: A sample of persons with MS (n=193) responded to the 21-item MFIS. Responses were calibrated using an item response theory (IRT) model. Subscales consisting of 2 and 4 items were created from 20 of the original items that best fit the model. Several strategies were used to develop the subscales: 1) content balancing, 2) item difficulty, 3) discrimination (most discriminating items) and 4) random selection. More than one subscale was built using each strategy. Subscales were calibrated using IRT and equated by anchoring on the item calibrations from the 20-item scale. Residuals were calculated (subscale score minus 20-item score). RESULTS: IRT-calibrated

1474 /Quality of Life and Course of Life in Young Adults with Metabolic Disease: Remarkable Differences Martha A. Grootenhuis, Pediatric Psychosocial, Academic Medical Center, University of Amsterdam, Amsterdam, Netherlands, Annet M. Bosch, Pediatrics, Academic Medical Centre, University of

A-21

Amsterdam, Amsterdam, Netherlands, Frits A. Wijburg, Pediatrics, University of Amsterdam, Amsterdam, Netherlands

measuring heterogeneous constructs (e.g., mental and physical domains), but high (>0.6;P<0.05) among subscales measuring homogenous constructs, suggesting good convergent & discriminant validity. The C-indexes suggest greater discrimination by the D39-T vs. SF-36 for all laboratory indicators (range: 0.6-0.7 vs. 0.5-0.6) and all diabetic complications. CONCLUSIONS: In this Taiwan sample, the D39-T demonstrates good scale reliability and the same factorial structure as the Western version. The favorable discriminative capability of the D39 relative to SF-36 for all clinical indicators suggests the D39-T may be acceptable for use in clinical practice.

AIMS: The clinical consequences of inborn errors of metabolism have been described in numerous publications. However, few studies have investigated the patients own experience of their disease and the effect of the metabolic disorder on quality of life, and educational outcomes. METHODS: We evaluated Health Related Quality of Life (HRQoL), Course of Life (CoL), and educational attainment of 63 Dutch patients with classical galactosemia (aged 1-41 years), and 53 Dutch patients with PKU (aged 18-31 years). All patients completed validated, age specific, HRQoL questionnaires, patients aged 18 and older also completed a CoL questionnaire (developmental milestones). RESULTS: No differences were found between the PKU patients and the controls on any of the Rand-36 scales or on the TAAQoL cognition scale. PKU patients had normal scores on all CoL scales. More PKU patients than controls attended special schools in primary education. However, patients with galactosemia had a delayed CoL with respect to both social and psychosexual development. More galactosemic patients than controls attended special primary education, and adult patients had a significantly lower educational attainment. Patients aged 8 and older had lower scores on the domain of cognitive function, and patients aged 16 and older had significantly lower scores on the domain of social function. CONCLUSIONS: In spite of the much more complicated diet and follow up of patients with PKU, which was considered to be more detrimental than the relatively simple diet of galactosemic patients, the CoL and HRQoL of patients with PKU appears not to be affected by their disease. In contrast, there is a strong negative effect on the CoL, HRQoL and educational outcome of patients with classical galactosemia. The long term complications, such as cognitive and language problems, appear thus to be more relevant to the HRQoL and CoL than the burden of a strict diet and follow-up. Evaluation and support of cognitive as well as of social development should be part of the follow up of patients with classical galactosemia.

1229 /Validation of the Insulin Delivery System Questionnaire (IDSQ) in Patients with Type 1 Diabetes Risa P. Hayes, Global Health Outcomes, Eli Lilly and Company, Indianapolis, IN, James T. Fitzgerald, Medical Education, University of Michigan, Ann Arbor, MI, Douglas B. Muchmore, Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, IN AIMS: The aim of this study was to validate the Insulin Delivery System Questionnaire (IDSQ), a 14-item instrument designed to provide patient-reported evaluations of insulin delivery systems. METHODS: Preliminary validation of the IDSQ was performed in patients with type 1 diabetes (N=137, 47% male, mean age=39 years, A1C mean=8.1%) participating in a randomized, crossover study comparing preprandial inhaled insulin with preprandial injectable insulin with respect to HbA1c. The IDSQ and Diabetes Treatment Satisfaction Questionnaire (DTSQ) were administered at Visit 2 (before lead-in), crossover, and endpoint. To confirm that insulin delivery system features evaluated by the IDSQ are important to patients with type 1 diabetes, participants were asked to rate the importance (1 [not at all] to 5 [very]) and rank the 12 IDSQ features at the same time points. Factor analysis identified three IDSQ subscales, Insulin Satisfaction (2 items), Ease of Use (5 features), and Lifestyle Impact (7 features). IDSQ subscale scores were correlated with DTSQ scores and change in HbA1C. Repeated measures ANOVAs were used to compare scores across time. RESULTS: IDSQ subscale scores were positively associated with both DTSQ scores at all study time points (r=0.39 to 0.65, p<0.001) and change in HbA1c from crossover to endpoint (r=0.20 to 0.26, p<0.001). IDSQ Insulin Satisfaction and Lifestyle Impact subscale scores of patients randomized to inhaled insulin at baseline improved significantly (p<0.05) with inhaled insulin treatment but decreased significantly (p<0.05) with return to injectable insulin. Mean importance ratings of all IDSQ features was >3.0. Of the 12 features, "Ease of blood glucose control" was ranked highest and dosing features ranked next highest at all study time points. CONCLUSIONS: Preliminary study indicates that the IDSQ is a valid and responsive instrument that provides an evaluation of insulin delivery systems on features that are important to patients with type 1 diabetes.

1718 /Development and Validation of a Quality of Life Instrument for Patients With Diabetes (D39-T) I-Chan Huang, Epidemiology and Health Policy Research, University of Florida, Gainesville, FL, Chyng-Chuang Hwang, Internal Medicine, National Tainan Hospital, Tainan, Taiwan, Taiwan, Albert W. Wu, Health Policy and Management, Johns Hopkins University, Baltimore, MD AIMS: Very few QOL instruments are available for diabetic patients in Asia. The Diabetes-39 (D39), a 39-item QOL instrument for diabetes, has been developed and used in the West, but it is unclear whether this instrument can be applied to Asia. We translated the D39 and tested its psychometric properties using samples in Taiwan. METHODS: This was a cross-sectional study. Data were collected from 280 patients treated in diabetes clinics of a teaching hospital in Taiwan. Forward/back translation and cognitive debriefing were used to develop Taiwan version of D39 (D39-T). Cronbach's alpha coefficient was used to estimate scale reliability. We conducted exploratory factor analysis to determine factor structure. We applied multitrait analysis to evaluate convergent and discriminant validity among subscales of the D39-T with SF-36. We calculated the Cindex (area under ROC curve) to examine known-groups validity of D39-T using laboratory indicators (HbA1C, fasting blood sugar, and high/low-density lipoprotein cholesterol) and presence of diabetic complications (retinopathy, nephropathy, neuropathy, diabetic foot and cardiovascular complications). RESULTS: Factor analysis yielded 5 factors for the D39-T, which is consistent with the Western version. Cronbach's alpha coefficients were acceptable for all subscales (>0.7). Among subscales of the D39-T and SF-36, Pearson's correlations were weak (<0.4;P<0.05) with subscales

1446 /Quality of Life Assessment in Diabetic Patients Attending Outpatient Clinic in a Tertiary Health Care Facility Ehijie F. Enato, Clinical Pharmacy & Pharmacy Practice, University of Benin, Benin, Edo, Nigeria, Ephraim I. Ekweanua, Pharmacy, Federal Medical Centre, Asaba, Delta, Nigeria, Azuka C. Oparah, Clinical Pharmacy & Pharmacy Practice, University of Benin, Benin, Edo, Nigeria, Evbade M. Arigbe-Osula, Ochuwa E. Aghomo, Clinical Pharmacy & Pharmacy Practice, University of Benin, Benin, Edo, Nigeria AIMS: Diabetes mellitus is one of the most prevalent endocrine disorders worldwide. The aim of this study was to evaluate quality of life (QoL) of diabetic patients attending outpatient clinic of a tertiary health care facility in Nigeria. METHODS: The study was carried out on 309 out of 400 consecutive patients presenting for routine

A-22

check at a diabetic clinic in Nigeria. Demographic data were collected, while QoL was assessed using diabetic-specific instrument - Diabetes 39 (D-39) developed and previously validated. Descriptive statistics on sample characteristics was computed. Differences between means were explored using Student’s t test or ANOVA. Chronbach’s coefficient alpha was used to determine reliability of the instrument, while QoL was assessed using a transformed score on a scale of ‘0’ to ‘100’ (highest to lowest QoL). RESULTS: Demographic and clinical profiles indicated: mean age of 60 ± 12.2 yrs, 45% males, a majority were business men/women (36%) and retired (34%), had at least secondary educational level (57%). 98% of them had type I diabetes, and 57% of the patients were previously diagnosed 5 yrs prior to their enrolments into the study. Hypertension (37%) was the most prevalent co-morbid disease. 74% of the patients were on oral hypoglycemic agents. The mean blood sugar level was 7.65 ± 2.74 mmol/L. The chronbach’s alpha of the different subscales ranged from 0.56 to 0.90. The QoL was found to be: Diabetes control, 33; Anxiety and worry, 42; Social burden, 20; Sexual function, 38, and Energy and morbidity, 39. Level of education or duration of illness did not affect the QoL of the respondents (p>0.05). However, females had significantly better QoL on sexual function than males (p<0.05). Also, on energy and mobility subscale, patients less than 60yrs had better QoL than those equal to or greater than 60yrs old(p<0.05). CONCLUSIONS: Though diabetic patients had better than average QoL, there is need to improve QoL of males and senior citizens with the disease in this health care facility.

children report on childrens QOL. Given that these findings suggest that there are systematic differences in response styles for children and parents, it is important that researchers consider this when interpreting scores on QOL. 1533 /Pre-School Children Quality of Life: Comparison of the Infant and Toddler Quality of Life Questionnaire (ITQOL) and the TNO-AZL Pre-School Children Quality of Life Questionnaire (TAPQOL) Hein Raat, Public Health, Erasmus MC - University Medical Center Rotterdam, Rotterdam, ZH, Netherlands, Ashna Mohangoo, Public Health, Erasmus MC, Rotterdam, ZH, Netherlands, Jeanne M. Landgraf, Qual of Life Measures, Health Act, Boston, MA AIMS: We compared the reliability and validity of the originally U.S. ITQOL (103 items, 12 scales) and the originally Dutch TAPQOL (43 items, 12 scales), two parent questionnaires to measure health-related quality of life of pre-school children. METHODS: Parents of 500 children from a random general population sample (2 months-4 years) and of 217 children with diagnosed respiratory disease of 2 outpatient clinics (5 months-5.5 years) were mailed ITQOL and TAPQOL questionnaires; a retest was sent after 2 weeks to a random subgroup. RESULTS: Response was 83.0%, respectively 79.7%. Mean child age was 24.6 (SD 13.8), respectively 34.5 (SD 16.4) months. Both ITQOL and TAPQOL showed few missing items: circa 1%. In the general population sample 4 ITQOL and 11 TAPQOL scales showed a ceiling effect, defined as >25% with the maximum score; somewhat less ceiling occurred in the clinical sample. All multi-item ITQOL scales showed internal consistency (Cronbach's alpha>=0.70), while the TAPQOL, with fewer items per scale, had 3 scales with alpha<0.70. Only 4 ITQOL and 4 TAPQOL scales showed Intra Class Correlation coefficients >=0.70 between test and retest. Six ITQOL scales and two TAPQOL scales showed medium or large effects (defined as Cohen's effect size >=0.50) between children with and without respiratory disease. Correlations between ITQOL and TAPQOL scale scores were relatively low in both samples; only one TAPQOL scale showed r>=0.50 with three ITQOL scales. CONCLUSIONS: The ITQOL and the TAPQOL, next to the PedsQl (not in this study) belong to the few health profile measures for pre-school children; relatively few reports concerning these measures have been published yet. Both ITQOL and TAPQOL showed to be feasible, but they have different content. Both can be applied depending on the research objectives. We recommend developing a shortened ITQOL version, and further evaluations of pre-school measures, especially of test-retest reliability and resposiveness to change.

Quality of Life in Children 1 1600 /Parent-Proxy and Child Self-Reports of Quality of Life: Determining the Cognitive Explanations for Discordance Elise C. Davis, Caroline Nicolas, Elizabeth Waters, Kay Cook, Health and Social Development, Deakin University, Burwood, Victoria, Australia, Angela Gosch, Munich University of Applied Sciences, Munich, Germany, Ulrike Ravens-Sieberer, Child and Adolescent Health, Robert Koch Institute, Berlin, Germany, Robert Cummins, Psychology, Deakin University, Burwood, Victoria, Australia AIMS: Quality of life assessed by parent-proxy reports or child selfreports is an important health outcome measure. Concordance between scores from parents and children tends to be poor, though reasons are unknown. Researchers have proposed that parents and children differ in their understanding and interpretation of the terms or there may be real differences in perspective however this has not been tested. The aim of this study was to identify and compare, using qualitative techniques, cognitive processes that parents and children use when reporting on children s QOL. METHODS: 15 childparent dyads (children aged 8-12 years) were recruited from primary schools within Victoria, Australia. Parents and children were given a new pan-European QOL questionnaire, KIDSCREEN (28 item), to complete in an interview setting. Parents and children were instructed to complete the questionnaire and to verbalise their thoughts whilst they were considering their responses. Parents and children were questioned separately to allow both of them to freely discuss the questionnaire. Interviews were recorded, transcribed, and analysed. RESULTS: Discordance between parent proxy and child self-reports of QOL were attributed to differences in interpretations of some items, but mostly to differences in using the rating scale and different experiences. Compared to parents, children were more likely to provide extreme responses (49.6% vs 23.5%), provide responses followed by explanations (65.7% vs 34.6%) and base their responses on one single example rather than multiple examples (38.3% vs 25.9%). CONCLUSIONS: This is the first study that examines, using qualitative methods, the cognitive processes by which parents and

1660 /Parental Opinions about the Expansion of the Neonatal Screening Programme Symone B. Detmar, Nynke Dijkstra, Marlies Rijnders, Esther J. Hosli, Quality of Life, TNO, Leiden, The Netherlands AIMS: Advances in genomics will open up new opportunities in the fields of genetic testing, early diagnosis and disease treatment. The field of application par excellence for these developments is neonatal screening. The debate about the potential benefits and drawbacks of neonatal screening on the quality of life of child and family is mainly theoretically driven and based on opinions held by professionals. The opinion of the general population is rarely sought. We conducted a qualitative study of the perceptions, preferences and needs of parents (to-be) with respect to various neonatal screening possibilities. METHODS: Seven focus group discussions were conducted with parents-to-be, parents with healthy children and parents with affected children. Four examples of existing or potential neonatal screening possibilities were discussed in depth using disease scenarios (Phenylketonuria, Cystic Fibrosis, Duchenne Muscular Dystrophy

A-23

and Coeliac Disease). All focus group discussions were audio-taped and content-analysed. RESULTS: It was found that the arguments participants used were related to the characteristics of the disease being considered, and that the majority of participants used various arguments for and against screening. Assuming the availability of effective early medical treatment, almost 100 percent of the participants were willing to participate in a screening programme. In the absence of this type of medical treatment, willingness was much lower and the arguments participants used in the absence of effective options for early medical treatment diverged considerably, and were related to the perceived impact on their quality of life. CONCLUSIONS: The divergence in attitudes and preferences reflects the complexity associated with considering new screening possibilities. A better understanding of the thinking of target groups (i.e. parents) is important for developing successful implementation and for directing applied research in genomics.

AIMS: To identify currently available generic and disease-specific HRQoL instruments for children and adolescents, to describe their content and to assess their psychometric properties. METHODS: Databases (Medline, ISI Science Citation Index, PsycLit, ProQolid and previous reviews, i.e. Eiser and Morse, 2001) were searched to find measures of HRQoL, health status or well-being . Generic and disease-specific instruments were analysed separately. Descriptive characteristics and psychometric properties (reliability, validity and sensitivity to change) were analysed following guideline recommendations of the Scientific Advisory Committee of the Medical Outcomes Trust. RESULTS: In total 24 generic and 48 disease-specific instruments were identified, 5 generic and 23 specific instruments were published since the last major revision in 2001. Most of them were developed in the USA (n=26) or UK (n=12). There was evidence that children as young as 6 years old could provide reliable and valid responses. Disease-specific instruments were produced for a total of 25 different diseases. There was substantial variability in the number and characteristics of the dimensions included. Reliability was generally acceptable, and the majority of instruments provided data on internal consistency (n=49), and to a lesser extent test-retest reliability (n=32). Nearly all of them reported construct validity, but only 16 (22.2%) analysed sensitivity to change. CONCLUSIONS: There are quite a large number of instruments available which are designed specifically to measure the HRQoL of children and adolescents, and the number is growing, particularly in terms of disease-specific instruments. Some psychometric characteristics have been reasonably well tested, but others, most notably sensitivity to change, have not been tested in the majority of instruments. The heterogeneity in content may indicate a lack of definition of the HRQoL concept in paediatric populations. Funding: Instituto de Salud Carlos III (Network of excellence IRYSS G03/202), in collaboration with the REpIER network.

1447 /Children with Special Health Care Needs (CSHCN) and Quality Of Life: Results of the European KIDSCREEN Project Angela Gosch, Michael Erhart, Nora Wille, Child and Adolescent Health, Robert Koch Institute, Berlin, Germany, Luis Rajmil, Research Area, Catalan Agency for Health Technology Assessment, Barcelona, Spain, Ulrike Ravens-Sieberer, Child and Adolescent Health, Robert Koch Institute, Berlin, Germany, and the European KIDSCREEN group AIMS: The objective of this European cross-sectional study using representative samples of children, adolescents and their parents in 11 European countries is to present data of children with special health care needs (SHCN), to examine differences by age, gender and the familial affluence, and to explore the relationship between the child's health status and health-related quality of life (HRQoL). METHODS: In 11 European countries (A, CH, CZ, D, E, F, GR, HU, NL, UK, PL) 16,730 parents were asked to fill in the translated and harmonised CSHCN Screener, a five item non-categorical, consequences-based and non-diagnostic based screening instrument to identify children with SHCN. The children and adolescents (8-18 years) answered the KIDSCREEN-27 questionnaire, a generic HRQoL instrument with 5 dimensions and questions assessing the familial affluence (FAS). RESULTS: Overall 10.7% of the children and adolescents experience one or more of five health or health need consequences due to an ongoing health condition. Percentages vary widely over the 11 European countries, with the highest in The Netherlands (16.1%) and the lowest in Greece (3.4%). Differences between genders (more males) are found as well as an impact of the familial affluence on the number of children with SHCN(lower FAS is associated with a higher rate of CSHCN). Children with SHCN show lower scores on the five KIDSCREEN-27 dimensions in comparison to non-CSHCN: they report a lower physical (d=.41) and psychological wellbeing (d=.29), reduced relationships with parents (d=.22) and peers (d=.27), and a lower wellbeing in school (d=.25). CONCLUSIONS: This is the first European study that examines, using the CSHCN screener and the KIDSCREEN-27, the rate of SHCN in children and adolescents aged between 8-18 years and the relationship to HRQoL. The study identifies differences between country specific overall rates of CSHCN across Europe. This result and the impact of determinants on SHCN will be discussed in detail. Implications for policy and programme planning will be addressed.

Understanding Minimally Important Differences for QOL Research 1396 /Linking Minimal Important Change (MIC) to Measurement Error to Determine the Usefulness of PatientReported Outcomes (PROs) for Daily Clinical Practice and to Estimate Sample Sizes for Research: An Example with WOMAC Data Caroline B. Terwee, EMGO Institute, VU University Medical Center, Amsterdam, the Netherlands, Leo D. Roorda, Rehabilitation and Psychology, Jan van Breemen Institute, Amsterdam, the Netherlands, Dirk L. Knol, Clinical Epidemiology and Biostatistics, VU University Medical Center, Amsterdam, the Netherlands, Michiel R. de Boer, Nutrition and Health, Faculty of Earth and Life Sciences, Free University, Amsterdam, the Netherlands, Lex M. Bouter, Henrica C. de Vet, EMGO Institute, VU University Medical Center, Amsterdam, the Netherlands AIMS: Although increasing attention is being paid to interpretation of scores and determination of MIC, the link with measurement error is seldomly made. We provide a practical example of how to relate MIC to measurement error to determine the usefulness of PROs for daily clinical practice and to calculate sample sizes required for detecting important changes in groups of patients. METHODS: As an example we used WOMAC data (range 0-100) from a cohort study (n=200) in patients undergoing hip or knee replacement. We calculated the standard error of the measurement (SEM) by taking the square root of the sum of the between measures variance and the residual variance from an ANOVA analysis in a subgroup of 35 stable patients who were assessed twice while on the waiting list for total hip of knee replacement. As a measure of MIC, we used the mean change in score in the subgroup of patients that reported to be

1611 /Health-Related Quality of Life (HRQoL) Measurement in Children and Adolescents: A Review of Instruments M Solans, S Pane, MD Estrada, V Serra-Sutton, S Berra, Catalan Agency Health Technology Assessment Research, Barcelona, Spain, M Herdman, J Alonso, Institut Municipal d'Investigació Mèdica (IMIM-IMAS), Barcelona, Spain, L Rajmil, Catalan Agency Health Technology Assessment Research, Barcelona, Spain

A-24

"better" on a global rating scale 6 months after joint replacement, as well as MIC values from the literature. Based on SEM and MIC we determined the number of patients required to detect the MIC within a group of patients and between two treatment groups. RESULTS: The SEM was 6.1 points for WOMAC physical functioning and 9.8 points for WOMAC pain. Given these SEMs, one can only distinguish changes of at least 24 points in physical functioning and 39 points in pain from measurement error in individual patients. MIC values were 27 points for physical functioning (n=41) and 30 points for pain (n=30) in this cohort of total joint replacement patients. In studies with patients undergoing rehabilitation, much smaller MIC values of 7-9 points are reported. CONCLUSIONS: The WOMAC can only be used in individual patients to detect large changes, such as those after total joint replacement. The much smaller effects of rehabilitation cannot be detected with the WOMAC in individual patients, but only in groups of patients.

feasibility for assessment of health status in patients with advanced NSCLC. METHODS: EQ-5D index and Visual Analog Scale (VAS), and Functional Assessment of Cancer Therapy-Lung (FACTL) were administered in phase II trial of combination chemotherapy (at baseline, start of each 3-week cycle, and end of patient's last cycle). Clinical assessments included radiological tumor response and Eastern Cooperative Oncology Group performance status (PS). Reliability, convergent and construct validity, and responsiveness were assessed using standard statistical tests. Anchor (PS and best tumor response) and distribution (0.5 standard deviation and standard error of measurement) methods were used to estimate MIDs. RESULTS: Data from 195 patients were analyzed. Cronbach's alpha for index was 0.68 at baseline. Test-retest Pearson correlation was 0.70 and 0.80 for index and VAS, respectively for 7-day interval. Convergent validity correlation between index and FACT-L subscale scores ranged from 0.15 to 0.57; the highest correlations were with physical and functional well-being. Index scores differentiated patients by PS and best tumor response (p<0.05) in cross-sectional analyses, demonstrating construct validity. EQ-5D was more responsive to PS than to best tumor response (see Table 1 for index results). MIDs were estimated as 0.12-0.15 for index and 8-10 for VAS. CONCLUSIONS: The EQ-5D demonstrated desirable psychometric properties. Our findings suggest that this brief questionnaire is useful for assessment of health status and physical/functional well-being in advanced NSCLC.

1420 /Minimally Important Change Determined by a Visual Method Integrating an Anchor-Based and a Distribution-Based Approach Henrica C. De Vet, Raymond W. Ostelo, Caroline B. Terwee, Nicole Van der Roer, EMGO-Institute, Dirk L. Knol, Clinical Epidemiology and Biostatistics, Heleen Beckerman, Rehabilitation and Medicine, Maarten Boers, Clinical Epidemiology and Biostatistics, Lex M. Bouter, EMGO-Institute, VU University Medical Center, Amsterdam, Netherlands

Stable Worsened Tumor Stable Progres PS PS response disease -sion

AIMS: Minimally important changes (MIC) in scores help interpret results from health status instruments. Various distribution-based and anchor-based approaches have been proposed to assess MIC. The aim of this study is to describe and apply a visual method, called the anchor-based MIC distribution method, which integrates both approaches. METHODS: Using an anchor, patients are classified into a group of persons with an important improvement, an important deterioration, or a group without important change. For these groups the distribution of the change scores on the measurement instrument are depicted in a graph. We present two cut-off points for an MIC: the ROC cut-off point and the 95% limit cut-off point. RESULTS: We illustrate our anchor-based MIC distribution method determining the MIC for the Pain Intensity Numerical Rating Scale in patients with low back pain, using two conceivable definitions of minimal important change on the anchor. The graph shows the correspondence between the anchor and the health status instrument, and the consequences of choosing the ROC cut-off point or the 95% limit cut-off point. CONCLUSIONS: The anchor-based MIC distribution method provides a general framework, applicable to all kind of anchors. This method forces researchers to choose and justify their choice of an appropriate anchor and to define minimal importance on that anchor. The MIC is not an invariable characteristic of a measurement instrument, but may depend, among other things, on the perspective from which minimal importance is considered and the baseline values on the measurement instrument under study. A balance must be struck between the practicality of a single MIC value and the validity of a range of MIC values.

Effect -0.08 -0.73 Size St. Response -0.08 -0.38 Mean

-0.27

-0.36

-0.44

-0.38

-0.34

-0.28

1480 /Computerized Assessment of QOL in the Oncological Routine: Which Criteria Should We Use for Monitoring Changes? Georg Kemmler, Psychiatry, Medical University of Innsbruck, Innsbruck, Austria, August Zabernigg, Klaus Gattringer, Internal Medicine, Kufstein County Hospital, Kufstein, Austria, Barbara Weber, Computer Science, University of Innsbruck, Innsbruck, Austria, Barbara Sperner-Unterweger, Bernhard Holzner, Psychiatry, Medical University of Innsbruck, Innsbruck, Austria AIMS: The Computer-based Health Evaluation System (CHES) is a PC software for the collection, processing and presentation of medical and psychosocial data, including quality of life (QOL). The programme has a feedback module providing the clinician with longitudinal QOL profiles for the individual patient. To further improve this module we plan to add a graphical display of relevant changes of a patient's QOL. Aim of this contribution is an empirical investigation of possible criteria for such relevant changes. METHODS: The investigation was based on a longitudinal QOL study of cancer patients receiving chemotherapy at the oncological out-patient unit of Kufstein County Hospital. QOL was assessed using the EORTC QLQ-C30. Several potential criteria for defining relevant changes (absolute differences, multiples of the standard deviation (SD)) were compared in terms of the relative frequency and statistical significance of the resulting "relevant" changes. RESULTS: QOL data of 88 cancer patients were used. The mean number of follow-up times per patient was 4.4, resulting in a total number of 390 change scores available for analysis. When applying criteria of relevant change suggested by the literature (e.g., 0.5*SD), high proportions of reputedly "relevant" changes were found (e.g.,

1005 /Validity, Responsiveness, and Minimally Important Difference (MID) Associated with Euroqol EQ-5D in Advanced Non-Small Cell Lung Cancer (NSCLC) Lucie Kutikova, Global Health Outcomes, Eli Lilly and Company, Indianapolis, IN, Murali Sundaram, Pharmaceutical Systems and Policy, West Virginia University School of Pharmacy, Morgantown, WV, Nathan H. Enas, Huan Lu, Global Statistics, Astra M. Liepa, Global Health Outcomes, Eli Lilly and Company, Indianapolis, IN AIMS: Given a lack of validity data for the EuroQol EQ-5D in cancer patients, we investigated its psychometric properties and its

A-25

using the 0.5*SD criterion, 41 % of all changes in the functioning scales were deemed "relevant"). The majority of these changes could not be distinguished from random fluctuations. Increasing the thresholds (e.g. to 1 SD) resulted in practically and statistically more useful criteria, but determination of optimal cut-offs remains problematic. CONCLUSIONS: Recommended thresholds of relevant QOL changes (usually based on group data) appear to be unduly low when considering changes in the individual patient. This should be taken into account when monitoring QOL changes. Further research is needed to set up criteria of relevant change in the individual that are both clinically and statistically meaningful.

Cancer Studies and Approaches in QOL 1 1279 /Added Cancer QOL Measures: Adding Value? Ding Chen, Chris J. O'Callaghan, NCIC Clinical Trials Group, Queen's University, Kingston, Ontario, Canada, Michael D. Brundage, Division of Cancer Care and Epidemiology, Queen's University Cancer Research Institute, Kingston, Ontario, Canada, Timothy J. Whelan, Radiation Oncology, Juravinski Cancer Centre at Hamilton Health Sciences, Hamilton, Ontario, Canada AIMS: The National Cancer Institute of Canada Clinical Trails Group (NCIC CTG) has compiled a database of 'QoL checklist items'. These are single-item questions developed for a specific trial to assess disease- or treatment-related aspects of QoL (beyond the core or disease-specific QoL questionnaires). We evaluated the added value of the QoL checklist items (in the context of a specific trial) by determining their sensitivity to change over time and treatment induced differences, and, to assess their reliability and validity. METHODS: This study was a secondary analysis of the QoL data from an NCIC CTG trial of tamoxifen versus tamoxifen plus octreotide LAR in post-menopausal women with primary breast cancer post definitive surgical management. QoL data from the first four months of treatment were analyzed. RESULTS: Compliance was high: 96% (635 of 659 eligible patients) at baseline and 89% (511 of 573 expected) at month 4. Some checklist items discriminated between patients differing in performance status (e.g. ECOG PS was significantly associated with the 'physical intimacy' item at all time points, Kruskal-Wallis chi-square=7.57, 10.28 at baseline, month 4 respectively; both p<0.02). Items related to treatment side effects discriminated between patients differing in relevant physician-rated toxicities (e.g. CTC toxicity scores associated with 'vasomotor symptoms', Z=-10.05 and -11.22 at month 1 and 4 respectively with both p<0.0001, Wilcoxon rank-sum). Checklist items selected on clinical grounds changed in the hypothesized direction over time (e.g. 'weight gain' worsened significantly on both arms) and were responsive to treatment-induced differences (e.g. 'abdominal bloating' scores were significantly higher on octreotide, p<0.001 at month 4). Some checklist items correlated significantly with the patient's global QoL. Overall, checklist scores provided data not addressed by the EORTC core questionnaire and QLQ-BR23 module. CONCLUSIONS: In conclusion, the selected checklist patientreported outcomes were clinically valid and added value to the standard measures in the context of this clinical trial.

1540 /Estimating the Minimal Clinical Important Difference of the Ocular Surface Disease Index®: Preliminary Findings Kimberly L. Miller, Outcomes Research, David R. Mink, Statistics, Susan D. Mathias, Outcomes Research, Ovation Research Group, San Francisco, CA, John G. Walt, Global Health Outcomes Strategy and Research, Allergan, Inc, Irvine, CA AIMS: To assess the minimal clinical important difference (MCID) of the Ocular Surface Disease Index® (OSDI®). METHODS: The OSDI is a 12-item patient-reported outcomes (PRO) questionnaire designed to quantify ocular disability due to dry eye disease. Overall OSDI and subscale scores range from 0 to 100; with scores categorized by severity, i.e. normal (0-12), mild (13-22), moderate (23-32) and severe (33-100). To estimate the MCID, we used data from the RESTORE Registry study, which collects clinical, efficacy, PRO data, including the OSDI, and safety data from patients with dry eye disease. We used an anchor-based method selecting a clinician's global impression (CGI) and subject global assessment (SGA) as anchors. Patients who completed the OSDI at baseline and at a follow-up visit and completed the SGA or their clinician completed the CGI were included in our analyses. We examined change scores for the OSDI; assessed the association between each anchor and OSDI scores; and computed the MCID using a one-way ANOVA. RESULTS: 77 patients were included in the analysis. Most were Caucasian (87%), female (87%), with mean age of 61.3 years. Using either anchor, most patients reported an improvement in OSDI scores at follow-up. Both the SGA and the CGI correlated well with OSDI scores (r = 0.278-0.509, p<0.05). MCID results were fairly consistent across scales and indicated that an 8-point change and 9-point change (on a 0 to 100 scale) was clinically meaningful for the CGI and the SGA, respectively. CONCLUSIONS: Although these results are preliminary, ongoing data collection from the RESTORE Registry will allow for future analyses of larger samples sizes to finalize the MCID estimates.

Anchor N CGI

77

SGA

66

Baseline Overall OSDI 38.9 +/23.29 39.0 +/23.92

1799 /The Relationship Between Cancer Patient Treatment Decision-Making Roles and Quality Of Life (QOL) Pamela J. Atherton, Health Sciences Research, Mayo Clinic, Rochester, MN, Tenbroeck Smith, American Cancer Society, Rochester, MN, Jeff Huntington, University of Utah, Rochester, MN, Mashele Huschka, Jeff A. Sloan, Health Sciences Research, Mayo Clinic, Rochester, MN

ANOVA Overall OSDI BetaChange Coefficient -10.3 +/- 19.38 -7.6

AIMS: It is unknown whether failing to meet patient expectations for participating in treatment decisions translates into QOL deficits. As part of a larger survey by the American Cancer Society, patients reported the role preferred and the role actually experienced during the treatment decision-making process. METHODS: Patients who were diagnosed with one of the ten most common cancers in 2000 completed the Profile of Mood States (POMS), the SF-36, and the Control Preferences Scale (CPS). Fisher's exact tests compared role preference distributions across demographic categories. Two-sample t-tests compared the POMS and SF-36 scores between patients whose preferred role preference was concordant with the role experienced and those with discordant preferred and actual roles. RESULTS: 33% of the 599 consenting patients preferred an active role in treatment decision-making, 52% preferred a collaborative role, and 13%

- 9.6 +/- 19.91 -8.9

A-26

preferred a passive role. The actual roles experienced were 33% active, 50% collaborative, 17% passive. Over 88% of patients had concordant preferred and actual roles. Patients with concordant roles had higher SF-36 physical scores (45 vs 40, p=0.004), higher vitality (50 vs 42, p=0.005), less fatigue (70.2 vs 60.1, p=0.001), better concentration (84 vs 79, p=0.008) and better overall mood (77 vs 73, p=0.006). Role preference differed across gender (p=0.0002) in that more women preferred a collaborative role than men (57.8% vs 45.5%) and fewer women preferred a passive role (9% vs 17.3%). Patients under age 50 experienced more active roles in treatment decisions than those aged 50+ (p=0.04). Patients reporting an active role had higher SF-36 physical scores (p=0.005) and higher POMS vigor subscale scores (p=0.04). CONCLUSIONS: Patients who experienced discordance between their preferred role and their experience reported substantial QOL deficits in both physical and emotional domains. Satisfaction with care and QOL can be improved by meeting patient expectations with respect to the amount of input they have in making treatment decisions.

Innsbruck, Tyrol, Austria, Barbara Sperner-Unterweger, Georg Kemmler, Psychiatry, Medical University of Innsbruck, Innsbruck, Tyrol, Austria AIMS: The Computer-Based Health Evaluation System (CHES) was routinely used for the asssessment of Quality of Life (QoL) of oncological outpatients receiving chemotherapy over a period of more than one year. Aim of this project was to study one rarely investigated clinically relevant side-effect, namely taste alterations including their impact on QoL. METHODS: Cancer patients, who were receiving chemotherapy at the out-patient unit of Kufstein County Hospital, were consecutively included. Inclusion criteria were German speaking and expected survival time of more than six months. At each visit the patients were asked to fill in the EORTC QLQ-C30 and two additional questions concerning taste alterations (with the help of CHES). RESULTS: The sample consisted of a total of 103 cancer patients (main diagnoses: 25.20% breast cancer 24.30% lung cancer, gender 53.4% female, age 58.6+/-18.4). QoL was assessed 4.8times on average per patient, 499times in total. The routinely computerized QoL with the help of the computer software CHES was well accepted by the patients and was found to be helpful for optimizing treatment of the individual patient (e.g. symptom management). 76.3% of the patients reported taste alterations in the course of chemotherapy (41.3%mild; 20.0%moderate and 15.0%severe).Taste alterations were significantly associated with a reduction in QoL and an increase of symptoms such as fatigue, sleep disturbances and pain. Analyses investigating the impact of the type of chemotherapy on taste alterations are currently performed and will be presented. CONCLUSIONS: Taste alterations are very common in patients undergoing chemotherapy.Therefore this kind of treatment side effect should be paid more attention in clinical trials as well as in the daily oncological practice. The use of a computerized assessment and evaluation of quality of life including clinically relevant sideeffects such as taste alterations has proven to be feasible and beneficial and has the potential of changing individual medical care.

1803 /Comparison of Baseline Quality of Life Between Minority and Non-Minority Patients Participating in Oncology Clinical Trials Angelina D. Tan, Paul J. Novotny, Health Sciences Research, Judith S. Kaur, Oncology, Jeff A. Sloan, Health Sciences Research, Mayo Clinic, Rochester, MN AIMS: Minority patients (MP) suffer deficits in access to care and socioeconomic status. This study uses a patient-level pooled analysis to explore whether these deficits translate into quality of life (QOL) differences between MP and non-MP on clinical trials. METHODS: Baseline QOL scores were combined from 47 clinical trials (6513 patients) conducted either at the Mayo Clinic Cancer Center or in the North Central Cancer Treatment Group. QOL scales used were the Uniscale, Linear Analogue Self Assessment (LASA), Symptom Distress Scale (SDS), Profile of Mood States (POMS) and Functional Assessment of Cancer Therapy - General (Fact-G). Fisher s Exact tests and linear regression adjusted for age, site, and performance score. Survival data was compared using the method of KaplanMeier. RESULTS: Eight percent (531) of patients self-reported as MP (0.45% American Indian/Alaskan Native, 0.7% Asian, 5% Black/African American, 1.5% Hispanic, 0.1% Native Hawaiian and 0.3% Other). MP had no meaningful deficits in overall QOL or the SDS and were slightly worse on FACT-G total score, physical, social/family, functional, and SDS nausea severity. MP among lung cancer patients reported greater nausea (58 vs 69) and sleep problems (34 vs 54), with neurological cancers reported worse emotional wellbeing (53 vs 74), and with GI tumors had lower social/family wellbeing (60 vs 67). Regression models confirmed these results. Median survival time was shorter for MP (198 vs 310 days, p=0.001) but was not significant after adjusting for disease severity. CONCLUSIONS: MP on these clinical trials did not report large QOL deficits at baseline relative to non-MP. MP did indicate small deficits in physical, social, and emotional subscales. MP experienced large tumor-specific deficits for a few QOL domains that might bear further attention.

1763 /Discovering Causal Structure in Dynamically Evolving Multidimensional QOL Carol M. Moinpour, SWOG/Cancer Prevention, Nigel E. Bush, Cancer Prevention, Vivian A. Markle, Transplant Clinic/Clinical Trials Support Office, Fred Hutchinson Cancer Research Center, Seattle, WA, Gary W. Donaldson, Anesthesiology, Pain Research Ctr/Cancer Prevention, University of Utah, FHCRC, Salt Lake City, UT AIMS: Why does Quality of Life (QOL) change during chronic health challenges? Controlled treatment trials address this question, but average effect sizes are typically small, with substantial individual variation. We propose observing multiple dimensions of QOL frequently and repeatedly to suggest their possible underlying patient-specific causal structure and establish targets for intervention. METHODS: Patients who received hematopoietic stem cell transplants at FHCRC responded frequently via Internet to a rotating battery of questions assessing nine dimensions of QOL. We analyze monthly data from 178 who met an average response rate of twice per week. Approximately 25% of these patients provided longitudinal data for a full year, with roughly uniform monthly durations for the remaining patients (to minimize attrition bias, we included all patients with at least one full month of reporting data). The analyses invoked the criterion of Granger causality, which identifies a possible cause with a lagged variable that predicts an outcome controlling for all other lagged variables in the information set. Mixed effects models and vector autoregression imposed a uniform qualitative structure while allowing patients to vary in their levels and pairwise regressions. RESULTS: We identified a specific pattern of Granger causes among the nine variables. The pattern is concisely

1388 /Results of The Assessment of QOL in the Oncological Routine Using a PC Software (CHES): Taste Alterations in Cancer Patients Receiving Chemotherapy Bernhard Holzner, Psychiatry, Medical University of Innsbruck, Innsbruck, Tyrol, Austria, August Zabernigg, Internal Medicine, County Hospital Kufstein, Kufstein, Tyrol, Austria, Gerhard Rumpold, Medical Psychology and Psychotherapy, Medical University of Innsbruck, Innsbruck, Tyrol, Austria, Klaus Gattringer, Internal Medicine, County Hospital Kufstein, Kufstein, Tyrol, Austria, Barbara Weber, Informatics, University of Innsbruck,

A-27

summarized as a cyclic dyadic graphical model. To illustrate, symptomatology was Granger-caused by Fatigue and Ability to Concentrate, suggesting that pacing and memory or concentration aids might be important modulators of developing symptoms (e.g., remembering medication schedules). In general, these relationships were small in practical terms, but random effects about the population means indicated much larger relationships for some patients. CONCLUSIONS: This methodology provides a new way to understand processes affecting QOL, and suggests intervention targets for the discovered mechanisms. For example, a modest intervention to augment memory might pay surprising benefits in controlling symptoms.

AIMS: Group-based chronic disease self-management programs (CDSMPs) have been found to induce response shift (RS) in participants. This may have consequences for the validity of classic pre-post assessment as participants may not only change their perspective over the course of an intervention but they may also develop a new understanding of questionnaire items. This study compares results from data derived from classic pre-post and results from retrospective assessments (post-retro). METHODS: 682 participants of CDSMPs completed the 8-dimension Health Education Impact Questionnaire (heiQ), a generic measure of CDSMPs outcomes. Change was assessed using a) post minus baseline and b) post minus baseline reported retrospectively. Additionally the heiQ-Perspective, a questionnaire measuring RS, was collected at post-test. People were grouped into -, 0 or + RS (i.e. they now realize they were worse, the same or better before the intervention than they thought they were). The course impact was assessed using a) mean change scores and b) proportions of people reporting MCID 0.5 units on the 1 to 6 unit scale. RESULTS: Across the two ways of assessing change, mean heiQ differences greatly differed such that marked differences in conclusions of CDSMP effectiveness would result. The overall mean improvement of participants across the 8 dimensions using the classic pre-post method was 0.37 compared with 0.48 using the post-retro method. However the proportion of people who scored >0.5 was similar (~37%) across methods. Across RS categories for the classic pre-post assessment the proportion scoring >0.5 for -, 0, + RS was 48%, 30% and 36% which was similar for the post-retro method. CONCLUSIONS: RS is frequent in CDSMPs. The results suggest that RS is strongly associated with estimation of CDSMP effectiveness and has a profound effect on pre-post assessment. Interestingly, the 0.5 MCID cut-off method is minimally affected by RS. A one-off post course assessment of CDSMPs seems to be an appropriate and cost-effective assessment method.

Methodology Developments in QOL 1 1781 /The Pros and Cons of PROs Paul Kind, Centre for Health Economics, University of York, York, England AIMS: This paper discusses the (dis)advantages of the PRO and the implications for quality of life (QoL) measurement and research. METHODS: The development of QoL measures was largely in response to the recognition that conventional clinical measures were narrowly focussed and failed to take account of major elements that were of value to patients. Generic measures such as SF-36, HUI and EQ-5D represent a line of research that can be traced back to the earliest work of Bush and colleagues in the early 1970s. A critical element in most such measures is the value associated with the items defined by its descriptive classification system. But we need to know more about the effect of treatment than its impact on patient's blood pressure or lung function. We need to know how the overall health status of the patient changes in response to treatment and how that change is valued. It is the incorporation of value that separates QoL measurement from other forms of observation. RESULTS: However, defining QoL measurement became a major problem for the US Food and Drug Administration in its deliberations on patient information leaflets and the concept of Patient Reported Outcomes (PROs) was proposed as a way of coping with this intellectual logjam. A consequence of this has been to place QoL measures in the same category as a group of unrelated measures, such as compliance and patient satisfaction. This taxonomy has been seized upon like water in the desert. The pharmaceutical industry and FDA now speak a shared language in which QoL measurement has been assigned a status equivalent to the measures it was designed to replace. A new marketing opportunity has opened up that promotes the new concept of "PROs". Guidance published by the FDA threatens to compromise decades of scientific and intellectual investment in developing the distinctive QoL measurement paradigm. CONCLUSIONS: This paper weighs the advantages and disadvantages of accepting the PRO definition of QoL measurement and explores the consequences for the international research community of allowing industry and government to dictate the direction of scientific endeavour.

1407 /A Comparison of Imputation Methods for Trial Data Collected by Reminder Shona A. Fielding, Peter M. Fayers, Public Health, Craig R. Ramsay, Marion K. Campbell, Health Services Research Unit, The University of Aberdeen, Aberdeen, Scotland AIMS: Several methods exist to deal with missing data but it is unclear which are most appropriate. The aim is to compare imputation results with actual quality of life (QoL) data collected after a reminder in studies from two different disease areas. METHODS: EuroQoL EQ5D scores obtained following a reminder were treated as missing. Reasons for non-response were investigated using logistic regression. Missing values were imputed using several methods (including last value carried forward (LVCF), next value carried backwards (NVCB), regression methods and hotdecking). The accuracy of these were investigated by calculating the bias and proportionate variance (PV). RESULTS: In a knee surgery trial, reasons for missingness were related to patient characteristics (e.g. age) and to previous or current QoL implying data were missing not at random (MNAR). At one year follow-up, the next value carried backwards (NVCB) method had smallest bias (-0.007, 95%CI (0.060,0.045)) with the regression based on age, type of knee arthritis and previous EQ5D score second smallest bias (-0.045, 95%CI (0.070,-0.019)). The PV of NVCB was 0.74, 95% CI (0.45,1.11). The PV for the regression method was much lower (PV=0.26, 95% CI (0.21,0.33)). These results were reproduced with a second trial of vitamin supplementation, with evidence of MNAR data and NVCB was reasonably accurate (both bias and PV). The hotdecking procedure showed the smallest PV. CONCLUSIONS: Logistic regression allowed the mode of missingness to be identified as potentially MNAR in the two trials. This conclusion was reached using reminder information and we strongly recommend that all trials

1191 /Is a One-Off Post Intervention Assessment a Valid Method to Assess Outcomes from Chronic Disease Self-Management Programs? Sandra Nolte, CIRCLE, RMIT University, Melbourne, VIC, Australia, Melanie Hawkins, Medicine, University of Melbourne, Parkville, VIC, Australia, Gerald R. Elsworth, CIRCLE, RMIT University, Melbourne, VIC, Australia, Andrew J. Sinclair, Exercise and Nutrition Sciences, Deakin University, Burwood, VIC, Australia, Richard H. Osborne, Medicine, University of Melbourne, Parkville, VIC, Australia

A-28

should collect and analyse such data. In two different trial settings, NVCB was preferable. A disadvantage is that NVCB cannot be used when no later observations occur. In this case LVCF may be appropriate. A disadvantage of regression methods is that given equal combinations of covariates, some imputed values are the same, causing an artificially reduced variance. However, caution must be exercised when imputing around the end of the treatment period.

reviews. METHODS: We conducted a systematic review of systematic reviews of HSMIs. An extensive literature search was performed in PubMed, EMBASE and PsycInfo. Articles were included if they met the following inclusion criteria: (1) systematic review, (2) HSMIs were reviewed, and (3) the purpose of the review is to identify all measurement instruments assessing (an aspect of) health status and to report on the measurement properties of these HSMIs. Two independent reviewers selected the articles and extracted the data. A checklist for assessing the methodological quality of the systematic reviews was developed and applied. RESULTS: PRELIMINARY: We found 100 systematic reviews that satisfied our inclusion criteria. In 84 reviews the search strategy was sufficiently described, 18 stated that the article selection was performed by at least two reviewers and 12 stated that the data extraction was done by at least two reviewers. Twenty-four reviews used only one database (mostly Medline/Pubmed), 17 used two databases, 52 three or more (up to 24) and 2 only stated 'literature search'. Relatively few reviews assessed the methodological quality of the individual studies. The application of criteria varied from ad hoc to very thoroughly. CONCLUSIONS: Systematic reviews of HSMIs published until now differ widely in their methodological quality. Guidelines for conducting and reporting such a systematic review should be developed.

1630 /Does Health Utility Change for Patients with Knee Injuries: It Depends on what Question you Ask? Michael R. Robling, Kerenza Hood, Christopher C. Butler, General Practice, Cardiff University, Cardiff, Wales, UK AIMS: Patients with mechanical knee problems may experience decreased health utility whilst awaiting arthroscopy but assessment may be biased by changes in patient's internal response scale response shift. This study aimed to detect the presence and impact of response shift in patients either waiting for diagnosis or following intervention for a knee injury. METHODS: Health utility (EQ-5D) was assessed over two interviews in two groups of patients with suspected or confirmed internal derangement of the knee. Group 1 included patients recently placed on a waiting list for MRI or an orthopaedic appointment. Group 2 included patients who at baseline were about to undergo arthroscopy. A retrospective pretest-posttest study design was used. EQ-5D was assessed at baseline (pretest) and at six months follow-up (posttest). At follow-up, EQ-5D was also assessed retrospectively - patients provided a renewed judgement of baseline state (thentest). Prospectively and retrospectively derived change scores were calculated. The difference between prospective and retrospective change was determined (response shift) . RESULTS: 39 patients were interviewed at baseline, 36 at follow-up. The overall mean improvement in EQ-5D was 0.16 (sd=0.29; t=3.15; p=0.004) using prospective change score. Change was only evident in the arthroscopy group and not the waiting list group (mean change of 0.26 and 0.05 respectively). When change was assessed using the retrospective score, the overall mean change of 0.12 was of borderline significance (sd 0.12; t=2.07, p=0.046). Change in EQ-5D for the arthroscopy group (mean=0.16; sd=32; t=1.59) was now not significant. Mean change score for the waiting list group had increased using the retrospective approach (mean=0.12), but was still not significant. CONCLUSIONS: Response shift was apparent for arthroscopy patients: re-appraising baseline state to provide a more positive retrospective rating than originally provided. In this study the choice of prospective or retrospective assessment critically determines the significance of observed change.

Mental Health and its Impact on QOL 1652 /Improving the Estimation of Unmet Need for Mental Health Care in the General Population: The European Study of Epidemiology of Mental Disorders (ESEMED) Project Jordi Alonso, Miquel Codony, Health Services Research, Institut Municipal d’Investigació Mèdica (IMIM-IMAS), barcelona, Spain, Viviene Kovess-Masféty, Foundation d'enterprises MGEN pour la Santé Publique, Paris, France, Matthias C. Angermeyer, Psychiatry, Leipzig Universitä, Leipzig, Germany, Josep Maria Haro, Research, Sant Joan de Déu, Serveis de Salut Mental, Sant Boi de Llobregat, Spain, Giovanni de Girolamo, Mental Health, 2nd University Psychiatric Clinic, Bologna, Italy, Gemma Vilagut, Josué Almansa, Health Services Research, Institut Municipal d Investigació Mèdica (IMIM-IMAS), barcelona, Spain, Ron de Graaf, Netherlands Institute of Mental Health and Addiction, Utrecht, The Netherlands, Koen Demyttenaere, Neurosciences and Psychiatry, University Hospital Gasthuisberg, Leuven, Belgium, Jean P. Lépine, Psychiatry, Hôpital Fernand Widal, Paris, France, Traolach S. Brugha, Psychiatry, Academic Unit, University of Leicester, Leicester, United Kingdom AIMS: High prevalences of mental disorders have fuelled a controversy about need for mental health services. Mere presence of a mental disorder may not be an adequate indicator of need for care. We aimed to estimate population level of unmet need for mental health care in Europe. METHODS: 2-phase, cross-sectional survey of representative samples of adult general population in 6 European countries: Belgium, France, Germany, Italy, Netherlands and Spain. 21,425 individuals participated (response= 61.2%), and 8,796 responded the 2nd phase. Mental disorders were assessed by certified lay interviewers using a computerised version of the Composite International Diagnostic Interview (CIDI 3.0), cross-culturally adapted in participating countries. Individuals with 12-month mental disorder that was disabling (interfered "a lot" with their life or activities in their lifetime) or had caused use of services, were considered in need for care. SF-12 and the WHO-DAS were administered. RESULTS: 6.4% of sample had need for mental health care. Individuals with need for mental health care had lower MCS mean scores than those with a 12-month disorder but without care need and those without a mental disorder (41.2, 45.6, and 51.5, respectively; p< 0.01). Differences in the same direction were found

1397 /Assessment of the Methodological Quality of Systematic Reviews of Health Status Measurement Instruments (HSMI) Lidwine B. Mokkink, Caroline B. Terwee, EMGO-Institute, VU University Medical Center, Amsterdam, Netherlands, Paul W. Stratford, School of Rehabilitation Science, McMaster University, Hamilton, Ontario, Canada, Jordi Alonso, Health Services Research Unit, Institut Municipal d'Investigació Mèdica (IMIM-IMAS), Barcelona, Spain, Donald L. Patrick, Department of Health Services, University of Washington, Seattle, Washington, Ingrid Riphagen, Lex M. Bouter, EMGO-Institute, VU University Medical Center, Amsterdam, Netherlands AIMS: A tool for selecting the best health status measurement instrument (HSMI) for research is a systematic review of HSMIs. In such a systematic review all existing measurement instruments on a well-defined concept are described and evaluated. The aim of this review is (1) to evaluate the methodological quality of systematic reviews of HSMIs, (2) to appraise how the methodological quality of individual studies is assessed in these reviews, and (3) to appraise how the results of the individual studies are evaluated in these

A-29

for the WHO-DAS WLD index (21.4, 15.9, and 6.1, respectively; p< 0.01). Mental health need for care was closely related to a complex definition of severity (correlation 0.504, asymptotic standard error=0.024, AUC=0.70). 48.2% with need for mental health care reported no formal use of services. So, 3.1% of adult population had unmet need for care CONCLUSIONS: Interference with daily life is a feasible and valid component indicator of population need for mental health care. Unmet need for mental care in Europe is high. Funding: EC (QLG5-1999-01042, SANCO 2004-0123); Piedmont Region, Italy; SIC III (FIS 00/0028), MCT (SAF 2000-158-CE), Generalitat de Catalunya, Spain; unrestricted educational grant from GlaxoSmithKline

AIMS: The purpose of this study was to examine how chronic mental health problems (MHP) affect life satisfaction. METHODS: This study had a two-group cross-sectional comparative design. Hundred and seven adults with mental health problems from the community health care system were recruited. In addition, data from a general population study including 1893 people was used as a comparison group.The Quality of Life Scale (Norwegian version) was used to measure life satisfaction (satisfaction with broader life domains) in both groups, and the Symptom Checklist-90 Revised was used to assess mental symptoms in the MHP group. General linear model univariate analysis of variance (UNIANOVA) was used to identify possible differences between the MHP group and the general population on life satisfaction controlling for age, sex, education, marital status, having children and cohabitation. Interaction terms were included in the model to test whether the variables differed in their effect on life satisfaction in the groups. Effect was calculated as the difference in the adjusted mean score between the MHP group and the general population in the UNIANOVA model divided by the standard deviation in the general population. RESULTS: Results showed that the mean total Quality of Life Scale score (maximum possible 112, minimum 16) was 71.3 for the MHP group and 83.1 for the general population (p<0.001; effect size 0.94). The groups also differed in subscales such as relationships and material well-being, health and functioning and personal, social and community commitment (p<0.001; effect sizes 0.85, 0.91 and 0.90, respectively). Age and the life satisfaction were more strongly positively correlated in the MHP group (r=0.40, p<0.001) than in the general population (r=0.05, p=0.03). CONCLUSIONS: Community residents with chronic MHP reported lower life satisfaction than the general population. Health care services for people with MHP should focus on life satisfaction as an important goal for treatment and care.

1122 /Quality of Life and Addiction Recovery: Patterns and Determinants Alexandre B. Laudet, Center for the Study of Addictions and Recovery, National Development and Research Institutes, Inc., New York City, NY, William L. White, Chestnut Health Systems/Lighthouse Institute, Bloomington, IL, Gordon R. Storey, Self Help Addiction Resource Centre Inc. (SHARC), Glenhuntly, Victoria, Australia AIMS: The desire for a better life is a key reason recovering persons cite for quitting drug use. The addiction field lags behind other biomedical fields in studying QOL outcomes. There is some evidence that QOL among active users is poor, but few studies have examined QOL and its determinants among recovering persons, especially persons with a history of drug (vs. alcohol) use. Moreover, existing studies use short follow-up periods that bear only on recovery initiation (vs. long-term recovery maintenance). This prospective study examines patterns and predictors of QOL and the association between recovery status and QOL. METHODS: Inner-city community-based individuals (N = 354) in recovery from severe substance use/dependence (primarily crack cocaine and heroin) ranging from 1 month to 10+ years were interviewed 3 times at yearly intervals (BL, F1 and F2) in New York City. Domains considered as QOL predictors in multivariate analyses were those identified from research conducted among general population and chronic disorder patients samples - e.g., stress, physical and emotional health, spirituality, social support. Analyses were conducted for the entire sample and by length of recovery subgroup to determine whether QOL predictors are similar or different at successive recovery stages. RESULTS: Longer recovery was consistently associated with higher QOL; QOL increased over time among persons who maintained recovery status (abstinence from drugs and alcohol) and higher QOL predicted sustained recovery a year later at each data collection point. F1 predictor domains accounted for between 19% and 36% of QOL variance at F2; longer recovery, lower stress and higher spiritual well-being were the most consistent predictors of F2 QOL for the total sample. CONCLUSIONS: This study represents a first step in considering QOL outcomes as a function of recovery from chronic drug addiction; findings that QOL increases over time suggest that the hope for a better life is a reality for many people in addiction recovery. Funded by NIDA Grant R01 DA14409 and by the Peter McManus Charitable Trust.

1318 /Responsiveness of the Adult ADHD Quality of Life Questionnaire Louis S. Matza, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Joseph A. Johnston, US Outcomes Research, Eli Lilly and Company, Indianapolis, IN, Douglas E. Faries, Outcomes Research, Eli Lilly, Indianapolis, IN, Karen G. Malley, Rockville, MD, Karin S. Coyne, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD AIMS: Childhood attention-deficit/hyperactivity disorder (ADHD) frequently persists into adulthood and continues to impair healthrelated quality of life (HRQL). The 29-item, self-report Adult Attention-Deficit/Hyperactivity Disorder Quality of Life Scale (AAQoL) was developed to assess HRQL among adults with ADHD. The measure has previously demonstrated good reliability and validity. The purpose of this analysis was to examine responsiveness of the AAQoL. METHODS: Data were from a randomized, placebocontrolled trial of atomoxetine among adults with ADHD. Patients completed the AAQoL, Conners' Adult ADHD Rating Scale (CAARS), SF-36, and Endicott Work Productivity Scale (EWPS) at baseline and week 8. Clinicians completed global measures of severity and improvement (CGI-ADHD-S, CGI-ADHD-I). Responsiveness of the AAQoL was examined through effect sizes and association with change in other measures (Spearman correlations, GLMs). RESULTS: Analyses included the 328 patients with AAQoL scores at baseline and week 8 (58.8% male; mean age = 36.9 years). All subscales (Life Productivity, Psychological Health, Life Outlook, Relationships) and the total score of the AAQoL reflected significant improvement from baseline to week 8 (p<0.0001). AAQoL change scores were significantly correlated with change in the CGI-ADHD-S (r = -0.37 to -0.50), EWPS (r = -0.43 to -0.63), and all CAARS symptom scores (r = -0.35 to -0.62) (all p<0.001). Correlations were also statistically significant, although generally smaller in magnitude, with change in psychosocial domains

1453 /Life Satisfaction Among People with Mental Health Problems Living in the Community Versus the General Population Eva Langeland, Health and Social Research, Bergen University College, Bergen, Norway, Astrid K. Wahl, Public Health and Primary Health, University of Bergen, Bergen, Norway, Monica W. Nortvedt, Health and Social Research, Bergen University College, Bergen, Norway, Kjell Kristoffersen, Berit R. Hanestad, Public Health and Primary Health, University of Bergen, Bergen, Norway

A-30

of the SF-36. AAQoL change scores significantly discriminated among patients with different levels of improvement as assessed by the CGI-ADHD-I at week 8, as well as change on the CGI-ADHD-S and CAARS. Effect sizes of AAQoL scales were in the moderate-tolarge range (0.65 to 1.07), and they were substantially larger than effect sizes for the SF-36 (0.03 to 0.43), a generic HRQL measure. CONCLUSIONS: The AAQoL was highly responsive to change in symptoms of ADHD, and it appears to be a useful outcome measure for treatments of ADHD in adults.

psychological or social dimensions of HRQL are included. However, it is unclear which variables predict outcome, to what extent, and whether there are differences in predictors between dimensions. The research question was: What are predictors of the physical and psychological dimension of HRQL after lung transplantation, and do they differ? METHODS: Results regarding the physical dimension (Nottingham Health Profile-energy scores), and the psychological dimension (ZUNG depression score) from 140 transplanted patients with a maximum follow-up of almost 10 years were analyzed using mixed model analysis. For both dimensions, the following variables were tested on their predictive value: Age, gender, diagnosis, year of transplantation, time spent on the waiting list, type of transplantation, presence of Bronchiolitis Obliterans Syndrome (BOS), and pre transplant HRQL scores. RESULTS: Patterns of change over time appeared to be comparable on both dimensions. With regard to the physical dimension, presence of BOS, age, gender, and pre-transplant energy score were significant predictors. BOS, and the pre-transplant depression score predicted the psychological dimension. CONCLUSIONS: Predictors differed between the physical, and the psychological dimension of HRQL. The physical dimension appeared to be influenced by both clinical and HRQL variables as well as demographic characteristics. The psychological dimension was soley influenced by clinical and HRQL features, and not by demographic characteristics. The present finding may be helpful when choosing, or developing interventions aimed at improving specific domains of HRQL after lung transplantation.

1338 /Outcomes of Depression in Primary Care: 6 Year FollowUp of the Lido Study in Barcelona PreliminaryFindings Ramona Lucas, Fundació Institut Català de l'Envelliment. UAB., Barcelona, Spain, Esther Jorda, Consorci Sanitari de Barcelona, Barcelona, Spain, Rosa Fernandez, Maria Luisa Rodriguez-Morato, CAP Horta, Barcelona, Spain, Donald Bushnell, Health Research Associates, Mountlake Terrace, WA, Mona Martin, Health Research Associates, Inc, Mountlake Terrace, WA AIMS: The Longitudinal Investigation of Depression Outcomes (LIDO) study was a cross-national observational study of major depression and its correlates in six countries worldwide. This presentation aims to report outcomes of Major Depression (MD) among primary care patients(PCP) in Barcelona, after a 6-year follow-up period. METHODS: The study population is a cohort of 195 subjects from three PCC who finished the LIDO study in 2000. Briefly, subjects were diagnosed with MD (1998-1999) and were followed for a period of 12 months, with follow-up evaluations at 6 week, 3, 9 and 12 months. Current follow-up started in August 2005. Subjects have been contacted by phone and invited to do the followup interview. Measures included: sociodemographic information, CES-D, QLDS, WHOQOL-BREF, SF-12, MHI-5, HSCL, a list of comorbid conditions, health care utilization, and socioeconomic variables. Subjects scoring 16 or greater on the CES-D complete the CIDI to assess current MD. All interviews are done face to face at the PCC. All participants have sign the inform consent. Analysis will be done using SPSS. RESULTS: Fifty-four percent have completed the current follow-up; 16% have moved out of the area; 3% have died; 6% did not want to participate; and the remaining 21% have to be scheduled for the interview. Fifty six percent of patients who completed the interview have CES-D <16, and 44% have significant depressive symptoms (CES-De16). Of these, two third met criteria for MD using the CIDI and one third did not. These are preliminary results. We expect to find worst outcomes among patients with MD compared to those without MD. CONCLUSIONS: The participation rate among subjects staying in the area is high. Longer follow-up periods of patients with depression might allow knowing other predictive variables of chronicity as well as long-term outcomes of MD. To our knowledge this is the first study in Spain in PCP with MD after a 6-year period.

1187 /The Relationship Between Measures of Asthma Control and Quality of Life Patricia M. Kenny, Madeleine T. King, Centre for Health Economics Research & Evaluation, University of Technology Sydney, Sydney, New South Wales, Australia, Guy B. Marks, Woolcock Institute of Medical Research, Sydney, New south Wales, Australia AIMS: It is generally expected that the health related quality of life (HRQOL) of people with asthma is associated with the control of asthma symptoms. Understanding this relationship may provide insights into the usefulness of various measures in different research contexts and indicate where to target interventions to improve HRQOL. This paper investigates this relationship using longitudinal data from a study supported by the Cooperative Research Centre for Asthma. METHODS: A stratified sample of 213 adults with asthma was recruited in 2002 via random population sampling and hospital emergency departments in New South Wales. Participants completed 6-monthly postal surveys for 3 years, with loss to follow-up of 6-10% each time (42% had missing data). Measures included: asthma control (sleep disturbance, use of short-acting beta agonists (SABA), activity limitation and asthma exacerbations), asthma-specific HRQOL (Sydney Asthma Quality of Life Questionnaire) and generic HRQOL (SF-36). The sources of HRQOL variation between individuals and within individuals over time were examined using multilevel models. RESULTS: Most of the variation in HRQOL was stable variation between people and some of this was explained by demographic factors. Together the asthma control variables explained 54-58% of the total variance in asthma-specific HRQOL, and 8-25% in generic HRQOL. Activity limitation was the main contributor to between person variation, explaining 48-49% in asthma-specific HRQOL and 9-27% in generic HRQOL. Sleep disturbance and SABA use were the main contributors to within person variation in asthma-specific HRQOL, explaining 28-40%. The asthma control measures explained minimal within person variance in generic HRQOL. CONCLUSIONS: As expected, asthma control was more closely related to the asthma-specific HRQOL measures than the generic measures. The within and between person variance analysis implies that sleep disturbance and SABA use are important asthma

Respiratory Diseases and QOL 1295 /Long-Term Health Related Quality of Life After Lung Transplantation: Different Predictors for Different Dimensions Karin M. Vermeulen, Office for Medical Technology Assessment, Wim van der Bij, Department of Pulmonary Diseases, Michiel E. Erasmus, Department of cardio-thoracic surgery, Elisabeth M. TenVergert, Office for Medical Technology Assessment, University Medical Center Groningen, Groningen, Groningen, The Netherlands AIMS: During the last decades, lung transplantation has proven to be an effective treatment option for patients with end stage lung disease with profound effects on both survival and Health Related Quality of Life (HRQL). When assessing HRQL, usually physical,

A-31

control measures for studies of treatment effectiveness (within person) and activity limitation for discriminating between groups.

interviewer-administered cohorts (inpatients and outpatients) and a postal cohort. There were two stages to the study; self-completion of the UKSIP and CFQoL in the first stage, and completion of the same two questionnaires seven to ten days later for the second stage. Sociodemographic and clinical data were extracted from patients' notes. RESULTS: The areas of HRQoL most impaired by CF were employment and concerns regarding the future. The UKSIP and CFQoL showed high reliability in the CF population; Cronbach's alpha internal consistency coefficients (0.865-0.929) and test-retest correlations (rs=0.57-0.94, p<0.005). Validity was variable; with the UKSIP showing discrimination across socio-demographic factors such as age and gender, whilst the CFQoL showed increased sensitivity to clinical variables like FEV1 (%predicted) and body mass index. Many parameters significantly influenced HRQoL with the greatest correlations seen with the Borg score (p<0.005); a measure of the sensation of dyspnoea. CONCLUSIONS: CF has a large impact on HRQoL, and its assessment as part of CF annual reviews is recommended to provide holistic patient care. Development of a Cardiff CF HRQoL instrument is needed to encompass both the generic and specific HRQoL areas highlighted in this study.

1714 /Initial Validation of the ASK-20 Survey Assessing Barriers to Medication Adherence Kristina S. Yu-Isenberg, Medical Affairs Commercial Operations, GlaxoSmithKline, RTP, NC, Louis S. Matza, Karin S. Coyne, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Jessica Psujek, Center for Integrative Medicine, Duke University Medical Center, Durham, NC, Maria Stoeckl, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Beth Skinner, Medical Affairs Commercial Operations, GlaxoSmithKline, RTP, NC, Ruth Quillian-Wolever, Center for Integrative Medicine, Duke University Medical Center, Durham, NC AIMS: Poor adherence with prescription medication is a widespread problem that contributes to deteriorating health and reduced healthrelated quality of life (HRQL). The patient-reported ASK-20 ("Adherence Starts with Knowledge") Survey assesses barriers to adherence among patients with various chronic conditions. This interim analysis is the first psychometric evaluation of the ASK-20. METHODS: Analyses were conducted with the first 48 patients recruited in a university outpatient setting. Participants were required to be prescribed medication for asthma, diabetes, or congestive heart failure for at least one year. Participants completed the ASK-20, Morisky Adherence Survey, SF-12 health survey (assessing HRQL), and additional questions regarding adherence behavior. Analyses examined ASK-20 distributional characteristics, reliability, and validity. RESULTS: The 48 participants were 74.5% female; 57.4% African-American with a mean age of 46.8 years. Means of the 20 items of the ASK-20 ranged from 1.4 to 3.3 on a 5-point scale (higher scores = greater adherence barrier). The mean total score was 44.9 (possible range: 20-100). Cronbach's alpha for the total scale was 0.75. The ASK-20 total score was significantly correlated with the Morisky (r = -0.60; p < 0.0001) and the SF-12 (r = -0.36; p < 0.05). The ASK-20 significantly discriminated among participants with different levels of adherence as assessed by the Morisky and groups differentiated by adherence behavior (e.g., the amount of time participants reported taking medications exactly as directed). CONCLUSIONS: In this interim analysis, the ASK-20 Survey detected barriers to medication adherence, while demonstrating adequate reliability, criterion validity, and discriminant validity. Data collection for this study is ongoing, and future analyses with a larger sample will further assess validity by examining the relationship between the ASK-20 and pharmacy refill rates. The ASK-20 may serve an important role in clinical settings to identify specific actionable barriers that are prevalent in patients with chronic conditions.

1817 /Quality of Life in Patients with Asthma Sara Miri, Zahra Pourpak, Immunology, Asthma and Allergy Research Institute, Tehran University of Medical Sciences, Tehran, Iran, Ali Montazeri, Mental Health, Iranian Institute for Health Sciences Research, Tehran, Iran, Hassan Heidarnazhad, Immunology, Asthma and Allergy Research Institute, Tehran University of Medical Sciences, Tehran, Iran AIMS: To study quality of life in patients with asthma. METHODS: This was a cross sectional study using the Iranian version of the Asthma Quality of Life Questionnaire (AQLQ) to measure quality of life in a sample of patients attending to an asthma clinic in Tehran, Iran. The AQLQ is a 32item asthma specific quality of life measure. Each item is rated on a 7-point Likert scale and the questionnaire provides one Overall Quality of Life score and four subscales namely: Activity Limitation, Symptoms, Emotional Function and Environmental Stimuli. Scores range from 1 to 7 and the higher scores for overall index and each subscales indicates a better condition. Data were analyzed using descriptive statistics. RESULTS: In all 92 patients were interviewed. The mean Overall Quality of Life score for patients was 3.6 (SD = 1.4) and for the Activity Limitation, Symptoms, Emotional Function and Environmental Stimuli subscales were: 3.6 (SD = 1.5), 4.1 (SD = 1.6), 3.7 (SD = 1.5), and 2.2 (SD = 1.9). There were no significant differences between men and women while significant differences was observed among sub-groups of patients that differed in the disease severity. The results are shown in the Table. CONCLUSIONS: The study finding suggest that asthma patients suffer more from poor environment and the disease severity has considerable effect on their quality of life.

1538 /Do Patient Reported Outcomes Improve Management of Adult Cystic Fibrosis? Sam Salek, Sarah Jones, Welsh School of Pharmacy, Cardiff University, Cardiff, UK, Meenu Rezaie, Carol Davies, Rebecca Mills, Ian Ketchell, Specialist Adult Centre for CF, Llandough Hospital, Cardiff, UK

Moder Intermittent Mild ate (n=23) (n=25) (n=19) Mean Mean Mean (SD) (SD) (SD) 4.0 3.2 Overall QOL 4.9 (0.91) (1.2) (1.0) 4.2 3.3 Activity-L 4.8 (1.0) (1.4) (0.87) 4.4 3.5 Symptoms 5.8 (0.87) (1.4) (1.4) Emotional 3.9 3.3 4.5 (1.4) Function (1.4) (1.6) 2.7 1.7 Enviromental-S 2.7 (2.1) (2.4) (1.3)

AIMS: Cystic fibrosis (CF) is an autosomal recessive disease that has a detrimental affect on multiple organs, resulting in premature fatality. The impact of CF on patients' HRQoL is liable to be great and cannot be assessed purely by physiological parameters, making CF patients ideal candidates for the application of HRQoL instruments. The aim of this study was to assess the affect of CF on HRQoL, to ascertain the validity and reliability of the UKSIP and the CFQoL in the adult CF population, and to examine the clinical and socio-demographic influences on HRQoL. METHODS: Over a ten week period, seventy participants were recruited from the Specialist Adult Centre for CF at Llandough hospital, Cardiff, UK; using

A-32

Severe (n=25) Mean (SD) 2.4 (0.76) 2.1 (0.79) 2.6 (0.87) 2.9 (1.5) 1.5 (1.3)

P <0.001 <0.001 <0.001 0.003 0.04

distress [Hospital Anxiety and Depression scale (HAD) and Impact of Event Scale (IES)]. RESULTS: Of all patients, only 17% reported pain from the endoscopy, but 65% reported discomfort from the procedure in general. Apart from an increase in sore throat (47% after endoscopy versus 22% before) the procedure did not cause symptoms. Patients’ distress levels (HAD) were increased in the week prior to the endoscopy compared to the week after. BE patients reported significantly less overall discomfort (59% versus 70% in the control group) and lower distress levels. Having had prior endoscopies only partly explained these differences. A higher perceived risk of malignancy was not associated with more discomfort or psychological distress. CONCLUSIONS: Upper GI endoscopy is burdensome for the majority of patients causing discomfort and moderate psychological distress. Patients under regular surveillance for Barrett’s esophagus reported significant, but less severe burden than the control group. Although adaptation may help subjects to cope, recommendations for endoscopic surveillance should be evidence-based and take into account the discomfort and distress of upper GI endoscopy for patients.

QOL and Gastrointestinal Disease 1014 /Cultural Adaptation and Application of the IBS-QOL in China:A Disease-Specific Quality-Of-Life Questionnaire Wen-wei Huang, Piwei Institute, Guangzhou University of Traditional Chinese Medicine, Shantou City, Guangdong Province, China, Fu-sheng Zhou, Piwei Institute, Guangzhou University of Traditional Chinese Medicine, Guangzhou, Guangdong, China, Donald M. Bushnell, Seattle Office, Health Research Associates, INC, Mountlake Terrace, WA, Chiaka Diakite, Public Health Department of Traditional Medicine, National Institute of Research in Bamako, Bamako, Mali, Xiao-huan Yang, School of Basic Medical Sciences, Guangzhou University of Traditional Chinese Medicine, Guangzhou, Guangdong Provinde, China AIMS: The aim of the study was to examine the validity of a translated and culturally adapted version of the Irritable Bowel Syndrome-Quality of life questionnaire (IBS-QOL) in patients suffering from IBS in China. METHODS: A structured procedure was used for the translation and cultural adaptation of the original English IBS-QOL into Chinese. The questionnaire was administered to 61 clinical patients with IBS and 70 healthy individuals. Psychometric testing for reliability, validity and responsiveness followed standardized procedures. Test retest reliability was tested in 61 randomly selected patients, who received the questionnaire a second time, 4 weeks later. 70 healthy subjects were not assessed at follow-up. All enrolled patients also completed the Short Form-36 Health Survey (SF-36) at the baseline visit. Responsiveness to treatment was assessed. RESULTS: The average length of time required to complete the questionnaire was short (5.63 minutes for IBS patients and 5.54 minutes for healthy subjects by selfadministration). Internal consistency (Cronbach s alpha) values ranged from 0.920 to 0.941 for the Chinese IBS-QOL subscales with the Relationships scale having the lowest alpha. The convergent and discriminate validity results comparing the Chinese translation of the IBS-QOL overall score and the SF-36 subscales confirmed our predicted hypothesis. The Chinese IBS-QOL scores are more highly correlated with Social Functioning, Vitality and General Health (SF36) and show weaker associations with physical functioning, Role Physical, Mental Health, and Bodily Pain (SF-36). The Chinese translation of the IBS-QOL was responsive to treatment. CONCLUSIONS: In general, the Chinese translation of the IBSQOL, after cultural adaptation and revision, possesses good reliability, validity and responsiveness. It is a reliable and valid instrument to assess the quality of life in Chinese patients suffering from IBS and will be prudent for further clinical trials or for related research projects in China.

1285 /Anxiety and Depression in Patients with Gastrointestinal Cancer Azadeh Tavoli, Mohammad Ali Mohagheghi, Cancer Research Centre, Cancer Institute, Tehran, Iran, Ali Montazeri, Mental Health, Iranian Institute for Health Sciences Research, Tehran, Iran, Rasool Roshan, Zahra Tavoli, Cancer Research Centre, Cancer Institute, Tehran, Iran AIMS: As part of a styudy on quality of life in patients with gastrointestinal cancer an investigation was carried out to examine anxiety and depression in this group of patients. METHODS: This was a cross sectional study of anxiety and depression in patients with gastrointestinal cancer attending to the Tehran Cancer Institute. Anxiety and depression was measured using the Hospital Anxiety and Depression Scale (HADS). This is a widely used valid questionnaire to measure psychological distress in cancer patients. The HADS is a 14-item questionnaire consisting of two sub-scales: anxiety and depression. Each item is rated on a 4-point scale, giving maximum scores of 21 for anxiety and depression. Scores of 11 or more on either sub-scale are considered to be a significant 'case' of psychological morbidity, while scores of 8 to 10 represent 'borderline' and 0 to 7 'normal'. Demographic and clinical data also were collected to examine anxiety and depression in sub-group of patients. RESULTS: In all 142 patients were studied. The mean age of patients was 54.1 (SD = 14.8), 56% were male, 52% did not know their cancer diagnosis, and their diagnosis was related to esophagus (29%), stomach (30%), small intestine (3%), colon (22%) and rectum (16%). The mean anxiety score was 7.6 (SD = 4.5) and for the depression this was 8.4 (SD = 3.8). Overall 25% and 23% of patients scored high on both anxiety and depression. There were no significant differences between gender, educational level, marital status, and cancer site and anxiety and depression scores whereas those who knew their diagnosis showed a significant higher degree of psychological distress (mean anxiety score: knew diagnosis 9.1 vs. 6.3 did not know diagnosis, P < 0.001; mean depression score: knew diagnosis 9.1 vs. 7.9 did not know diagnosis, P = 0.05). CONCLUSIONS: Contrary to expectation psychological distress was higher in those who knew their cancer diagnosis. It seems that the way we provide information for cancer patients has an important role in their improved or decreased quality of life.

1389 /The Burden of Upper Gastrointestinal Endoscopy Marie-Louise Essink-Bot, Michelle E. Kruijshaar, Public Health, Marjon Kerkhof, Gastroenterology and Hepatology, Ewout W. Steyerberg, Public Health, Peter D. Siersema, Gastroenterology and Hepatology, Erasmus MC, Rotterdam, The Netherlands AIMS: Regular endoscopic surveillance is recommended for patients with Barrett’s esophagus (BE) for early detection of esophageal cancer. However, upper gastrointestinal (GI) endoscopy is an invasive procedure that may cause pain, discomfort and psychological distress. We examined the burden of regular endoscopic surveillance by comparing BE patients under regular surveillance and a control group. METHODS: A total of 394 patients (180 BE patients and 214 patients with non-specific GI symptoms) filled out questionnaires one week before, on the day of, one week after and one month after upper GI endoscopy. Burden of endoscopy was assessed by three variables: 1) pain and discomfort experienced during endoscopy, 2) symptoms, 3) psychological

1016 /Multivariate Quality of Life (QOL) Prognostic Factor Analysis in Hepatocellular Carcinoma (HCC) Franck Bonnetain, Biostatistics Epidemiology, FFCD & Centre regional de lutte contre le cancer, Dijon, France, Michel Doeffel, Gastroenterology, CHU Strasbourg, Strasbourg, France, Patrick

A-33

Arveux, DIM, Centre Georges Francois Leclerc, Dijon, France, Olivier Bouche, Gastroenterology, CHU REIMS, REIMS, France, Laurent Bedenne, Gastroenterology, CHRU le Bocage Dijon, Dijon, France

6.61) in patients who received DLB. Mortality and morbidity were, respectively, 7.8 % and 58 % for PR, and 2.6 % and 42 % for DLB. QoL decreased more after PR than after DLB. The DLB-group recovered quicker, reaching pre-operative QoL levels after 3 months, and were less impaired when discharged. The LAD-subgroup and the MD-subgroup presented with equal levels of QoL. CONCLUSIONS: QoL analysis revealed favorable QoL-data after DLB. Additionally, the survival rates of the two groups did not differ significantly, but morbidity and mortality rates in the PR group were elevated. Therefore, the use of PR for advanced pancreatic cancer needs to be carefully evaluated.

AIMS: The aim of this study was to explore whether QoL predicts survival in patients with non-operable HCC. METHODS: Between 1995 and 2002, 123 cirrhotic patients of Child-Pugh class A or B with unresectable HCC of Okuda stage I or II were included in a randomized multicentric phase III trial (FFCD 9402). At baseline, QoL was assessed with Spitzer index (0 to 10). The overall survival (OS) was calculated until death or last follow-up using the KaplanMeier method. Analyses of potential pronostic factors was performed using univariate Cox model to calculated hazard ratio (HR) with 95% CI. Multivariate analysis was performed using backward stepwise Cox proportional hazards regression modeling (p = 0.05 to add and 0.10 to sort) RESULTS: According to Spitzer QoL Index (N= 94) the median of OS ranged from 2.4 (Spitzer = 6) to 21.8 (Spitzer =10)months. It was respectively 3.9 and 18.6 months among patients with QoL < 9 (N=36) and >= 9 (N=58). Factors included for multivariate analyses were : Spitzer (ordinal score, p< 0.0001), WHO performance status > 0 (p =0.015), sex (NS), Okuda stage (II vs I, p< 0.0001), included patients by center < 6 (NS), treatment (Chemoembolization vs Tamoxifen, NS), alcohol cirrhosis (NS), Child-Pugh class (B vs A, p< 0.0001 ), hepatomegaly (p =0.03), hepatalgia (p< 0.0001), age e 65 years, (NS), AFP e 400 ng/ml (p< 0.0001 ), icterus (p=0.01), Diameter of the main nodule e 5cm (p=0.14), tumor stage multinodular (p =0.014), and portal vein thrombosis (NS). Seven independent prognostic factors were retained by multivariate analysis: Spitzer QoL Index (HR = 0.79 [0.6 - 0.97]), Hepatalgia (HR = 2.31 [1.3 - 4.3], sex (HR = 0.49 [0.2 - 0.97], ChildPugh (HR = 2.01 [1.2 - 3.44], AFP (HR = 2.47 [1.5 - 4.11], Center (HR = 2.46 [1.3 - 4.60]) and hepatomegaly (HR = 1.92 [1.2 - 3.12]) CONCLUSIONS: Our results highlighted the pronostic value of Spitzer QoL index. We plan to performe a pooled analysis of 3 trials among HCC patients to confirm our results and to explore the pronostic value of QlQ-C30 scores.

Psychometrics: New Results and Insights 1 1683 /Evaluation of Using WHOQOL-BREF and SF-36 in Measuring Quality of Life Among Abused Women Kjersti Alsaker, Faculty of Health and Social Science,, Bergen University College, Bergen, Norway, Bente E. Moen, Department of Public Health and Primary Health, Section for Occupational Medicine, Bergen, Norway, Kjell Kristoffersen, Department of Public Health and Primary Health Care, Faculty of Medicine, Section for Nursing Sciences, Bergen, Norway AIMS: The aims were to determine what WHOQOL-BREF and SF36 reveal about HRQOL in a group of abused women, how they correlate with each other and how they correlate with experienced physical and psychological violence. The possible use of the instruments in clinical practice is discussed as well. METHODS: A cross-sectional study was performed at all women s shelters in Norway from October 2002 to May 2003. All 50 shelters agreed to participate, and 42 had residents who met the inclusion criteria in the study period: understanding Norwegian and staying in the shelter for at least 1 week as a result of violence from an intimate partner. RESULTS: Results The WHOQOL-BREF and the SF-36 Health Survey were used to study the quality of life of women at women s shelters in Norway. One week after arriving a women shelter in Norway and one year later. Violence was measured using the Severity of Violence against Women Scale (SVAWS) and the Psychological Maltreatment of Women Index (PMWI). The mean WHOQOL-BREF and SF-36 scores were all low at baseline. The correlations between WHOQOL-BREF and SF-36 were mostly high. The WHOQOL-BREF domains explained 46% of the variance in overall health and 40% of the variance in overall quality of life in WHOQOL-BREF. The SF-36 domains explained 66% of the variance in overall health and 27% of overall quality of life. SVAWS correlated more strongly with SF-36 domains than with WHOQOLBREF domains. PMWI correlated more strongly with WHOQOLBREF than with SF-36, showing the more psychological foundation of WHOQOL-BREF CONCLUSIONS: Both instruments gave an important view of the quality of life of abused women. As the SF-36 has more domains that give more specific information, it is probably the best instrument to discover the need for intervention. Because of the many barriers to screening for domestic violence, the use of SF36 or WHOQOL-BREF in clinical settings could be an uncontroversial way to discover women who need professional health care.

1079 /QoL Analyses After Bypass Surgery Versus Palliative Pancreaticoduodenectomy in Patients with Advanced Ductal Adenocarcinoma of the Pancreatic Head Bodo Schniewind, Surgery, Beate Bestmann, Reference Center on Quality of Life in Oncology, University of Schleswig-Holstein, Campus Kiel, Kiel, Germany, Doris Henne-Bruhns, Visceral and Transplantation Surgery, University Clinic of Ulm, Ulm, Germany, Fred Faendrich, Bernd Kremer, Surgery, University SchleswigHolstein, Campus Kiel, Kiel, Germany, Thomas Kuechler, Reference Center on Quality of Life in Oncology, University of SchleswigHolstein, Campus Kiel, Kiel, Germany AIMS: In some centers, palliative resection (PR; partial pancreaticoduodenectomy) is, in selected cases, promoted in preference to double loop bypass (DLB) surgery for advanced pancreatic cancer. This prospective study compares PR with DLB, placing particular focus on patients quality of life (QoL). METHODS: From 01/1993 to 09/2004, 167 patients were analyzed in a prospective single center study of palliative surgical treatment of advanced ductal adenocarcinoma of the pancreatic head. Thirty-eight underwent PR and 129 underwent palliative DLB. Patients undergoing DLB were divided into: 1) locally advanced disease (LAD-subgroup; n=42; 33%) and 2) metastasized disease (MDsubgroup; n=87; 67%). QoL was assessed using the EORTC QLQC30 questionnaire supplemented by a pancreatic cancer specific module. QoL-data were collected pre-operatively and for up to 12 months after surgery. RESULTS: Median survival was 7.0 months (95% CI: 4.09; 9.91) in PR patients and 6.0 months (95% CI: 5.39;

1223 /Validation of a Scale to Measure Patient-Perceived Barriers to Medication Use David G. Marrero, Patrick O. Monahan, Kathleen A. Lane, Medicine, Indiana University School of Medicine, Indianapolis, IN, Risa P. Hayes, Global Health Outcomes, Eli Lilly and Company, Indianapolis, IN

A-34

AIMS: Medication adherence may explain why patients show very different clinical outcomes. Previous assessments of adherence have used refill rates and pill counts. Few studies have investigated patient-identified barriers to using medications as prescribed. This is particularly true for persons with diabetes, most of whom are using poly-pharmacy regimens. We created a questionnaire to measure patient perceptions of barriers to medication adherence (PBMA) targeting a predominately low income, inner-city minority population. METHODS: Twenty items (Likert scale) leading with "I sometimes don't take my diabetes medicines because..." were developed from 5 focus groups (N=48). A questionnaire including these items was mailed to 1000 patients who have diabetes. RESULTS: Using data from 267 respondents (Mean age=58, 74% female, 43% non-Hispanic Caucasian, 77% income<$15,000), exploratory factor analyses with varimax rotation identified 5 factors, that may contribute to poor medication adherence: personal access to medications (F1); communication with providers (F2); understanding or appropriately following the prescribed regimen (F3); side effects (F4) and system factors that inhibited access to medication (F5). Cronbach alphas ranged from .73 to .83 for the five factors and was .92 for total scale score. No relationships were found between total scores and gender, race, or income. Greater perception of barriers was significantly (p<0.01) associated with being younger (r= -0.21), being bothered more by physical (r= -0.40) and emotional side effects (r= 0.43), and less satisfaction with control of blood sugar by diabetes medications (r= 0.45). CONCLUSIONS: This instrument is reliable, factorially valid, and consistent with clinical observation regarding factors known to be associated with patient medication adherence. Although study participants were patients with diabetes, the PBMA may be applicable to other therapeutic areas.

1691 /Quantitative Aspects of an Operational Procedure to Assess Cross-Cultural Equivalence: Description and Applications Antoine Regnault, LASS, Université Claude Bernard Lyon 1, Lyon, France, Christine de La Loge, Psychometrics and Statistics, Benoit Arnould, Patient Scales for Clinical Practice, MAPI Values, Lyon, France, Michel Lamure, LASS, Université Claude Bernard Lyon 1, Lyon, France AIMS: Cross-cultural validity is an essential property of PatientReported Outcome instruments when used in an international context. This paper describes and illustrates the quantitative aspects of an operational procedure allowing the detection of cross-cultural equivalence limitations. METHODS: The procedure, based on Herdman's Universalist model, explores six types of equivalence (conceptual, item, semantic, operational, measurement and finally the resulting functional equivalence). Advanced quantitative methods are included in the procedure to consider various issues: Confirmatory Factor Analysis (CFA) for conceptual equivalence and Differential Item Functioning (DIF) detection techniques for item and semantic equivalence. The procedure was applied to the Psychological General Well Being (PGWB) questionnaire using data from the large international IQOD database. The French (n=849) and Japanese (n=500) versions were compared to the original US English version (n=2567). RESULTS: The CFA allowed a deficiency in the conceptual equivalence between the Japanese and the US English versions of the PGWB to be identified. The conceptual equivalence of the French and US English versions was found to be acceptable by CFA. Then, DIF was detected, indicating a lack of item or semantic equivalence, for several items of these versions. CONCLUSIONS: The quantitative methods included in the procedure proved their capacity to flag potential cross-cultural differences, both at the general level (conceptual equivalence) and at the local item level (item or semantic equivalence). Applying the complete assessment procedure, using the quantitative methods together with linguists' or sociologists' expertise, would enable the conceptual dissimilarities or potential sources of bias that may be responsible for a distorted functional equivalence of PRO instruments to be detected and understood.

1383 /Social Isolation: Its Impact on Quality of Life Graeme Hawthorne, Psychiatry, University of Melbourne, Parkville, Victoria, Australia AIMS: Social isolation is a major consequence of demographic inequalities and health status, and it has a profound influence on selfreported quality of life. All too often, however, its effects are difficult to appraise or include in outcomes or quality of life research. This paper introduces the Friendship Scale, a short, user-friendly measure of social isolation. METHODS: The Friendship Scale, consisting of 6 items, was administered as part of a routine population survey to 3000 South Australians representative of the Australian population. RESULTS: The results showed that 7% of respondents were socially isolated and 9% were somewhat isolated. When putative predictors of social isolation were examined using logistic regression, older respondents were less likely to be isolated, those without a partner were more isolated, those out of the workforce were also more likely to be isolated, as were those with poor general health. Those classified with depression were 12 times (ORadj = 12.33) more likely to be socially isolated when compared with non-depressed respondents. Regarding the impact of social isolation on health status and quality of life, the socially isolated consumed significantly more health care resources and obtained significantly lower scores on the PCS and MCS of the SF36V2. The socially isolated also obtained significantly lower mean utility scores on the AQoL, EQ5D, and HUI3 that were between 0.15 0.19 utilities below those of the socially connected. Regression modelling of the data suggested that inclusion of Friendship Scale scores increased the explanatory power (R2) of putative predictors of health status and quality of life by up to 0.15%. CONCLUSIONS: The Friendship Scale is a highly sensitive and user-friendly measure of social isolation. Not only is it responsive to putative predictors of social isolation, but inclusion of it has a significant effect on improving regression models predicting health status and quality of life. It seems likely that social isolation should be included in multivariable models of quality of life research.

1569 /Validation of the Revised Patient Perception of Migraine Questionnaire for Use in Clinical Trials: Measuring Satisfaction with Acute Migraine Treatment Dennis A. Revicki, Miriam Kimel, Wen-Hung Chen, Julie McCormack, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, Maryland, Steven P. Burch, Global Health Outcomes, GlaxoSmithKline, Research Triangle Park, North Carolina AIMS: To evaluate in a clinical trial setting the psychometric properties of the revised Patient Perception of Migraine Questionnaire (PPMQ-R) - a patient satisfaction measure previously validated in clinical practice. METHODS: The PPMQ-R was tested in 1,304 migraineurs randomized to participate in two identical phase 3, single-attack trials comparing a fixed-dose, single-tablet combination of sumatriptan 85 mg formulated with RT Technology and naproxen sodium 500 mg to placebo. Study medication was taken orally within 1 hour of head pain onset and when pain was mild. Primary endpoints were pain and associated symptoms assessments. The PPMQ-R was completed 24 hours post dosing. The PPMQ-R includes satisfaction with efficacy, functionality, ease of use, cost, and tolerability subscales. Because cost is not a factor in clinical trials, the 2-item scale was not measured. Internal consistency reliability and construct and discriminant validity were evaluated using Cronbach's alpha, Pearson or Spearman correlations, and analysis of variance, respectively. RESULTS: The sample was 88% female and 86% white, with a mean age of 40 years (SD=11).

A-35

PPMQ-R scale scores and total score (efficacy, functionality, and ease of use) showed very good internal consistency reliability (alpha 0.84 to 0.99). Efficacy, functionality and total PPMQ-R scores showed large, inverse relationships with migraine pain severity, number of migraine symptoms, and work ability (r -0.62 to -0.75; all P< 0.0001) and moderate, inverse relationships with migraine recurrence measures and additional medications (r -0.33 to -0.51; all P< 0.0001). All but the tolerability scale discriminated among migraine pain severity levels and number of migraine symptoms 4 hours post dosing (all P< 0.001). CONCLUSIONS: The PPMQ-R is a valid and reliable instrument for measuring patient satisfaction with acute migraine treatment, an important benchmark of quality and effective care, in both clinical trial and practice settings.

HIV infected patients receiving HAART enrolled in the ATHENA cohort. Patients completed the Medical Outcomes Study HIV (MOS HIV), a symptom checklist, and the Beliefs about Medication Questionnaire (BMQ) every six months until June 2005. Growth curve models with random intercepts and slopes were used to investigate change over time. RESULTS: A total of 391 out of 517 (76%) eligible patients participated in the study of whom 90% were males and 51% were antiretroviral naïve. The mean age was 41 years. QoL improved significantly over time regarding both physical health and mental health. Additionally, most disease- and treatment-related symptoms decreased significantly over time. However, several symptoms increased significantly, i.e., pain in legs, sore muscles, numb feeling in fingers or toes, tingling hands or feet, abnormal body fat distribution, decreased interest in sex, difficulty seeing, and pain when urinating. Overall, patients endorsed the benefits of HAART. For instance, 87% of patients believed that their present health depends on HAART and 88% believed that their health in future depends on HAART. The two most frequently expressed concerns were worry about long-term effects of HAART, expressed by 73% and worry about getting an adverse effect of HAART in the near future, expressed by 54%. Patients' beliefs about the necessity of HAART and concerns about potential side effects of HAART remained stable over time. CONCLUSIONS: The observed overall improvement in long-term QoL adds to the well-established clinical benefits of HAART. Generally, the perceived benefit of HAART outweighed the burden of HAART.

AIDS and QOL: Key Results 1469/ HIV/AIDS Patients' Quality of Life Appraisal Depends on Their Personal Meaning Of Quality of Life and Frame of Reference Yuelin Li, Bruce Rapkin, Psychiatry & Behavioral Sciences, Memorial Sloan-Kettering Cancer Center, New York, NY AIMS: This study aims to show that patients' reponse to conventional QOL assessments may be moderated by their frame of reference. METHODS: Data for this study were gathered as part of an ongoing evaluation of Medicaid-supported care for people living with HIV/AIDS in New York City. As part of a longitudinal survey, we gathered data on patients' quality of life using the MOS-SF36. Also administered was the QOL appraisal battery discussed by Rapkin and Schwartz (2004). The brief assessment battery is designed to assess the criteria that individuals use in responding to QOL measures including aspects of their frames of reference (i.e., concerns and priorities most important to their well-being). RESULTS: Two hundred respondents completed the assessments: mean age = 47 with 10 years since HIV Diagnosis; 43% were female, 54% African American, and 43% Hispanic. Many patients reported physical symptoms: fever (64%), trouble in mouth (88%), nausea and vomiting (61%), rash (81%), and feeling depressed (43%). The symptoms had a detrimental impact on patients' Physical and Mental Component Scores in MOS-SF36 (all p < 0.05). However, patients experienced less detrimental impact in PCS and MCS if they focused on life goals related to survival, comfort/contentment, personal growth and reminiscence. Linear regression showed subtle yet statistically discernible moderating effects on PCS and MCS above and beyond what was predicted by physical symptoms (incremental R^2 = 6% in PCS and 8% in MCS, both p < 0.05). CONCLUSIONS: QOL assessments should consider the subtle effects of patients' personal meaning of QOL. There is a need to develop a Rasch-based QOL appraisal model to incorporate these parameters as part of QOL measurement.

1743/ Quality of Life of the Positive HIV Patients and Individuals with Risk for HIV Infection Aparecido Diniz, Angélica Belasco, Lucila C. Vianna, Nursing, Ricardo Sesso, Medicine, Dulce Barbosa, Nursing, Federal University of São Paulo, São Paulo, São Paulo, Brazil AIMS: This study aimed to assess the QoL in asymptomatic positive HIV and negative individuals with risk of HIV infection to correlate dimensions of SF-36 and HAD. METHODS: We studied positive HIV and negative individuals n=3l with risk of HIV infection n=30.This study carried out in the Center for Immunological Deficiencies Control.Their QoL the two groups was evaluated by the SF-36 and HAD. RESULTS: 52% the patients were male and in the contacts 23%.Among the HIV positive group 55% were single against 2% of the contacts and 32% married against 67%,74% were heterosexual against 100% of the contacts. Monthly individual and family income of the positive patients were significantly higher than the contacts. Analysis of QoL showed a trend to present better scores in physical aspect, pain, general aspects, emotional and mental health and poorer in physical capacity, vitality and social aspects, in the positive patient group, in spite of not being significant. Regarding HAD, the group of positive patients also showed lower indexes of anxiety and depression when compared to contacts. A negative correlation between mental health and anxiety r=-0.57 and between mental health and depression r=0.38 and positive correlation between anxiety and depression r=0.66, considering error µ 0.01. There was a statistically significant difference regarding depression of HAD when compared to positive patients p=0.05. Anxiety measured significantly correlates to the mental health in HIV-infected patients and their contacts. Significant association was observed between anxiety and vitality, social issues and psychotropic use. Patients depression was significantly correlated with vitality, general health indicators, and marital status. CONCLUSIONS: The use of indicators to evaluate patients and their contacts QoL may help direct the general health management approach to prevent psychologic disturbances. The search for a good QoL of HIV infected patients and their contacts must encompass proper therapy, social and emotional integration.

1343/ Antiretroviral Therapy for HIV-1 Infection: Patients' Long-Term Quality of Life Pythia T. Nieuwkerk, Medical Psychology, Academic Medical Center / University of Amsterdam, Amsterdam, the Netherlands, Mirjam A. Sprangers for the ATHENA study group, Medical Psychology, Academic Medical Center, Amsterdam, the Netherlands AIMS: The advent of highly active antiretroviral therapy (HAART) for HIV-1 infection has been accompanied by significant reductions in HIV related morbidity and mortality. However, toxicities and pill burden associated with most HAART regimens could diminish patients' quality of life (QoL). We investigated long-term QoL and trade-offs between burden and benefit of therapy among HIV-1 infected patients on HAART. METHODS: Between 1998 and 2000,

A-36

school (60%). The arithmetic mean value of respondent's own health state was 0.55[0.51-0.58; 95% CI]. The scores given to the health states were: SHI 0.63[0.60-0.65]; MIADI 0.39[0.37-0.42] and MAADI 0.15[0.13-0.17]. After ranking the health states, 66(44%) individuals revised their health status to 0.66[0.62-0.69]. Death was scored equal to zero in 99% of the cases and 26(17.3%) participants considered MADDI as bad as death. MOS-HIV had good response and acceptability. Factor analysis revealed that physical health (PH) and mental health (MH) components together accounted for 57% of the variance. Overall PH had the highest factor loadings from vitality, pain and physical subscales and for overall MH the highest factor loading was quality of life. Scale scores were lower for general health perceptions, pain and quality of life. CONCLUSIONS: Older individuals provided lower scores for self-assessement, reconsideration of health status, and for the MOS-HIV scale scores of social and role functioning, and for vitality. Age did not discriminate for the predetermined HIV/AIDS health states assessment using the VAS. It is important to assess if these differences are also found in HIV-infected individuals receiving antiretroviral therapy and whether they influence the findings of clinical trials.

1177/ ISSQoL: A New Questionnaire For Evaluating The Quality of Life of People Living with HIV in the HARRT Era Raffaella Bucciardini, Dipartimento Del Farmaco, Istituto Superiore di Sanità, Rome, Italy, Italy, Rita Murri, Dipartimento delle Malattie Infettive, Università Cattolica del Sacro Cuore, Rome, Italy, Europe, Stefano Vella, Dipartimento Del Farmaco, Istituto Superiore di Sanità, Rome, Italy, Europe AIMS: To design a Health-related Quality of Life (HRQoL) instrument for the new needs of HIV-infected people in the new era of HAART. METHODS: The self-administered questionnaire was built by an Italian network including researchers, physicians, people living with HIV, national institutions and community-based organizations (CBO). Several steps were undertaken to develop the questionnaire: i) review of existing HRQoL literature and questionnaires for HIV-infected people; ii) selection of the still sound domains, and identification of further domains related to new aspects of HRQoL in the era of HAART; iii) pre-test analysis carried out in two consecutive steps. Each step was characterized by structured interviews (n=100) to Italian HIV-positive people equally distributed throughout the country. The objective of the first step was to verify the usefulness of the questionnaire, to construct a form easily understandable to everyone, to define the domains and their meaning; the second pre-step aimed at evaluating and readapting first step outcomes by means of statistical analysis; iv) validation analysis, a cohort of 350 people with HIV was recruited. RESULTS: The internal consistence reliability (Cronbach's alpha) was >=0.70 for all domains. Most domains had Cronbach's coefficient >0.80. All domains demonstrated convergent and discriminant validity. The final version of ISSQOL includes two sections: HRQoL Core Evaluation Form (9 domains) and Additional Important Areas for HRQoL (6 domains). New domains of HRQoL were found to be essential in PLWAH in the era of HAART such as: sexual life, impact of body changes on physical and mental health, impact of antiretrovirals on well-being, short and long term life planning, motherhood/fatherhood. CONCLUSIONS: Data at present available provide preliminary evidence of the reliability and validity of the ISSQoL questionnaire for the measurement of HRQoL in HIVinfected people. The direct involvement of HIV-positive people in all the phases of the project was a key aspect of our work.

Preferences Research 1 1245 /The Predictive Validity of Two Measures of Health-Related Quality of Life: Mortalilty in a Longitudinal Population-Based Study David H. Feeny, Kaiser Northwest Center for Health Research, Health Utilities Incorporated, Portland, OR, Mark S. Kaplan, Community Health, Portland State University, Portland, OR, JeanMarie Berthelot, Health Analysis and Modelling Group, Statistics Canada, Ottawa, ON, Canada, Bentson McFarland, Psychiatry, Oregon Health & Science University, Portland, OR, Saeeda Khan, Health Analysis and Modelling Group, Statistics Canada, Ottawa, ON, Canada AIMS: Aims. To examine longitudinally and with a populationbased sample the association between baseline health-related quality of life (HRQL) and mortality risk, and to compare the predictive ability of Health Utilities Index Mark 3 (HUI3) with self-rated health (SRH) after adjusting for key covariates. METHODS: Methods. A 1994/95 nationally representative sample of 12,334 women and men aged 18 years and older participating in the Canadian Longitudinal National Population Health Survey were followed for 8 years. The main outcome measure was all-cause mortality. The predictive validity of baseline overall HUI3 scores and SRH were analyzed using a Cox proportional hazards model, controlling for standard determinants of health. RESULTS: Results. Hazard rates for respondents reporting good health, fair health, and poor health at baseline were, respectively, 1.45 (95% Confidence interval 1.05 to 2.01), 1.94 (1.33 to 2.84), and 3.13 (2.04 to 4.83) compared to those subjects who reported excellent health. In a model excluding SRH, the effect of the baseline HUI3 score on mortality was strong and significant (hazard ratio rate = 0.50; 95% CI 0.35 to 0.72) when controlling for a wide array of potential confounders. When HUI3 and SRH were considered simultaneously, the effect of HUI3 on mortality was somewhat attenuated, but still significant (HR = 0.66, 0.45 to 0.97). CONCLUSIONS: Conclusions. Both SRH and HUI3 are strong predictors of subsequent mortality controlling for potential confounders. HUI3 adds to the mortality prediction ability of SRH.

1673/ Assessment of Quality of Life in HIV Infected Ugandans Using the Visual Analogue Scale (VAS) and the MOS-HIV Health Survey Antonieta Medina Lara, HIV/AIDS Knowledge Programme, Liverpool School of Tropical Medicine, Liverpool, UK, Barbara Nyanzi Wakholi, Social Science, Medical Research Council Unit in Uganda, Entebbe, Uganda, Agnes Kasirye, Statistics, Medical Resarch Council Unit in Uganda, Entebbe, Uganda, Christine Watera, Entebbe Cohort, Medical Research Council Unit in Uganda, Entebbe, Uganda, Charles F. Gilks, HIV/AIDS, World Health Organization, Geneva, Switzerland, Heiner Grosskurth, HIV/AIDS, Medical Research Council Unit in Uganda, Entebbe, Uganda AIMS: To evaluate the use of Visual Analogue Scale (VAS) and MOS-HIV in individual self-assessment health status and evaluation of predetermined HIV/AIDS health states by HIV-infected Ugandans. METHODS: HIV-infected individuals (n=159 adults (>18 years)) from the Entebbe Cohort clinic were enrolled. Each participant was interviewed face-to-face in Luganda and was asked to: 1)answer the MOS-HIV survey; 2)rate his/her own health state; 3)score three predetermined health states for HIV/AIDS (symptomatic HIV infection (SHI), minor AIDS defining illness (MIADI) and major AIDS defining illness (MAADI) and 4)reconsider their first own health evaluation. RESULTS: Participants were mostly females (76%)with mean age of 36.6 years that have with completed primary

1723 /A Comparison of Three Indirect Utility Measures and Two Disease-Specific Instruments for Measuring Quality of Life in Asthma Carlo A. Marra, Helen McTaggart-Cowan, Larry D. Lynd, Judith A. Soon, Victorine Ssozi, Pharmaceutical Sciences, Aslam Anis, Health

A-37

Care and Epidemiology, University of British Columbia, Vancouver, British Columbia, Canada, J.Mark FitzGerald, Clinical Epidemiology, Vancouver Coastal Health Research Institute, Vancouver, British Columbia, Canada, Tony R. Bai, Respiratory, Pacific Lung Health Center, UBC, Vancouver, British Columbia, Canada, Gillian Currie, Economics, University of Calgary, Calgary, Alberta, Canada

interviewed (response rate 40%) and 168 of these returned DCE questionnaires. A further 110 from the cold sample returned DCE questionnaires. The TTO model best predicted actual TTO values, followed closely by the rank model (mean absolute errors of 0.051 and 0.065 respectively). The DCE models from the warm and cold samples were very similar and predicted higher values for better states and lower values for worse states. CONCLUSIONS: Rank and DCE data offer promising alternatives to the more challenging valuation tasks of standard gamble and time trade-off.

AIMS: To assess management strategies, instruments used to measure the health-related quality of life (HRQL) of asthmatics must be able to discriminate between levels of control. The study objectives were to examine the cross-sectional construct validity of the HUI-3, SF-6D, and EQ-5D in terms of self-reported asthma control and severity, and to compare these with two valid and reliable disease-specific instruments, the standardized version of the Asthma Quality of Life Questionnaire (AQLQ(S)) and the Asthma Control Questionnaire (ACQ). METHODS: One hundred fifty-seven asthmatics (35.0+7.9 years of age) completed all five questionnaires and provided self-reported HRQL, asthma severity, and asthma control level. Each participant also underwent pulmonary function testing and provided information on mediation and healthcare use. Construct validity was assessed by comparing mean scores for each instrument across levels of self-reported asthma severity. Convergent validity of the instruments was evaluated using Spearman s correlations. RESULTS: Strong correlations were identified between the three generic instruments (rho = 0.61 to 0.72) and between the AQLQ(S) and ACQ (rho = 0.82). However, there were only weak correlations between the generic and disease-specific instruments (rho = 0.19 to 0.46). A monotonic decline was not observed in for the utilities with increasing severity or worsening control; further work is needed to determine the etiology of this trend. CONCLUSIONS: There was no consistent relationship between the health utilities derived from the generic, preference-based instruments and increasing asthma severity, asthma control, or asthma-specific quality of life. This could have a significant impact on the results of any costutility analysis in asthma.

1775 /Effects of Comorbidities on Preference-Based Scores: Results from the U.S. EQ-5D Valuation Study Lauren J. Lee, Michael B. Nichol, Pharmaceutical Economics and Policy, University of Southern California, Irvine, CA AIMS: Assessing the association between co-morbidity and preference is necessary because more than 50 million U.S. residents have more than one medical condition. METHODS: The U.S. EuroQOL (EQ-5D) Valuation Study was a large prospective study (n = 4,048). A multi-stage probability sampling method ensured that the study sample was generalizable. Participants self-reported medical conditions and completed the EQ-5D and Health Utilities Index Mark 2/3 questionnaires. Incremental disutilities were compared by subtracting scale scores from those with no disease to those with single diseases; single diseases to disease pairs; and no disease to disease pairs. Half standard deviations (SD) in preference scores were defined as clinically significant. Statistical significance was derived from Wilcoxon rank-sum test with Bonferroni adjustment. RESULTS: The mean age was 45 (SE = 0.5) years, 52% were female, and 35% reported no medical condition. The most commonly reported single disease and disease pair were hypertension (n = 148) and hypertension plus arthritis (n = 55). Marginal disutilities for single diseases to no disease were statistically significant for arthritis and depression across all 3 preference-based measures; and clinically significant for depression for the EQ-5D only (-0.085, p<0.0001). When marginal disutilities associated with disease pairs were compared to single diseases, adding arthritis or depression as a second disease produced more clinically and statistically significant disutilities than adding any other diseases. CONCLUSIONS: Depression and arthritis consistently produced clinically and statistically significant reductions in preference. Given that only a small number of subjects reported single diseases (n <148) and disease pairs (n <55) in the U.S. EQ-5D Valuation Study, additional analyses using a larger population-based sample is warranted.

1669 /Using Rank and Discrete Choice Data to Estimate Health State Utility Values for QALYS John E. Brazier, Health Economics and Decision Science, University of Sheffield, Sheffield, South Torkshire, UK, Yaling Yang, Health Economics and Decision Science, The University of Sheffield, Sheffield, South Yorkshire, UK, Aki Tsuchiya, Health Economics and Decision Science, University of Sheffield, Sheffield, South Yorkshire, UK

1819 /Assessing Differences in Utility Scores: A Comparison of 4 Widely Used Preference-Based Instruments Nan Luo, Centre for Health Services Research, National University of Singapore, Singapore, Hwee-Lin Wee, Rheumatology, Singapore General Hospital, Singapore, David Machin, Biostatistics Unit, National Cancer Centre Singapore, Singapore, Wai-Chiong Loke, SingHealth Polyclinics, Singapore, Singapore, Shu-Chuen Li, Pharmacy, National University of Singapore, Singapore, Yin-Bun Cheung, MRC Tropical Epidemiology Group, London School of Hygiene and Tropical Medicine, London, United Kingdoms, David Feeny, Institute of Health Economics, Edmonton, Alberta, Canada, Kok-Yong Fong, Julian Thumboo, Rheumatology, Singapore General Hospital, Singapore

AIMS: There is increasing interest in using ordinal methods for eliciting health state values. Previous work has been hindered by the problem of placing results on the full health (=1) and dead (=0) scale required for QALYs. This paper presents a solution to this problem in for ranking and discrete choice experiment (DCE) data. METHODS: A representative sample of the UK general population (n=308) were interviewed and asked to rank from best bto worst health states derived from a condition specific quality of life instrument (AQLQ5D). The TTO technique was then used to elicit values for each health state. Four weeks after the interview, consenting individuals received a postal questionnaire containing a DCE with 6 pair-wise comparisons of AQLQ-5D states and 2 of AQLQ-5D states against being dead. An additional sample was sent the same questionnaire without the interview beforehand. TTO data were analysed by random effects, rankings by rank ordered logit and DCE by random effect probit. All models assumed additivity between dimension levels, and included a dummy variable for being dead. Rank and DCE results are transformed onto the QALY scale by rescaling the coefficients by dividing the beta coefficients for each dimension level by the coefficient on being dead RESULTS: 308 individuals were

AIMS: To characterise differences in utility measurements (dUTY) among four commonly used preference-based HRQoL instruments, evaluate the potential impact of these differences on cost-utility analyses (CUA) and determine if sociodemographic/clinical factors influenced the magnitude of these differences. METHODS: Consenting Chinese, Malay and Indian subjects, age 21 years and above, recruited from a primary care facility in Singapore were

A-38

interviewed using Singapore English, Chinese, Malay or Tamil versions of the EQ-5D, HUI2and HUI3, SF-6D. Agreement between instruments was assessed using Bland-Altman (BA) plots. Changes in incremental cost-utility ratio (ICER) from dUTY were investigated using eight hypothetical CUA decision trees. The influence of sociodemographic/clinical factors on dUTY between instrument pairs was studied using multiple linear regression (MLR) models for English-speaking subjects (circumventing structural zero issues). RESULTS: In 668 subjects (median age 48 years, 59% female), median utility scores ranged from 0.796 (95% CI: 0.796, 0.848) for the EQ-5D to 0.886 (0.882, 0.892) for the SF-6D. BA plots: Mean differences (95% CI) exceeded the minimum clinically important difference of 0.04 for 4 of 6 pair-wise comparisons, with the exception of the HUI2/EQ-5D (0.029, CI: 0.017, 0.041) and SF6D/HUI2 (0.015, CI: 0.006, 0.023). Decision trees: The ICER ranged from $94,661/QALY (6.3% difference from base case) to $100,693/QALY (0.3% difference from base case). MLR: Chronic medical conditions, marital status and FFM scores significantly (p<0.05) influenced dUTY for several instrument pairs. CONCLUSIONS: Although clinically important differences in utility measurements were present for different preference-based instruments, the impact of these differences on CUA appeared negligible. The presence of chronic medical conditions, marital status and FFM influenced the magnitude of these differences.

AIMS: To assess the impact of chronical health conditions and acute restriction in daily life on the different domains of adolescents HRQoL in a national representative sample of German adolescents. METHODS: Adolescents aged 11-17 (N=1700) of a national representative sample from the German Health Survey for Children and Adolescents were classified as having a chronical health condition or not according to parent data. Furthermore a report on sick days (missing school or being confined to bed) in the past month was considered. Self-reported HRQoL was measured using the KIDSCREEN-27 instrument. Mean scores were compared using ttests. Effect size (ES) was assessed by dividing mean difference with standard deviation. RESULTS: About 12.2% of the adolescents were classified to have a chronic condition or disability. 15.1% of the children missed 1-3 days 2.8% 4-6 days and 1.8% >7 school days in the last month. 16.7% of the children were confined to bed for 1-3 days, 2.9% for 4-6 days and 1.1% for >7 days. Children who suffered from a chronic condition displayed statistically significant lower HRQoL in the KIDSCREEN-27. This difference was especially pronounced on the dimensions physical- (ES=.43, p<.01) and psychological wellbeing (ES=.37, p<.01). With an increasing number of missed school days or bedridden days HRQoL on all subdimensions declined constantly. The HRQoL scores declined with ES=0.16-0.38 (0.15-0.45) for children with 1-3 missed school days (bedridden days), ES=0.43-1.07 (0.39-1.05) in children with 4-6 missed school days (bedridden days) and with ES=0.49-1.11 (0.531.42) for those >7 days absent from school (bedridden days). CONCLUSIONS: The KIDSCREEN-27 proved to be a valid measure to detect the theoretical expected impact of chronic conditions and restrictions in daily life activities on adolescents HRQoL. It was discussed which mechanism affected the physical as well as the psychosocial aspects of HRQoL.

Quality of Life in Children 2 1490 Health-Related Quality of Life in Children Following Completion of Cancer Treatment Heleen Maurice-Stam, Martha A. Grootenhuis, Bob F. Last, Psychosocial Department, Emma Children's Hospital Academic Medical Center, Amsterdam, The Netherlands

1365 /A Retrospective Cohort Study of Quality of Life, Pulmonary Function, and Radiographic Measures in Children with Congenital Scoliosis: an Evaluation of Patient Outcomes after Traditional Spine Surgery Hiroko Matsumoto, Pediatric Orthopaedic Surgery, Columbia University, New York, NY, Michael G. Vitale, Pediatric Orthopaedic Surgery, Michael R. Bye, Pediatric Pulmonology, Columbia University, New York, New York, Joshua E. Hyman, Pediatric Orthopaedic Surgery, Columbia University, New York, NY, Jaime A. Gomez, Whitney A. Booker, David P. Roye, Jr., Pediatric Orthopaedic Surgery, Columbia University, New York, New York

AIMS: Completing therapy is one of the major transitions in care in the practice of paediatric oncology and therefore deserves special consideration. The purpose of the study was to assess Health-related Quality of life (HRQoL) in paediatric patients shortly after the end of successful treatment. METHODS: HRQoL of patients aged 1-15 years was investigated on average two months after the end of successful treatment. A total of 54 parents of patients aged 1-5 years completed the TNO-AZL Preschool Quality of life Questionnaire. The TNO-AZL Children`s Quality of life Questionnaire was completed by 18 parents of patients aged 6-7 years, and by 52 patients aged 8-15 years. HRQoL in patients was compared with aged-matched and sex-matched norm data from the general Dutch population by means of one sample t-tests, one sample sign-tests or binomial tests. Effectsizes were calculated according to Cohen. RESULTS: All age groups, except patients aged 8-11 years, experienced significantly worse (p<0.005) HRQoL than the norm with respect to motor functioning. In addition, preschool patients were rated significantly worse (p<0.005) on Sleeping, Appetite, Stomach, Skin, Problem Behaviour, Anxiety and Liveliness, and patients aged 6-7 years on Autonomy and Cognitive functioning. With the exception of Skin all effectsizes were large (d >0.8). CONCLUSIONS: A few months after the end of successful cancer treatment, patients appeared to experience worse HRQoL than the norm to a clinically relevant extent. Supporting patients and parents should not stop when treatment ends.

AIMS: Children with congenital scoliosis are at an increased risk for severe progression of their spinal curvature, which will eventually impact their quality of life (QOL) and pulmonary function (PF). Recently, new surgical options have become available to treat these conditions. It is important to evaluate the outcomes of children who were treated with traditional methods before any comparative studies can be conducted. The purpose of this study is to evaluate QOL and PF of children who were treated with traditional method for progressive congenital scoliosis and to compare them to those of healthy children. METHODS: Twenty four patients (average 14± 5 years) with a diagnosis of congenital scoliosis who were treated with traditional surgery were evaluated using PF testing and QOL surveys using Child Health Questionnaire Parent Form (CHQ-PF). The CHQ scores and PF were compared to those of healthy children using onesample t tests. Pearson correlation was also examined to see if there were any relationships between PF and CHQ scores. RESULTS: Forced vital capacity (FVC), forced expiratory volume in one second (FEV1), total lung capacity, and vital capacity were significantly lower than those in healthy children (p<0.01). The average percent predicted FVC and FEV1 were 68.3% and 67.6% respectively. CHQ scores in our study patients were significantly lower than healthy children in physical function, general health, and physical summary

1621 /Chronic Health Condition, Restricted Participation in School and HRQOL: Results from the German Health Survey for Children and Adolescents Nora Wille, Michael Erhart, Jennifer Nickel, Ulrike Ravens-Sieberer, Child and Adolescent Health, Robert Koch-Institut, Berlin, Germany

A-39

and significantly higher in bodily pain and impact on parent emotion (p<0.05). The FVC and bodily pain score had a positive correlation (p=.049). CONCLUSIONS: Compared with healthy peers, patients with congenital scoliosis treated with traditional surgery have significantly worse QOL scores and PF. Children who had lower FVC had higher bodily pain. Current efforts are underway to ascertain the relative effects of spinal deformity and characteristics of the surgery on QOL and PF in this population.

four forms to a U.S. general population sample and supplemental disabled and chronic condition samples of caregivers for children 6 months- 17 years. Participants recruited through an online panel were randomly assigned to internet (N= 3,856) or mail administration mode (N=2,197). Model requirements of unidimensionality and local dependence were evaluated by categorical CFAs for the total sample and within respondent subgroups (age, mode, disability status). Item response curves were evaluated with non-parametric techniques. IRT methods will be used to link items across forms and develop normcentered item calibrations. Item calibrations and information functions will be reviewed to determine the final item pool. RESULTS: Many PF items were found to fit sub-domains of movement, personal care/ dexterity, and toileting skills. Significant ceiling effects were identified; the top response option was selected by at least 80% of respondents in 64% of field test items. In the final data collection, a subset of items met IRT model requirements and will be calibrated. CONCLUSIONS: Calibrated item banks allow IRT-based scores to be estimated from any set of items, so those most relevant for a particular child can be administered and children across a range of functional ability can be scored on the same continuum. Including items from current surveys in a bank allows IRT-based functioning scores to be reported in original scale metrics and results compared across different scales.

1557 /Social Determinants of Quality of Life in Polish Children on the Basis of KIDSCREEN Study Joanna Mazur, Hanna Kololo, Epidemiology, Agnieszka Malkowska, School Medicine, Ewa Mierzejewska, Epidemiology, Institute of Mother and Child, Warsaw, Poland AIMS: To analyze health related-quality of life (HRQL) in Polish children in broader social context and to test the hypothesis that HRQL depends more on social relations than on family wealth. METHODS: Analysis is based on data from representative sample of 1260 children aged 8-16 years and their mothers, participating in mail survey conducted in Poland in 2003. HRQL was measured using KIDSCREEN-27 version (Ravens-Sieberer et al., 2001) indexed in 5 dimensions. International T-values, developed with Rasch model, less than 40 were considered as low HRQL. Parental education level, Family Affluence Scale (Currie, 1998), Oslo Social Support Scale (Dalgard, 1996), quality of family relations (adapted from Weissman Social Adjustment Scale, 1996), parental tobacco smoking and alcohol drinking were analyzed as social determinants obtained mostly from parents form. Five stepwise logistic regression models were estimated using SPSS v.12. to asses the probability of normal HRQL scores. All models were adjusted for age (continuous), gender and domicile. RESULTS: Final models specific for five HRQL dimensions included 4 or 5 from 10 potential variables. Father education, domicile and tobacco smoking were excluded from all models. Quality of family relations remained significant in all models, and odds ratio ranged from 2.16 to 6.91 (dimension: Relations with Parents). Moreover perceived social support remained significant in all models, and odds ratio ranged from 1.89 to 2.75 (dimension: Psychological Well-being). Alcohol problems in family were recognized as independent predictor of low HRQL in dimensions concerning psychological well-being and school environment. Although crude HRQL score are higher in boys and younger children, age and gender appeared to be insignificant after adjusting for social factors, especially in the dimensions of parents and home life. CONCLUSIONS: Factors related to social relations and parental risk behaviors are stronger predictors of HRQL than demographic characteristics and traditional variables describing socio-economic status.

Understanding Pain and QOL: What is the Latest Progress? 1138 /The ICF Core Sets for Low Back Pain: Do They Tell Us What Matters to Patients ? Ricky Mullis, Julie Barber, Elaine M. Hay, Primary Care Musculoskeletal Research Centre, Keele University, Keele, Staffordshire, UK AIMS: The International Classification of Functioning, Disability and Health (ICF) provides a framework for the description of healthrelated states through domains classified from body, individual and societal perspectives. The Core Sets were developed by consensus to provide a link between the salient ICF categories and specific conditions or diseases. The ICF core sets for low back pain (comprehensive and brief) were intended to define the typical spectrum of problems in LBP patients and to enable the rating of patients included in clinical studies. The aim of this work was to establish the extent to which the Core Sets encompass key items identified by LBP patients as relevant to the problems they encounter. METHODS: 406 patients recruited into a LBP treatment trial (Hay et al. 2005) were asked to identify : 1) a specific activity or task made most difficult by their back pain, 2) one thing they really enjoy doing usually but are unable to do because of back pain. Responses were recorded as free text. The key components were extracted, and mapped onto the ICF by two independent researchers. The response categories were then checked for inclusion within the ICF Core Sets. RESULTS: On Q1, 397 patients were able to identify a most difficult activity. Of these, 395 (99%) were contained within the Core Set for LBP, and 329 (81%) within the Brief Core Set. Activities and Participation accounted for 385 (95%) items. On Q2, 312 patients were able to identify a usually enjoyed activity. Of these, 311 (99%)were contained within the Core Set, but only 54 (17%) fell within the Brief Core Set. CONCLUSIONS: Tasks identified by patients as "most difficult" had a high level of inclusion within the Core Sets, which is in keeping with the aims of the ICF. However, in Q2 most items were not included within the Brief Core Set for LBP. These data were drawn from participants in a clinical study i.e. the very people for whom the Brief Core Set was devised. This limitation could be significantly improved by inclusion of a single additional item (d920 - recreation and leisure).

1800 /Development of an Item Bank for Assessing Physical Functioning in Children Anastasia E. Raczek, QualityMetric Incorporated, Health Assessment Lab, Waltham, MA, Renee N. Saris-Baglama, Michael A. DeRosa, QualityMetric Incorporated, Waltham, MA, John E. Ware, Jr., QualityMetric Incorporated, Health Assessment Lab, Waltham, MA AIMS: To develop a calibrated Physical Functioning (PF) item bank for children. METHODS: Preliminary studies: 1) Literature review and content analysis of items from 20 PF scales. 2)Field test with 100 parents (54 children with disabilities, 46 without) completing a survey with 66 existing and newly-drafted PF items. Descriptive statistics and categorical confirmatory factor analyses (CFA) were conducted to evaluate item characteristics. 3) Thematic content analysis of cognitive interviews with clinicians (N=5) and parents (N=10) to evaluate item wording and content. Large-scale data collection: 253 existing and new PF items were administered across

A-40

relationship between chronic pain and QOL considerably (from std. beta 0.21 to 0.08, p<0.001), while stress-related symptoms showed a strong relationship with QOL (std.beta -0.54, p<0.001). The addition of stress-related symptoms into the equation increased the explained variance (Adj R Square) from 10% to 35%. The effect of chronic pain on QOL disappeared when stress-related symptoms were controlled for in the analyses. No interaction effects were found between sociodemographics and stress-related symptoms in relation to QOL. CONCLUSIONS: These results indicate that stress-related symptoms have a mediator effect on the relationship between chronic pain and QOL.

1094 /Measuring Quality of Life in Patients with Chronic Pain Disorders – Recurrent Versus Stable Conditions Wolfgang Weidenhammer, Centre for Complementary Medicine Research, Technical University of Munich, Munich, Germany AIMS: To investigate the impact of different pain conditions with respect to intermittent or permanent occurrence on assessment of health status and quality of life. Is referring to the previous 4 weeks period a reliable and valid approach in patients suffering from clearly definable pain attacks? METHODS: Data on intensity of pain (VAS) and quality of life (SF-36) were collected from out-patients with chronic headache, low back pain and osteoarthritic pain, undergoing a special research program on acupuncture in Germany. Patients with recurrent (10 or less days with pain per 4 weeks) versus stable (20 or more days with pain) pain conditions were compared regarding distributions for graded pain intensity and for SF-36 scores, including subdimension Bodily Pain (traditional and norm based scaling). Furthermore, correlations between scores before and after treatment were analyzed. RESULTS: 1,853 patients with recurrent (86% female; age 46±14 yrs) and 1,533 patients with stable pain conditions (78% female; age 61±13 yrs) were compared. Both groups differed markedly with regard to diagnosis. Patients with permanent pain showed smaller variance in estimating average and maximum pain intensity during the previous 4 weeks but not with regard to minimal pain level. Differences in variability were greatest for subdimensions Bodily Pain (recurrent pain: SD=8.3; permanent pain: SD=5.8) and Role Physical (recurrent pain: SD=14.0; permanent pain: SD=10.5) using SF-36 standard scores. Test-retest correlations (after treatment) were smaller in patients with recurrent pain (r=.16 to .27) compared to patients with permanent pain (r=.36 to .50). CONCLUSIONS: The differences between both patient subgroups have impact on the evaluation of treatment effectiveness and support the demand for bigger sample sizes when looking at patients with recurrent pain. It remains unclear whether the findings truly reflect different pain conditions, or are due to biased measurement. Instructions to patients should clearly address the problem of different patterns of occurrence of pain.

1711 /Mapping the MOS Sleep Problems Index-II to the SF-36 Health State Utility Index (SF-6D) Mark R. Kosinski, Lincoln, Rhode Island, Dominique Dubois, Health Economics and Pricing, Janssen Pharmaceutica, Beerse, Belgium, Xiaowu Sun, QualityMetric Incorporated, Lincoln, Rhode Island AIMS: Deriving preference scores for the Medical Outcomes Study (MOS) Sleep Problems Index-II (SLPI) would enable its use in cost effectiveness analyses. The aim of this study was to map scores of the SLPI to a preference-based health-state utility index (SF-6D) scored from the SF-36 Health Survey. METHODS: The SF-6D was regressed onto the SLPI in a random half sample (developmental sample) of chronically ill patients from the MOS (N=1,413). The best fitting regression model was then applied to the other random half sample (cross-validation sample) of chronically ill patients from the MOS (N=1,412) and a sample (N=199) of chronic low back pain patients participating in a 7-week, open-label, non-comparative, repeated-dose, trial of once-daily OROS® hydromorphone who completed both the SF-36 and SLPI. RESULTS: The best fitting multiple regression model included a quadratic term for the SLPI and explained approximately 34% of the variance in SF-6D scores based on the MOS developmental sample. Prediction errors were greatest at higher SLPI score levels (more sleep problems). Addition of other regression covariates such as demographic and disease status variables to the model explained minimal incremental amounts of variance (<4%) in SF-6D scores. These results were replicated in the MOS cross-validation sample. In both developmental and crossvalidation MOS samples mean predicted and observed SF-6D scores were nearly identical. In the trial sample, the correlation between estimated and observed SF-6D scores was 0.49 (p<0.001). Mean predicted SF-6D scores were 0.09 points higher than observed SF-6D scores at both baseline and final visits of the trial while the change in predicted and observed SF-6D scores were identical. CONCLUSIONS: Results indicate that it is possible to map the SLPI to the SF-6D yielding useable preference-based scores essential for cost-effectiveness analyses, with some caution for those scoring at the highest SLPI score levels.

1556 /Relationships Between Chronic Pain and Stress-Related Symptoms and Quality of Life (QOL) Within the General Norwegian Population Astrid K. Wahl, Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway, Tone Rustøen, Faculty for Nursing Education, Oslo University College, Oslo, Norway, Berit R. Hanestad, Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway, Anners Lerdal, Faculty of Health, Buskerud University College, Drammen, Norway, Øistein Knudsen jr, Directorate for civil protection, Tønsberg, Norway, Christine Miaskowski, Department of Physiological Nursing, University of California, San Francisco, CA

1679 /Evaluating Methods for Linking Pain Items from Two Studies Using Item Response Theory Analysis Wen-Hung Chen, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Jin-Shei Lai, Ceneter on Outcomes, Research and Education, Northwestern University, Evanston, IL, Karon F. Cook, Department of Rehabilitation Medicine, University of Washington, Houston, TX, Dagmar Amtmann, Department of Rehabilitation Medicine, University of Washington, Seattle, WA, Dennis A. Revicki, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD

AIMS: The aim of this study was to evaluate the relationships between chronic pain and stress-related symptoms and QOL . METHODS: A total of 4000 Norwegian citizens, aged 19 to 81 years, were randomly drawn from the National Register by Statistics Norway and sent a questionnaire packet. A total of 1912 questionnaires were returned after the two mailings and1893 had complete data (response rate 48.5%). The PTSS-10 was used to assess stress-related symptoms. QOL was measured using the QOLS -N. After controlling for age, sex, education, and marital status, multiple linear regression analyses were performed and possible interaction effects was assessed with general linear model univariate analyses of variance. RESULTS: Sociodemographic characteristics explained 5% of the variance (Adj. R Square) in QOL. Adding chronic pain to the equation increased the explained variance to 10%. Entering level of stress-related symptoms in the equation, reduced the

AIMS: Development of comprehensive item banks for pain and other domains require approaches to link different sets of items administered to different samples. This study evaluates two methods for linking items from two pain surveys using item response theory (IRT) analysis to form several item banks with a common

A-41

measurement scale. METHODS: We completed a secondary analysis of two independent surveys: IMMPACT Main Survey (959 chronic pain patients; 43 pain items) and Module 2 (N=148; 37 additional pain items); and NU-CORE Survey (402 cancer patients; 61 pain items). There were 19 pain interference items common across the IMMPACT and NU-CORE data sets, and the 37 Module 2 items were included in NU-CORE. Two linking methods were examined, the simultaneous (or joint) IRT resulting in same measurement scale and separated IRT calibration of the two data sets, requiring scale transformation. RESULTS: The simultaneous IRT calibration approach worked best and produced stable parameter estimates across the two sets of pain items based on the common pain interference items. The separated IRT approach produced item parameter estimates that were less consistent or stable across the two samples. The correlation between estimated Theta scores based on different methods was 0.99 and correlation between slopes was 0.92. For pain quality, the separated calibration was not supported due to the smaller number of subjects completing the common items. Simultaneous calibration yielded more stable results. CONCLUSIONS: The simultaneous IRT calibration method appeared to produce the best item parameters across independent samples of pain interference and quality items. If simultaneous calibration is not applicable, separated calibration with scale transformation may produce similar results if there are sufficient number of common items and large sample sizes. The simultaneous approach for linking across independent samples is recommended for developing comprehensive item banks.

baseline for most scales. The RT plus PCV chemotherapy arm showed a significant and clinically meaningful increase in both fatigue (p=0.0016) and a decrease in physical functioning (p=0.0063) during chemotherapy, compared to the RT alone arm. RT plus PCV chemotherapy also led to an increase in nausea/vomiting, during and shortly after chemotherapy but this was not clinically meaningful. The long term results (after one year) showed no difference between arms. CONCLUSIONS: This study shows that, when compared to RT alone, RT plus adjuvant PCV chemotherapy increases progression-free survival, but leads to a short-term negative impact in two (fatigue and physical functioning) of the seven selected HRQOL endpoints. However, longer term results show both groups to have comparable HRQOL. 1706 /Impact on Health-Related Quality of Life (HRQL) of Radical Prostatectomy and Prostate Brachytherapy in Patients with Low-Risk Clinically Localized Prostate Cancer Montse Ferrer, Health Services Research Unit, Institut Municipal d'Investigació Mèdica (IMIM-IMAS), Barcelona, Spain, Ferran Aguiló, Servei d'Urologia, C.S.U. Bellvitge, L'Hospitalet de Llobregat, Barcelona, Spain, Olatz Garin, Àngels Pont, Health Services Research Unit, Institut Municipal d'Investigació Mèdica (IMIM-IMAS), Barcelona, Spain, Pablo Fernández, Oncología Radioterápida, Instituto Oncológico de Guipuzkoa, San Sebastián, Spain, Alfonso Mariño, Oncología Radioterápica, Centro Oncológico de Galicia, A Coruña, Spain, Asunción Hervás, Hospital Ramón y Cajal, Madrid, Spain, Humberto Villavicencio, Urologia, Fundació Puigvert, Barcelona, Spain, Ismael Herruzo, Serv. Oncología Radioterápica, Hospital Regional Carlos Haya, Málaga, Spain, Jordi Craven-Bartle, Oncologia Radioterápica, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain

Cancer Studies and Approaches in QOL 2 1421 /An International EORTC Phase III Randomized Controlled Trial on the Role of Adjuvant Chemotherapy in Newly Diagnosed Anaplastic Oligodendrogliomas: HealthRelated Quality of Life and Symptom Results M. Taphoorn, Neurology, Medical Center Haaglanden, The Hague, The Netherlands, M. Van Den Bent, Oncology, Erasmus University Medical Center, Rotterdam, The Netherlands, M. Mauer, C. Coens, Quality of Life Unit, EORTC Data Center, Brussels, Belgium, J.Y. Delattre, Neurology, CHU Pitie-Salpetriere, Paris, France, A. Brandes, Oncology, Ospedale Bellaria, Bologna, Italy, C. Van der Rijt, Neurology, Erasmus University Medical Center, Rotterdam, The Netherlands, H. Bernsen, Neurology, Canisius Wilhelmina Hospital, Nijmegen, The Netherlands, M. Frenay, Medicine, Center Antoine Lacassagne, Nice, France, C. Tijssen, Neurology, St. Elisabeth Hospital, Tilburg, The Netherlands, D. Lacombe, A. Allgeier, Brain Unit, A. Bottomley, Quality of Life Unit, EORTC Data Center, Brussels, Belgium

AIMS: Prostate cancer is one of the most common in men. Results on cancer control with radical prostatectomy (RP) and brachytherapy (BT) that are the principal treatment options are similar, but there is no consensus on which is the treatment of choice. Our objective was to compare HRQL between RP and BT up to one year after treatment. METHODS: Longitudinal prospective comparative study with treatment decided by consensus between the patient and his physician. The study included 514 patients recruited at 8 Spanish centres. HRQL data was gathered using computer-assisted telephone administration, pre and post-treatment (months 1,3,6, and 12). Generic HRQL questionnaires administered were SF36 and FACT. Specific instruments included were: International Index of Erectile Function (IIEF), and the Expanded Prostate Cancer Index Composite (EPIC), which were specifically designed to measure the impact of treatments for localized prostate cancer. HRQL data was obtained from 394 patients (76% of participants): 255 received BT and 139 RP. RESULTS: Mean age was higher in patients treated with BT (65 vs 67 years, p<0.001). Both treatment groups showed very similar HRQL before treatment. One month after treatment, all HRQL questionnaires showed deterioration in both groups, but RP patients presented statistically worse outcomes in almost all the dimensions of the SF-36; FACT-P; EPIC Urinary and Sexual; and some IIEF scores. Generic questionnaires showed a recovery in HRQL between months 3 and 6, and differences between treatment groups disappeared. However, on the specific questionnaires, these differences remained until one year after treatment: EPIC urinary (73 vs 77, p=0.006), EPIC sexual (34 vs 48, p<0.001), and IIEF Sexual overall satisfaction (5.9 vs 7.3, p<0.001). CONCLUSIONS: This finding supports BT since it shows a lower impact on HRQL of patients compared with surgery. Funded by: DURSI-GENCAT (2005-SGR-00491), AATRM 086/24/2000, FIS PI020668 and IRYSS Network (G03/202 and FIS PI052403).

AIMS: Little is known about the Health-related Quality of Life (HRQOL) of patients treated for newly diagnosed anaplastic oligodendrogliomas. We investigated the impact of adjuvant procarbazine, CCNU, vincristine (PCV) chemotherapy following radiotherapy (RT) compared to RT alone on both short and long term HRQOL. METHODS: Adult patients with anaplastic oligodendrogliomas, were randomly assigned to receive RT alone or RT plus PCV chemotherapy. HRQOL was assessed with the EORTC QLQ-C30 and BN20. Seven pre-specified HRQOL endpoints were selected: Global HRQOL, fatigue, nausea/vomiting, physical and social functioning, communication deficit and seizures. Assessments were performed at randomization, at the end of RT, then every three months for the first year following RT and then every six months until progression. RESULTS: 368 patients were randomized, with 58% males, median age 49 years. Compliance with HRQOL questionnaires was 78% at baseline, and reasonable (55-71%) for each assessment up to 2.5 years post RT. Initial analysis indicated both groups showed substantial (but comparable) impairment at

A-42

Measures were: Dyadic Adjustment Scale; MOS ‘Social Support Survey; Sexual Adjustment Questionnaire. Analysis used paired ttests and analysis of variance. All findings below are significant at p<.05 or less. RESULTS: Dyadic adjustment was lower than norms for both survivors and spouses. Compared to wives, survivors reported more social support, confided more in their spouses, and did more to maintain their marriages. Survivors also had less regret and frustration about their marriage, but were less sexually satisfied. No ethnic differences were found for sexual adjustment; however, compared to API survivors, Caucasian survivors had better dyadic adjustment, were more satisfied with their relationship, and perceived more social support. The worst adjustment was found in couples where the survivor was API and the spouse was Caucasian. No differences were found according to treatment. CONCLUSIONS: Overall dyadic adjustment was negatively impacted by prostate cancer, with wives affected more than husbands. Cultural factors appear to influence adjustment, with API men having lower ratings, particularly if their wives are of a different ethnic background. These findings suggest that supportive interventions are particularly needed for spouses of these cancer survivors, and that cultural expectations and values may affect long-term adjustment to prostate cancer and marital relationships.

1566 /Patient Reported Outcomes (PROs) Support Change in Practice in U.S. Palliative Radiation Therapy for Osseous Metastases (RTOG 9714) Deborah W. Bruner, Nursing, University of Pennsylvania, Philadelphia, PA, Jennifer James, Statistics, Radiation Therapy Oncology Group, Philadelphia, PA, William Hartsell, Radiation Therapy, Advocate Good Samaritan Cancer Center, Downers Grove, IL, Andre Konski, Radiation Therapy, Fox Chase Cancer Center, Phila, PA, Walter Curran, Radiation Oncololgy, Thomas Jefferson University Hospital, Philadelphia, PA, Mack Roach, Radiation Therapy, University of California-San Francisco, San Francisco, CA, Desiree Doncals, Radiation Therapy, Akron City Hosp, Akron, OH, Benjamin Movsas, Radiation Therapy, Henry Ford Hospital, Detroit, MI, Ding-Jen Lee, Radiation Thearpy, Johns Hopkins Med. Ctr, Baltimore, PA, Charles Scarantinio, Radiation Therapy, Rex Healthcare, Raleigh, NC AIMS: To determine whether there is a difference in health-related quality of life(HRQL) for 8Gy in a single fraction as compared to 30Gy in 10 fractions for patients(pts) with painful bone metastases from breast or prostate cancers. METHODS: Primary results of RTOG9714 are reported elsewhere and indicate that a single fraction is equivalent to multiple fractions in terms of pain control. European studies have demonstrated similar results but have not previously changed practice in the US because US physicians reportedly voiced concerns over less than optimal measures of HRQL in previous trials. RTOG9714 took these concerns into account and PROs were measured using Functional Assessment of Cancer TherapyGeneral(FACT-G) to assess HRQL and Brief Pain Inventory(BPI) to assess pain. Pts were stratified according to treatment arm, gender, and marital status. Area under the curve(AUC) analysis was used to compare overall HRQL and pain for the first 3 mos following treatment. The Wilcoxon rank test was also used to test the 3 mo change score differences for each PRO measure. Pts must have completed the pretreatment and the 3 mo follow-up assessment to be included for analysis. RESULTS: Of the 907 pts randomized to RTOG9714, 351(39%) completed the FACT-G and 380(42%) completed the BPI assessments at both timepoints. Pts included in the analysis were more likely to have a partner(p=0.04), be white(p=0.01), and had higher baseline KPS(p<0.01) than those with missing data. Pretreatment characteristics between analyzable pts on the 8Gy and 30Gy arms did not differ. There was no difference between the 3 mo. and baseline assessments in pain control or HRQL by treatment arm or marital status. There were statistically significant differences in HRQL by gender. CONCLUSIONS: Based on this data, there is no difference in overall HRQL between the two treatment arms from baseline to 3 mos. further supporting the use of a single fraction for painful bone metastasis. This is a case study of PROs providing evidence for change in clinical practice.

1312 /Prospective Evaluation of Quality of Life in Prostate Cancer Patients and Their Spouses Beate Bestmann, Reference Center Quality of Life in Oncology, University Hospital Kiel, Kiel, Germany, Volker Rohde, Wolfgang Weidner, Urology, University of Giessen, Giessen, Germany, Thomas Kuechler, Reference Center Quality of Life in Oncology, University Hospital Kiel, Kiel, Germany AIMS: Predominant symptoms in patients receiving treatment for prostate cancer are erectile dysfunctions and urinary problems. These symptoms may impair the patients overall QoL but especially their sexual life and their spouses sexual life as well. Therefore we conducted a study on QoL of as well prostate cancer patients as their spouses. METHODS: For assessment of QoL in prostate cancer patients, the EORTC QLQ-C30 plus a validated, prostate-specific module were used. The spouses' questionnaires included the SF-12 plus the global health scale from the EORTC QLQ-C30. All questionnaires were administered prospectively before primary treatment, 3-, 6-, 12- and 24 months after primary treatment. QoL data of 439 patients and 244 spouses were included the analysis. Additionally, QoL scores from a reference sample of 1,185 male persons without prostate cancer were included. RESULTS: In terms of general QoL, prostate cancer patients showed surprisingly good scores that were comparable to the reference sample. All QoL scores remained stable during the time observed. The same trend was found for the scales 'nutrition', 'heat' and 'psychic strain' on the prostatespecific module. Nevertheless, a significant increase of symptoms was found on the module scales dealing with sexual functioning ('erectile dysfunction', 'sexual problems', 'problems with partner'). While sexual problems and problems with partner recover over time, the impairment in terms of erectile dysfunction remained stable. The spouses however, reported constantly high QoL scores. CONCLUSIONS: The results of this study show that treatment for prostate cancer may have tremendous impact on the patients' sexual functioning and their sexual relationship/ partnership. This information should be included in pre-operative counseling. The fact that no impairment of treatment was found on the EORTC QLQ-C30 but only on the module underlines the importance of disease-specific QoL measures for prostate cancer patients.

1793 /Dyadic Adjustment in Multiethnic Prostate Cancer Survivors and Their Spouses Carolyn C. Gotay, Yen-Chi Le, Cancer Research Center of Hawaii, University of Hawaii, Honolulu, Hawaii AIMS: Prostate cancer survivorship confers many challenges, but few studies have focused on long-term marital relationships or nonCaucasians. We examined social support, sexual adjustment, and dyadic adjustment in multiethnic prostate cancer survivors and spouses. METHODS: A population-based tumor registry identified 5-11 year survivors of localized prostate cancer and their spouses (N=190 couples). Treatments were ‘watchful waiting’ (n=17), surgery (n=84), or radiation (n=80). 75% of patients and spouses were Asian/Pacific Islanders (APIs), and 25% Caucasian. Ethnic groups did not vary in age, treatment, or time since diagnosis. Data were collected by mailed questionnaires (response rate 51.4%).

A-43

established and over time. Ten THC patients were interviewed at stages of the THC experience. Interviews were recorded/transcribed. Data were analyzed using classical methods described by Glaser and Strauss (1967). Conceptual categories and their properties were identified from substantive coding using the constant comparative method. Analysis of quantitative data from the SF36 is being conducted and will be reported. RESULTS: Quality of life was maintained. Participants credited THC with preventing disease exacerbation and said ongoing connection to providers furthered understanding/management of their illness. The over-arching process of their THC experience was that they were: Living in a Comfort Zone. The 4 major categories sequenced into overlapping phases were: Making a connection, Focusing on vital signs, Understanding illness, and Taking control. In addition, as a major innovation, THC implementation is more about change management than technology. Providers and patients grasped technology quickly and reported few problems. CONCLUSIONS: THC has the potential to promote and maintain quality of life for community-dwelling adults with chronic illness. The cost effectiveness of THC delivery is yet to be determined.

Cardiovascular Disease and QOL 1484/ Underlying Factor Structure of the Minnesota Living with Heart Failure Questionnaire (MLHFQ) in Eight Country Versions Olatz Garin, Montserrat Martinez-Alonso, Montse Ferrer, Angels Pont, Health Services Research Unit, Institut Municipal d'Investigació Mèdica, Barcelona, Spain, Eric Van Ganse, Faculte de Medecine Lyon Sud, Lyon, France, Ingela Wiklund, Outcomes Research, AstraZeneca R&D, Moelndal, Sweden, Jordi Alonso, Health Services Research Unit, Institut Municipal d'Investigació Mèdica, Barcelona, Spain AIMS: The MLHFQ is the most widely used instrument to assess quality of life in patients with heart failure. The purpose of this study is to confirm the structure of 2 dimensions (physical and emotional, 8 and 5 items) and evaluate their metric properties in 8 country versions METHODS: Data of 5 studies including 2587 patients from Canada, Brazil, Sweden, Poland, Italy, Great Britain, France and Germany was pooled in a database. Confirmatory Factor Analyses (CFA) (weighted least squares method over the polychoric correlations) was conducted. To evaluate the fit of the underlying structure of the MLHFQ versions with the postulated two-factor model the following indices were used: comparative fit index (CFI), non-normed fit index (NNIF), and Root Mean Square Error of Approximation (RMSEA). The factor loading invariance across countries was assed with the Chi-square difference test. Reliability and validity across versions were also evaluated to confirm the scales' metric properties RESULTS: The CFA within each country provided significant factor loadings for all the items. The goodness of fit CFI and NNFI indices were acceptable (>0.90). In contrast, the Root Mean Square Error was only acceptable for Canada and Germany, ranging within 0.066-0.146 for the rest of countries. The factor loading equality across the 8 countries was significant (p < 0.0001). Regarding the dimensions' metric properties, Cronbach's alpha coefficient was above 0.7 for all countries and scales. No difference was shown across countries in the scale score means, and mean differences between patients in NYHA class I-II and III-IV followed the expected pattern CONCLUSIONS: The CFA shows that although factor loadings were statistically significant in all MLHFQ versions, the fit of the model was only completely adequate for Canada and Germany. In spite of the excellent reliability and validity showed by scales across versions, further research on the MLHFQ structure is needed. (Funding: EC ALRI-CT-2000-00551; SAF2002-11009-E; DURSI-GENCAT 2005-SGR-00491)

1646 /The Austrian Web - Based Outpatient Cardiac Rehabilitation Quality Assessment Project: Patient - Rated Outcomes Stefan Hoefer, Medical Psychology and Psychotherapy, Medical University Innsbruck, Innsbruck, Austria, W Benzer, A Philippi, K Mayr, H Ocenasek, H Harpf, R Pokan, Austrian Working Group of Outpatient Card. Rehab., Feldkirch, Austria AIMS: Health-related quality of life (HRQL) is an important outcomes parameter in cardiac rehabilitation (CR). However paper versions of instruments are time and resources consuming. We therefore developed a health outcome measurement system for cardiology (HOM-C) using the internet and assessed its feasibility and validity. METHODS: Methods: We included three wellestablished self-administered instruments, the Short-Form 36 (SF36), the MacNew Heart Disease Health-related Quality of Life Questionnaire (MacNew), and the Hospital Anxiety and Depression Scale (HADS), in the HOM-C. A computer interface allows patients to self-assess their current health status by using a touch screen or computer mouse. We implemented the HOM-C in a pilot phase in two outpatient CR centres to assess both its feasibility and the validity of the software. RESULTS: The HOM-C was used by 131 patients referred for CR (76.2% male, mean age 54.1 years). Complete baseline data was available for 99% of the patients. Factor analysis confirmed the three factor structure of the MacNew (physical, emotional and social) and explained 67.4% of variance. Baseline levels for the MacNew were relatively high and ranged from 5.5 ± 1.2 to 5.8 ± 1.2. Seven factors explained 73.7% variance for the SF-36. With 22.8% of the 131 patients screened positively for anxiety and 14.1% for depression, 73.4% of variance was explained by only one factor for the HADS. CONCLUSIONS: Among the computerized versions of the three measures, the MacNew performed most consistently with previous results. The HOM-C provides an easy, useful and cost effective way of implementing standardised HRQL measures in CR.

1526 /Telehomecare Maintains Quality of Life for CommunityDwelling Adtuls with Chronic Illness: Preliminary Research Janet E. Jeffrey, Diane L. Duff, Nursing, York University, Toronto, ON, Canada, Lynda Atack, Applied Arts and Health Sciences, Centennial College, Toronto, ON, Canada, Mina Singh, Nursing, York University, Toronto, ON, Canada AIMS: One goal of this preliminary research is to explore the impact of telehomecare (THC) on quality of life for community-dwelling adults who have chronic illness. Telehomecare uses telecommunication technology to provide video-based face-to-face communication, observation, and remote monitoring of heart and lung sounds, blood pressure, pulse, oxygen saturation etc. This technology links patients with health care professionals (providers); a THC visit replaces a home or clinic visit. METHODS: This descriptive exploratory study uses mixed methods. In phase 1, indepth interviews of patients (primarily with diabetes, CHF, and stroke) and providers were used to develop a grounded theory of participant concerns/experiences. Participants also completed a satisfaction survey. In phase 2, patients complete SF36 as THC is

1780 /Satisfaction with a Single Pill Treatment of Hypertension and Dyslipidemia Elizabeth P. Merikle, Outcomes Research, Pfizer Canada Inc., Montreal, Quebec, Canada, Gretchen Marcucci, Chris Evans, Mapi Values, Mapi Values, Boston, MA AIMS: The goal of the present study was to assess the satisfaction of patients with a single pill (CADUET) for treating both hypertension and dyslipidemia using the Expectations and Satisfaction with

A-44

Treatment Questionnaire-Short Form (ESTQ-SF) and to establish its psychometric properties. METHODS: JEWEL was a 16-week, international, multicentre, open-label, titration-to-goal study in patients with hypertension and dyslipidemia (N=2225). Patients were administered the ESTQ-SF at baseline and at follow-up visits. The ESTQ-SF is a shorter version of the previously validated ESTQ. Patients were grouped according to their treatment history (hypertension, dyslipidemia, hypertension and dyslipidemia, and treatment naïve). Psychometric analyses were performed on satisfaction scale and expectation items of the ESTQ-SF. Mixed models were used to assess significance of changes in satisfaction scores over time and to assess differences based on treatment history. Paired t-tests were used to assess changes in treatment expectations. RESULTS: Internal consistency for the satisfaction scale was high (alpha=0.77). Test-retest reliability ranged from 0.53 to 0.61. Patients' level of expectation about their pill burden increased when switching or incorporating CADUET into their treatment regimen. This improvement was statistically significant at the first follow-up visit (p < 0.0001) and was maintained throughout follow-up. Patients meeting treatment goals had significantly higher satisfaction at the end of the trial. A positive and statistically significant trend in satisfaction was observed across all groups (p < 0.0001). Satisfaction scores varied by previous treatment history with treatment naïve patients having the highest levels of satisfaction (p < 0.0001). CONCLUSIONS: The ESTQ-SF demonstrates adequate reliability and is sensitive to changes in treatment satisfaction that occur in patients treated with CADUET for hypertension and dyslipidemia. These findings suggest that a single pill treatment for hypertension and dyslipidemia will help in their effective management.

Quality of Life from the Caregivers Perspective 2 1783 /Proxy Versus Self-Reported Responses on the Pediatric Functional Assessment of Cancer Therapy (PEDSFACT) Jin-shei Lai, CORE, ENH & NWU, Evanston, IL, Jennifer Beaumont, CORE, ENH, Evanston, IL, Stewart Goldman, Hematology/Oncology, CMH & NWU, Chicago, IL AIMS: Using parent proxies to report children's health related quality of life (HRQL) is common as proxies are thought more valid and sensitive to detecting change over time than children's selfreport. However, HRQL is a subjective perception emphasizing individual's phenomenological experiences. Thus, proxies may not truly reflect children's HRQL. This study attempts to understand the relationship between proxies and self-reported HQOL using various analytic approaches. METHODS: 96 cancer patients (S) age 8-17, and their parents (P), completed the pedsFACT. The pedsFACT measures physical function (PF: 8 items), positive illness experience (PI: 4 items), emotional distress (ED: 6 items) and social/ family function (SF: 5 items). Associations between S & P were examined using Cochran-Mantel-Haenszel (CMH) statistics. Agreement was examined using Kappa statistics. Paired t-tests were used to examine response differences between S & P. Proxy was considered comparable to self-report when it was associated (CMH, p<.05), agreeable (Kappa >=.3) and not significantly different (ns t-test, p >.05). RESULTS: In PF, all 8 items had significant CMH; 7 items had Kappa > 0.3; 5 had ns t-test. In PI, all had significant CMH; 2 had Kappa=0.3; 2 had ns t-test. In ED, 3 items had significant CMH; 2 had Kappa>0.3; 4 had ns t-test. In SF, 3 had significant CMH; 1 had Kappa>0.3; 4 had ns t-test. Among 8 items with significant t-test, P responded more severely than S on all but 1 PI item. 7 items (4 PF, 1 PI, 1 ED, 1SF) satisfied all 3 criteria and were considered replicable between P & S. CONCLUSIONS: Inconsistent results were found across various analyses. S & P responses were significantly associated for most items. Though in general Kappa statistics were not high, better agreement was found in PF, which is consistent with the literature. Proxy reports on a subset of 7 items are sufficiently representative of self-report, yet the use of proxies to measure children's HRQL using the complete pedsFACT is not recommended. Despite a small sample, these results are promising.

1277 /Assessing the Construct Validity of the Health Utilities Index Mark2 (HUI2) and Mark3 (HUI3) in Cardiac Care Maria J. Santana, Public Health Sciences, University of Alberta, Edmonton, Alberta, Canada, David H. Feeny, Health Utilities, Inc, Dundas, On, Canada AIMS: The purpose is to assess the cross-sectional construct validity of HUI2 and HUI3 in heart transplant patients. METHODS: Data were collected in a randomized controlled clinical trial to asses the effects of including health-related quality of life measures in routine clinical care. Baseline data was collected for 52 pre- and post-heart transplant patients visiting the out-patient clinic at the University of Alberta Hospital, Edmonton. A priori hypotheses were specified about the relationships between single attribute utility scores, overall HUI utility scores, Hospital Anxiety and Depression Scale (HADS) scores and results from Six Minute Walk Test (6MWT). Standards to classify the degree of association were taken from Guyatt et al (1987): strongly correlated>0.50; moderately correlated, 0.35-0.50; weakly correlated, 0.20-0.34; negligible, 0.00-0.19. A priori hypotheses:1. HUI2 emotion should be strongly correlated with anxiety from HADS. 2. HUI3 emotion should be strongly correlated with depression from HADS.3. HUI2 mobility should be moderately correlated with 6MWT.4. HUI3 ambulation should be moderately correlated with 6MWT.5. HUI2 self-care should be negligibly correlated with anxiety from HADS.6. HUI3 pain should be weakly correlated with depression from HADS.7. HUI3 speech should be negligibly correlated with depression from HADS.8. HUI3 ambulation should be negligibly correlated with depression from HADS.9 Overall HUI3 should be moderately correlated with depression from HADS.Pearson's correlation was used to assess the degree of association. RESULTS: 69%of the sample were male with a mean age of 54. Of the nine a priori hypotheses formulated, 7 were supported. Hypotheses 2 and 9 were off by one category. Correlations were significant at the 0.05 level (2-tailed). CONCLUSIONS: Evidence from this study supports the cross-sectional construct validity of HUI2 and HUI3 in heart transplant patients.

1369 /Evaluation of Needs of Families with a Child with Renal Failure Jeanet Bruil, Symone Detmar, Gert Jacobusse, Esther Hosli, Child Health, Prevention and Physical Activity, TNO Quality of Life, Leiden, NL AIMS: Aim of the study was to describe the need for support in families with children with renal failure as input for a family program. METHODS: After 27 families had been interviewed, a written questionnaire was developed including items on background characteristics, family support (received and needed), the Impact on Family Scale and health-related quality of life. Descriptive analyses and univariate variance analyses were carried out to determine needs within the families and differences between age groups. Children older than 10 years old filled in a questionnaire themselves. 246 families received a letter from their hospital of which 146 were willing to participate (59%). RESULTS: 112 parents (77%) and 78 adolescents participated in the study. Most of the children (n=85; 75%) had a transplant. One third had additional health problems. These children as well as older children who received dialyses had significantly more often financial support (p<.05). Parents of younger children reported significantly more problems on the Impact on Family Scale, especially on the social subscale (p=.014), personal stress and total scale (p<.05). Parents often reported financial problems, tiredness and worries. 42 parents (37%) needed more

A-45

support than received. More support was mainly needed in finances, administration, school, holiday, caretaking of the other children in the family. Parents of younger children needed more instrumental support in taking care of family and child. Adolescents reported most problems on school, being able to do things in free time and problems in sports. In addition, they also had financial problems and would like to have more contact with other adolescents with renal failure. CONCLUSIONS: Although medical care is organised multidisciplinary, non-medical issues should be given more attention. Specifically finances, instrumental support in taking care of the child at home, school and holidays are areas where more support is needed. Adolescents need more contacts with other patients and more support in school, free time and sports.

(n=77) of caregivers of non-elderly HD patients. Their quality of life was evaluated by the SF-36 questionnaire, the Caregiver Burden scale and the Cognitive Index of Depression. RESULTS: The majority of the caregivers of the elderly were woman (78%), aged 55+/-l5yrs and sons/daughters (40%) of the patients. Cargivers`Emotional Aspect, Vitality and Mental Health were the most affected dimensions. The Mental Component Summary adjusted mean score was worse for caregivers of PD than HD elderly patients (36.3 vs. 44.8, P&It00.l). The most affected Cargiver Burden scale dimensions were General Strain and Environment. 32% of the cargivers presented depression signs. CONCLUSIONS: Caregivers of elderly dialysis patients, especially of those on PD, experience a significant burden and adverse effects in their quality of life. Educational, social and psychological support interventions should be considered to improve the quality of life of both, cargivers and patients.

1499 /The Needs & QOL of ALS Family Caregivers Mary Tederous-Williams, Counseling, School & educational Psychology, University at Buffalo, Orchard Park, NY, James P. Donnelly, Counseling, School & educational Psychology, University at Buffalo, Buffalo, NY, Tomas Holmlund, Neurology, Dent Instiution, Amherst, NY, Michael J. Battaglia, Neurology, Buffalo Medical Group, Buffalo, NY

1130 /Quality of Life & Family Burden in Caregivers of Schizophrenic Patients Selwyn Stanley, Social Work, Bishop Heber College, Tiruchirappalli, Tamilnadu, India, Shwetha Sundar, Social Work, Sowmanasya Hospital, Tiruchirappalli, Tamilnadu, India

AIMS: The goals of this study were to develop a model of the needs of the Amyotrophic Lateral Sclerosis family caregivers,to assess the QoL of family caregivers,to examine the relationship of identified needs,& QoL to patient & family characteristics. METHODS: This study was conducted in three phases:1)ALS family caregivers(N=19) seen in a multidisciplinary ALS clinic were asked to identify their needs since the time of ALS diagnosis of their family member,2)The resulting 109 unique needs were printed on individual cards and sorted by 12 of the caregivers, 3)The needs were rated on importance by the same 12 caregivers who completed the sorting. The caregivers also completed a brief standardized measure of health-related quality of life (SF-8).The sort data were analyzed in two steps:1) Nonmetric multidimensional scaling(MDS);2)Cluster analysis of the MDS values for each need.The SF-8 data was analyzed in relation to the needs ratings, patient and caregiver characteristics. RESULTS: The MDS produced an interpretable solution with a stress value of .36. The cluster analysis resulted in a four map of ALS caregivers needs:Stage 1: Early Coping & Adjustment;Stage 2:Maintenance; Stage 3:Transition to End Stage; and Stage 4:Coping with Change & Loss. Results further indicated that the symptom presentation of ALS in the caregiver's family member was associated with lower levels of mental and physical health relative to national means,particularly for patients with initial bulbar symptoms (p=.03). Additionally, caregivers who resided with their family member with ALS and responsible for 24-hour care had poorer mental(p=.04)and physical health (p=.02)than ALS family caregivers that did not live with the ALS family member. CONCLUSIONS: The data from this study of caregiver needs suggest, there is a predictable pattern of specific needs, and that QoL for caregivers is related to characteristics of the patient's disease,living arrangements and other differences.Future studies should examine the relationship of need satisfaction to QoL as well as the potential of caregiver support programs to meet identified needs and improve QoL.

AIMS: Schizophrenia is a severe mental illness, which is stressful not only for patients, but also for family members. Providing care to family members dealing with chronic illness may result in feelings of burden or strain for caregivers that can diminish their quality of life (Sales, 2003). The study was conducted at the community care center, Gunaseelam, in Tamilnadu, India which is attached to a Hindu temple. The site is a known place for religious healing of the mentally ill for the past 200 years. Now a modified integrated treatment approach is offered here exclusively for schizophrenic patients which is a blend of traditional temple rituals and modern psychiatric intervention. A comparative pre and post intervention analysis to determine the extent of change in caregivers of schizophrenic patients attending a community based rehabilitation programme with regard to their family burden and perceived quality of life was undertaken. METHODS: An integrated intervention package comprising of pharmacotherapy, psychosocial therapies and spiritual therapy was provided for a period of 48 days to the patients and their caregivers. Thirty primary caregivers were included as respondents. The Family Burden Scale by Pai and Kapur and the Perceived Quality of Life Scale by Patrick et al., were administered at the point of admission of the patients to their primary caregivers as well as 90 days after discharge during the first follow up contact. RESULTS: Results indicate significant reduction in caregiver burden and enhanced QOL following intervention. Further, a significant negative correlation between the family burden and QOL scores of the caregivers was obtained indicating that these two dimensions mutually influence one another. CONCLUSIONS: It is suggested on the basis of this study that spiritual therapy specific to the religious orientation of patients combined with pharmacotherapy and other psychosocial therapies may enhance the QOL of caregivers in schizophrenia.

1579 /Quality of Life of Caregivers of Elderly Patients on Hemdialysis or Peritoneal Dialysis Angélica Belasco, Dulce Barbosa, Lucila V. Carneiro, Nursing, Domingos Belasco Junior, Fisioterapy, Ricardo Sesso, Medicine, Federal University of São Paulo, São Paulo, São Paulo, Brazil

Methodology Developments in QOL 2 1719 /Analytic Approaches to Dealing with Missing QOL Data in HIV Clinical Trials Albert W. Wu, Health Policy and Management, Johns Hopkins University, Baltimore, MD, Herbert Thijs, Centre for Statistics, Universiteit Hasselt, Diepenbeek, Belgium, I-Chan Huang, Epidemiology and Health Policy Research, University of Florida, Gainesville, FL, Henrik W. Finnern, Health Economics & Outcomes Research, Boehringer Ingelheim GmbH, Ingelheim, Germany, Diane

AIMS: The objective of this study was to describe caregivers`characteristics and to evaluate their burden and quality of life. METHODS: We studied caregivers of elderly patients in hemodialysis (HD)n=84, in peritoneal dialysis (PD)n=40 and a group

A-46

L. Fairclough, Preventive Medicine and Biometry, University of Colorado, Denver, CO

changed these to -0.016 (age), -0.22 (sex), and-0.28 (Black). Plots and regression analyses of EQ-5D, HALex, HUI2, and SF-6D scores on QWB-SA score (which displayed little ceiling effect), showed a flattening of curves at QWB of approximately 0.6. Probit and tobit models of the 4 indices on QWB yielded very consistent results for all indices, with values corresponding to the 0.6 QWB-SA as the cutpoint for dichotomizing and censoring. CONCLUSIONS: Four of the five indices display ceiling effects in the general population. All five indices load highly on a overall construct. Censoring began in four indexes above QWB-SA=0.6. As censoring occured across a range, differential ceiling effects in sub-attributes may play a role

AIMS: Missing QOL data in clinical trials can impact conclusions but the effect has not been studied in HIV. Complete case analysis (CC) and last observation carried forward (LOCF) are commonly used to handle missing data but require strong assumptions. We used several analytic approaches in a QOL dataset under assumptions of missing completely at random (MCAR), missing at random (MAR) and missing not at random (MNAR). METHODS: QOL data were combined from 2 large open-label multinational HIV clinical trials comparing treatments A and B over 48 weeks. Patients failing on the comparator regimen could be treated in a roll-over study. Inclusion in the QOL analysis required completion of the MOS-HIV at baseline and at least one follow-up visit (wk 8, 16, 24, 40, 48). Primary outcomes were change from wk 0 to 48 in MHS and PHS score analyzed using CC, GEE and LOCF (assuming MCAR), direct likelihood (MAR) and sensitivity analyses using joint mixed effects model and MCMC multiple imputation (MNAR). Time and treatment were included in the likelihood-based models and MCMC. Both baseline and longitudinal variables (adverse event, reason for discontinuation, time to response) were used in the imputation model. RESULTS: At baseline 524 patients randomized to TrtA and 491 to TrtB completed the MOS-HIV. By wk 48, 71% of patients on TrtA and 31% on TrtB completed the MOS-HIV. Examining changes within each treatment group, CC and MCMC produced the largest or most positive changes. The joint model was most conservative; direct likelihood and GEE produced intermediate results; LOCF showed no consistent trend. There was greater spread for within-group changes than between-group differences (Within MHS TrtA -0.1 to 1.8, TrtB 0.4 to 1.9; Between -0.5 to 0.3; Within PHS TrtA -1.5 to 0.6, TrtB 1.7 to 0.0; Between 0.2 to 1.0). All between-group differences were small and < 0.1 SD. CONCLUSIONS: Likelihood-based models that do not require MCAR can be recommended for handling missing QOL data. Sensitivity analyses using auxiliary information can help to investigate how missing data can influence the results.

1785 /Effects of Norm Selection on PCS and MCS: A Tale of Two Studies Colleen M. Renier, Irina V. Haller, Education and Research, SMDC Health System, Duluth, MN, Thomas E. Elliott, Education and Research, Duluth Clinic, Duluth, MN, Jeanette A. Palcher, Education and Research, SMDC Health System, Duluth, MN, Ana M. Fernandez, Rheumatology, Duluth Clinic, Duluth, MN AIMS: The physical and mental health summary scales, PCS and MCS, of the SF-series of quality of life (QoL) instruments are weighted sums of subscale z-scores dependent on the sample used for standardization. We sought to evaluate the potential impact of using age and gender specific samples to standardize QoL scores in disease specific research. METHODS: Two studies were standardized using U.S. population total sample (ttl), gender (mf), age (agp), and age within gender (mfxagp) means and standard deviations (Ware, et. al.). In the first (OPK), the SF-8 was completed by participants of a case-control study of men with osteoporosis (OP) or osteopenia and women with OP, ages 40-79, diagnosed between 7/1/01 and 6/30/03, group matched to controls from a general patient population by gender, age group and primary care location. In the second (PAIN), the SF-36 v.1 was completed by patients of a pain clinic admitted between 10/29/99 and 7/2/04. Correlations of PCS and MCS with age and gender were analyzed and predictive regression models compared. RESULTS: In OPK, age was significantly correlated (sig. corr., <.01) with PCSttl, PCSmf, MCSttl and MCSmf, and gender was sig. corr. with PCSmf and PCSmfxagp. In PAIN, age was sig. corr. with PCSagp, PCSmfxagp, MCSttl and MCSmf, and gender was sig. corr. with PCSmf and PCSmfxagp. In OPK, the strongest predictive models for PCS were for PCSttl controlled for age and PCSagp (R2=.361 for both) and the strongest model for MCS was for MCSttl controlled for age (R2=.237). In PAIN, the strongest predictive model for PCS was for PCSttl controlled for age and gender (R2=.241) and the strongest model for MCS was for MCSttl controlled for age (R2=.367). CONCLUSIONS: Since disease can evenly impact QoL, sig. corr. between PCS, MCS, age and gender remained after standardizing for them, with standardization also creating variation. In light of this and the fact that total population standardization, adjusted for age and/or gender, provided the strongest predictive models, use of gender and/or age normative samples should be used with caution.

1769 /Ceiling Effects and Unidimensionality for Five Health Utility Indices Mari Palta, Population Health Sciences, Dennis G. Fryback, Medicine, Nancy C. Dunham, Population Health Sciences, Jing Xu, Popul;ation Health Sciences, Jennifer M. Buechner, Population Health Sciences, University of Wisconsin, Madison, WI, Ronald D. Hays, Medicine, UCLA, Los Angeles, CA AIMS: To evaluate ceiling effecs and the extent to which preference scores from the QWB-SA, EQ-5D, HUI2, HALex and the SF-6D form a unidimensional latent construct in the U.S. general population. METHODS: The National Health Measurement Study obtained a national unequal probability sample (n = 1262) of the noninstitutionalized US population ages 35-89. Four indices were administered in random order by telephone interview followed by the HALex. We used factor analysis to estimate loadings on a latent construct, accounting for survey weights and possible censoring. Relationships among indices were also examined graphically and using regression. RESULTS: Population means of EQ-5D, HALex, HUI2, QWB-SA and SF-6D were 0.86, 0.79, 0.84, 0.65, and 0.82, respectively. Only one eigenvalue exceeded 1.0 and the one-factor confirmatory categorical? factor analysis fit the data well (RMSEA = 0.xx), with loadings for the five indices of 0.84, 0.79, 0.83, 0.79 and 0.85, respectively. Considering the skewed distributions of EQ-5D, HALex, HUI2, and SF-6D to be censored at 1.0 changed loadings to 0.81, 0.76, 0.81, 0.81 and 0.86, and estimated mean scores to 0.92, 0.81, 0.85, 0.65, and 0.83. Regression analysis of the latent construct gave coefficients for age (-0.014 per year), gender (-0.19 for women) and race (-0.31 for Blacks vs. other). Accounting for censoring

1196 /Disentangling Health and Impairment in Measuring Quality of Life: How Can Response Shift Theory Help? Carolyn E. Schwartz, DeltaQuest Foundation, Concord, MA, Elena M. Andresen, College of Public Health & Health Professions, University of Florida, Gainesville, Gainesville, FL, Margaret A. Nosek, Physical Medicine & Rehabilitation, Baylor College of Medicine, Houston, TX, Gloria L. Krahn, Child Development and Rehabilitation Center, Oregon Health & Science University, Portland, OR AIMS: Recent advances in models of disability have made a conceptual differentiation between health and disability, and have

A-47

recognized that the environments in which people with impairments live can contribute to many of the disability-related limitations they experience. When one seeks to measure the health status of individuals with disability, it becomes necessary to consider that including function in the broader construct of health complicates the measurement problem. By design, most health status measures equate function with health, necessarily leading to worse measured health status in people with functional impairments regardless of their level of health. This paper discusses how response shift theory and appraisal can help to unconfound measurement of health status in disabled persons. METHODS: We discuss theoretical and conceptual distinctions building on response shift theory and other current developments in quality-of-life research. RESULTS: The components of the Sprangers and Schwartz model take on a distinct meaning for people with disabilities, and some of these components may be causal indicators that must be treated differently in measurement development. CONCLUSIONS: We make recommendations for future measurement development research to: (1) address the so-called disability paradox by developing disabilityneutral items; (2) explicitly measure appraisal processes; (3) implement fundamental longitudinal research that documents the prevalence of response shift in people with disability; and (4) to include selected relevant causal indicators without compromising the measurement model.

Rheumatology and Key Issues in QOL 1495 /The Development of an Orthopaedic Waiting List Prioritisation Questionnaire: Meeting the Challenges of Clinical Priority, Gaming and Equitable Access Richard H. Osborne, Centre for Rheumatic Diseases, Dept Medicine (RMH), Kerry Haynes, Centre for Rheumatic Diseases, Medicine (RMH), The University of Melbourne, Melbourne, Vic, Australia, Peter Fayers, Public Health, Aberdeen University., Aberdeen, Aberdeen, UK, Stephen E. Graves, Medicine (RMH), Peta Chubb, Centre for Rheumatic Diseases, Debra Robbins, Center for Rheumatic Diseases, Dept Medicine (RMH), The University of Melbourne, Melbourne, Vic, Australia AIMS: We aimed to develop a clinically sensitive system for prioritization of people who may require a joint replacement. Challenges included fair assessment for patients across demographic groups, assignment of queue position, as well as gaming (false high scorers) & stoic scorers. METHODS: To identify priority criteria, stakeholder consultations were undertaken & 4 workshops each were held with surgeons & patients. Domains were pain, activity limitations, psychosocial wellbeing, economic impact and deterioration. Draft Guttman items were cognitively tested with patients & surgeons. 38 items survived appraisal & were administered to 600 metropolitan and rural patients. 11 items survived clinimetric and statistical item reduction. Validation included co-administration with standardised measures (N=960 patients), further cognitive interviews, surgeon rating & test-retest. Surgeons (N=96) weighted items using Discrete Choice Experiments, which involved prioritizing patient vignettes derived from MAPT items. RESULTS: The DCE scaling generated an interval scale which clearly ranked patients across the disease continuum. The MAPT differentiated people on or not on waiting lists (p<0.001), and was highly correlated with other measures, MAPT vs: WOMAC (r=0.75); Oxford Hip/Knee (r=0.80/0.75); AQoL (r=-0.69); EQ5D (r=-0.76); HADS Depression (r=0.57) and Anxiety (r=0.55); SF-36 PCS (r=0.62) and SF-36 Mental Health (r=-0.43). Test-retest (N=80) was high 0.75 as was Cronbach alpha; 0.84. Patient MAPT and MAPT score from clinical assessment (198 pairs) were highly correlated (r=0.74) suggesting high criterion validity and minimal gaming. CONCLUSIONS: The MAPT is short, clinically relevant and is highly endorsed by patients and clinicians. It is a single broad measure of arthritis severity and surgical priority. It has very strong concordance with standard disease-specific, generic and clinical ranking. Potential uses include priority assignment (incl. referral), monitoring, resource allocation and clinical trials

1181 /Do We Have a Psychological Foundation to Conduct Importance Weighting? Chia-huei Wu, Psychology, National Taiwan University, Taipei City, Taiwan (R.O.C.), Grace Yao, Psychology, National Taiwan University, Taipei, Taiwan (R.O.C.) AIMS: Importance weighting is a common idea in measurements for QOL, job satisfaction and self-esteem, etc. However, few studies discussed the psychological meaning of importance weighting. This study was aimed to discuss the psychological foundation of importance weighting from Locke’s range-of-affect hypothesis. Range-of-affect hypothesis indicated that the satisfaction evaluation of an item was determined by the have-want discrepancy, importance and their interaction, implying that satisfaction evaluation has incorporated the judgment of importance, therefore, weighting satisfaction score with importance score is unnecessary. The purpose of this study was to review studies in different fields to show that range-of-affect hypothesis depicts a general phenomenon for satisfaction evaluation and provides a theoretical account for the uselessness of importance weighting. METHODS: Related studies in job satisfaction literature were reviewed. Further, two QOL-related studies conducted by the authors were reviewed. Finally, a new study was conducted in the context of self-esteem. Two hundred and sixtynine undergraduate students at National Taiwan University participated in this study. They completed questionnaires measuring domain actual-want self-discrepancy, domain self-satisfaction and global self-esteem. Moderated regression analysis was used to test (1) whether domain actual-want self-discrepancy and domain importance satisfaction have a significant interaction effect on domain satisfaction, and (2) whether domain importance and domain satisfaction have a significant interaction effect on global self-esteem. RESULTS: All results revealed that (1) satisfaction evaluation has incorporated the judgment of importance and (2) weighing satisfaction with importance did not contribute to predicting global satisfaction. CONCLUSIONS: The phenomenon depicted in rangeof-affect hypothesis is general and the procedure of importance weighting is unnecessary. Further, the range-of-affect hypothesis also provides an implication on QOL promotion from the perspective of self-regulation theory.

1203 /Reproducibility and Responsiveness of the Functional Handicap Score (FHS) and Dutch Arthritis Impact Measurement Score (AIMS2) for Patients with Wrist or Hand Symptoms in Primary Care Marinda N. Spies-Dorgelo, General Practice, Caroline B. Terwee, EMGO Institute, Wim A. Stalman, Daniëlle A. van der Windt, General Practice, VU University Medical Center, Amsterdam, the Netherlands AIMS: We determined the clinimetric properties of two questionnaires assessing symptoms (FHS) and physical functioning (subscale hand and finger function of the AIMS2) in a Dutch general practice population. The FHS is designed for patients with carpal tunnel syndrome, the Dutch-AIMS2 for patients with rheumatic diseases. We examined if these questionnaires are also applicable in a less specific group of patients who consult their GP for hand and wrist symptoms. METHODS: The first 84 participants in a one-year follow-up study on the diagnosis en prognosis of hand and wrist problems completed the FHS and Dutch-AIMS2 twice within 1 to 2

A-48

weeks. These data were used to assess test-retest reliability (ICC) and smallest detectable change (SDC, based on the standard error of measurement (SEM)). For assessing responsiveness change scores between baseline and 3 months follow-up were related to an external criterion to assess minimal important change (MIC). We calculated the group size needed to detect the MIC beyond measurement error. RESULTS: For the FHS the ICC was 0.68 (95% CI: 0.54-0.78). The SDC at individual level was 1.00 and at group level 0.11, both on a 5point scale. The MIC was 0.23, exceeding the SDC at group level. The group size required to detect a MIC beyond measurement error was 19 for the FHS. For the Dutch-AIMS2 the ICC was 0.62 (95% CI: 0.47-0.74). The SDC at individual level was 3.80 and at group level 0.42, both on a 11-point scale. The MIC was 0.31, smaller than the SDC at group level. The group size required to detect a MIC beyond measurement error was 150. CONCLUSIONS: Our study confirms that the performance of a questionnaire is often dependent on the characteristics of the population in which it is used. In our heterogeneous general practice population the FHS was suitable to assess severity of symptoms, whereas the Dutch-AIMS2 was less suitable to measure physical functioning in persons with hand and wrist problems.

Ontario, Canada, John Flannery, Physiatry, Toronto Rehabilitation Institute, Toronto, Ontario, Canada, Marta Krywonis, Etobicoke Community Care Access Centre, Etobicoke, Ontario, Canada, Rhona McGlasson, Total Joint Network, Toronto Western Hospital, Toronto, Ontario, Canada AIMS: To determine if WOMAC pain and function outcomes and patient satisfaction were maintained after the institution of a new standardized care pathway for people undergoing primary hip or knee replacement (TJA). Based on a partnership of 23 institutions, this care pathway was designed to provide a seamless transition from acute to rehabilitation sectors by pre-surgery streaming of patients to inpatient or home-based rehabilitation and to reduce length of stay (LOS). METHODS: Rehab destination and LOS were tracked against bench marks of minimum 50% discharged home and an inpatient rehab stay of 7 days. Subjects recruited to the evaluation completed the WOMAC pre-surgery with follow-up including a satisfaction questionnaire 3 months later. The sample of 500 ensured the 99% confidence interval (CI) of the mean of WOMAC pain and function would be within 2 and 5 points respectively of the results of a randomized trial where TJA patients received home-based or inpatient rehab. RESULTS: On average greater than 50% of people have been discharged home from acute care and over 50% of those receiving inpatient rehab have been discharged in 7 days, with 80% discharged by day 10. To date, 478 of the patients (mean age 67, 295 females) have had their surgery and 86 have completed 3 month follow-up (follow-up will be complete in July 2006). To date, those discharged home have mean pain and physical scores of 80.5 and 76.6% (high scores are better pain relief and function) and those receiving inpatient rehab have pain and function scores of 77.5 and 72.5%. These values are within our a priori CI for outcome. Only 5% of individuals reported dissatisfaction with the rehabilitation planning and care. CONCLUSIONS: This model of care has achieved the goals of increasing the proportion of people discharged home following TJA and decreased rehab LOS with increased system capacity for rehab. The patient outcomes are maintained. The remainder of the participants will be followed and variances from the care pathways will be evaluated.

1739 /Comparison of Habits of Fracture Group Due to Osteoporosis and Non-Fracture Group Using EQ-5D Nobuo Koinuma, Health Administration and Policy, Tohoku University, Sendai, Japan, Yasutake Monma, Michiya Ito, Sayuri Kaneko AIMS: Osteoporosis is the main cause of fracture of the elderly and is a disease which has a possibility of spoiling the QOL greatly, by long-term treatment, concurrence of other diseases, deterioration of ADL, bedridden, and loss of a definite aim in life etc. This study is aimed at clarifying the preventive effect against fracture due to osteoporosis by improvement of the eating habits from a viewpoint of QOL. METHODS: The group of fracture due to osteoporosis and the non-fracture group using the pair-mach method. The questionnaire was distributed to the local residents of three areas aged 55 and over, and patients with fracture due to osteoporosis admitted in 40 hospitals. The questionnaire consists of about 90 items concerning eating habits in the age of 20 years old, 40 years old, and the present. As investigation of QOL, we used Euro QOL and VAS. RESULTS: The data from 2,076 local residents (response rate: 60.9%) and 55 inpatients (100%) was analyzed with the adjustment of sex and age. The items as which the significant difference was regarded between the fracture group and the non-fracture group are weight, menopause age, the number of years after menopause, apoplexy, the medicine against osteoporosis, depressor, milk, fermented soybeans, bread, meat, sesame oil, seaweed, and polished rice. It is a level of movement and usual activity that there was a significant difference statistically by the five items of Euro QOL. In the value of utility or VAS, there was no significant difference between two groups. It is reported that there is no significant difference in utility in two groups and our study is equal to the previous studies. However, there might be a limit in catching QOL of patient with fracture in respect of sensitivity since EQ-5D is not a tool to measure QOL of the disease specifically. CONCLUSIONS: We carried out a field survey to compare habits with fracture group with non-fracture group. In EQ5D, it is movement and usual activity that the significant difference was observed in two groups and there was no difference in utility.

Qualitative Approaches and Challenges in the QOL Field 1084 /Being Normal: Quality of Life Domains for Persons with a Mental Illness Deborah J. Corring, Rehabilitation Sciences, University of Western Ontario, London, Ontario, Canada AIMS: Several authors argue that quality of life has a major subjective element and can only be assessed by self-report. The purpose of this study was to explore the construct of quality of life from the perspective of individuals diagnosed with severe and persistent mental illnesses such as schizophrenia. METHODS: Qualitative research strategies, specifically in-depth interviews and focus groups, were used. Fifty-three individuals participated in data collection,were for the most part single, unemployed and diagnosed with schizophrenia. Trustworthiness strategies employed included member checking, triangulation, peer debriefing, thick description and an audit trail. Analysis of the transcripts was done using the constant comparative method from verbatim transcriptions. RESULTS: There were two dominant themes permeating these results- stigma, and the pervasive fear of the major positive symptoms of psychoses. In addition, four domains were identified, specifically the experience of illness, relationships, occupation and sense of self. CONCLUSIONS: Most current quality of life measures do not sufficiently explore the domains identified in these results. Stigma toward those with mental illness has been identified as a

1238 /Patient Reported Outcomes as a Quality Indicator of Standardized Care Plans for Total Joint Replacement Aileen M. Davis, Outcomes and Population Health, Toronto Western Research Institute, Toronto, Ontario, Canada, Nizar N. Mahomed, Orthopaedic Surgery, Toronto Western Hospital, Toronto, Ontario, Canada, Heather Brien, Toronto Community Care Access, Toronto,

A-49

major challenge for some time and many calls have been made to further understand it and eliminate it. Participants in the study talked about the desire to be seen as normal human beings. The issue of what is normal is a complex one. They speak of wanting a normal ordinary life, much like the rest of society, but in actual fact many have lowered expectations. Many persons with mental illness are willing to settle for the basics in life - mental and physical health, supportive relationships, meaningful occupations and a positive sense of self - believing that acquisition of these basics will lead to a more satisfactory quality of life. Ensuring that they are able to do so requires action on all of our parts.

of British Columbia, Vancouver, British Columbia, Canada, Stefan Cano, Neurology, University College London, London, England, United Kingdom, Constance M. Chen, Peter G. Cordeiro, Plastic and Reconstructive Surgery, Karen E. Hurley, Psychiatry and Behavioral Sciences, Memorial Sloan-Kettering Cancer Center, New York, NY AIMS: Cosmetic and reconstructive breast surgeries are elective procedures that women undergo primarily to enhance quality of life (QOL). Identifying the components of QOL that underlie women's motivation for breast surgery can improve patient care and satisfaction. Thus, we explored factors that may influence decisionmaking about breast surgery (motivation, perceived barriers) among women who had undergone breast augmentation, reduction or reconstruction to identify key QOL issues. METHODS: Qualitative in-depth semi-structured interviews were conducted with a convenience sample of women who underwent breast surgery within the past three years. Interviews were conducted as part of the item generation phase for a surgery-specific scale to assess QOL in breast surgery. Forty-eight women were recruited (breast augmentation n=20; reduction n=15; reconstruction n=13) from surgeons in Western Canada. Interviews were transcribed, coded, and analyzed using grounded theory. RESULTS: Across all patient groups, women identified lower "self-confidence" and "self-esteem" due to feeling "different" because of breast size or appearance as key motivations for surgery. Breast reconstruction patients emphasized "feeling whole again" and "regaining a sense of normalcy" after breast cancer; similarly, breast augmentation and reduction patients also expressed a desire to look "normal" or "proportional." All three groups identified "stigma" associated with plastic surgery as a key barrier, as well as concern about "seeming vain" or "narcissistic." CONCLUSIONS: Breast augmentation, reduction and reconstruction patients are distinct patient groups who pursue surgery under different life circumstances. Our findings, however, suggest similar underlying QOL concerns and similar barriers to surgery in all three groups. Understanding women's QOL concerns, including tensions relating to women's desire to look "normal," without seeming "vain," can promote improved physician-patient interactions and overall quality of care for breast surgery patients.

1427 /Patients' And Health Care Professionals' (HCPs) Perspectives on the Most Important Quality of Life Issues in Bone Metastases Kristin Harris, Sukirtha Tharmalingam, Radiation Oncology, Toronto Sunnybrook Regional Cancer Centre, Toronto, ON, Canada, Andrea Bezjak, Radiation Oncology, Princess Margaret Hospital, University of Toronto, Toronto, ON, Canada, Jackson Wu, Radiation Oncology, Tom Baker Cancer Centre, University of Calgary, Calgary, Alberta, Canada, Michael Barton, Faculty of Medicine, CCORE, University of NSW, Liverpool Hospital, Liverpool BC, NSW, Australia, Jesmin Shafiq, CCORE, Liverpool Hospital, Liverpool BC, NSW, Australia, Richard Eek, Medical Oncology, Liverpool Hospital, Liverpool, Sydney, NSW, Australia, Galina Velikova, Cancer Research UK Clincial Centre, St. James's Hospital Leeds, Leeds, UK, Andrew Bottomley, Quality of Life Unit, EORTC, Brussels, NONR, Belgium, Edward Chow, Radiation Oncology, Toronto Sunnybrook Regional Cancer Centre, Toronto, ON, Canada AIMS: To identify important QOL issues in bone metastases from patients’ and HCPs' perspective. METHODS: Patients with bone metastases and HCPs involved in their care were asked to choose the most important 5 to 10 QOL issues from a questionnaire listing 61 items generated from literature review and interviews with patients and HCPs. The items related to: symptoms (13), function (15), treatment side effects (3), psychosocial domains (27), treatment expectations (2) and other (1). Interviews took place in 3 Canadian cancer centers and 1 center in Australia. RESULTS: A total of 328 patients and 106 HCPs were interviewed in Canada from Toronto Sunnybrook Regional Cancer Centre, Princess Margaret Hospital and Tom Baker Cancer Centre. Most patients were from outpatient clinics. HCPs included 36 radiation oncologists, 26 nurses, 22 medical oncologists and 11 palliative care doctors. Both patients and HCPs identified the following issues: long-term/chronic pain, difficulty carrying out usual daily tasks, able to perform self-care, able to perform role functioning. Additional QOL issues from patients were: lack of energy, worry about becoming dependent on others, worry about loss of mobility compromising independence, difficulty in carrying out meaningful activity, worry about disease progression deterioration in condition and future complications, financial burden due to the illness. Additional QOL issues from HCPs were: short-term (acute) severe pain, uncontrolled unmanageable pain not relieved by painkillers, pain at night preventing sleep, limited movement due to pain, pain at rest, hope for sustained pain relief. So far 20 patients and 22 HCPs have been interviewed in Australia with similar outcomes. Complete results will be updated. CONCLUSIONS: HCPs identified symptoms (7 related to pain), functions and treatment expectations as important QOL issues. Patients indicated financial burden and psychosocial issues as equally important as symptoms (only 1 related to pain) and functions.

1653 /Quality of Care and Quality of Life in Persons with Disabilities Using Focus Groups DIS-QOL Project Ramona Lucas, Fundaciò Institut Català de l'Envelliment. UAB, Barcelona, Spain, Sara Domenech, Fundació Institut Català de l' Envelliment. UAB, Barcelona, Spain, Jaume Sastre, Hospital de Día "Fundació Esclerosis Múltiple", Barcelona, Spain, Beatriz Garvía, Psychology, Fundació Catalana Síndrome Down, Barcelona, Spain AIMS: This presentation aims to report information about what people with intellectual and physical disabilities (ID, PD), carers and professionals believe are important to quality of life (QoL) and quality of care (QC) for people with disabilities. Also, we report information on attitudes towards disabilities (AD). This is part of DIS-QOL project financed by the EC. METHODS: Focus groups (FG) were conducted with participants from the Down Syndrome Catalan Foundation and one Day Hospital for Multiple Sclerosis located in Barcelona, Spain. Participants were adults with PD and ID; caregivers and health and social care professionals. Number of participants was 5 to 6. We collected information on sociodemographic variables and health and disability status in people with disabilities. The WHOQOL-BREF was used as the basis to elicit information about QoL. All FG lasted about two hours. All FG were audio taped. All participants signed the consent form. Analysis will be done using qualitative techniques. RESULTS: Five FG have been conducted: two with adults with disabilities, one with carers and two with professionals. Mean age of the sample was 37 (SD 9.5) with 69% being female. All WHOQOL-BREF areas and most of the facets

1315 /What Motivates Women to Undergo Breast Surgery? A Comparative Study of Body Image, Barriers, and Quality of Life in Three Surgical Groups Lisa R. Rubin, Psychiatry and Behavioral Sciences, Andrea L. Pusic, Plastic and Reconstructive Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, Anne Klassen, Pediatrics, University

A-50

were mentioned in most of the groups. Facets emphasized by ID subjects were self-esteem; autonomy, ability to work, independency, family and social relations. Facets emphasized by PD subjects were mobility, ADL, dependency, scarcity of health and social resources and transport. Families from PD subjects emphasized environmental facets including financial resources; and professionals emphasized scarcity of resources, psychological needs, autonomy and relationships. Attitudes from society more commonly mentioned were lack of respect and discrimination. All participants enjoyed participation. CONCLUSIONS: Information collected will allow establishing the experiences and needs of this population, which ultimately may help to improve the quality of care and their quality of life.

bowel function. But how to interpret these scores? What is 'normal' for ageing men? METHODS: Participants in the European Randomized study for Screening on Prostate Cancer (ERSPC) from the Rotterdam region, the Netherlands, received - attached to the invitation for the 2nd or 3rd screening round - a questionnaire on health (EPIC urinary and bowel modules, 12 single items on sexual functioning, SF-36 Mental Health and Vitality, and the EQ-5D VAS on self-rated health). Sexual dysfunction was defined as (almost) always having problems in achieving or maintaining an erection if wished to, or not being sexually active because of erectile problems. Response to the questionnaire was 90%, mean and median age 67, range 58-87 with 7 respondents aged >75. We report results of screen-negative men. RESULTS: Preliminary results (n=2,378) indicated erectile dysfunction in 19% of respondents, ranging from 13% in the youngest (i.e. 58-61 years) till 27% in the oldest group (i.e. >70). We observed worse urinary function and more urinary bother in older agegroups. Self-rated health (EQ-5D VAS) and SF-36 Vitality were also lower, i.e. worse, in older men. CONCLUSIONS: Erectile and urinary function were worse in oldest age groups. Remarkably, levels of bowel function and bother, and SF-36 Mental Health were similar in all agegroups. The average SF-36 scores were higher, i.e. better, than U.S. and Dutch norm scores, which may be explained by the high social economic status of participants to the screenings trial. We documented age-related patterns of changing sexual, urinary and bowel function in men without prostate cancer to enable better interpretation of outcomes after prostate cancer treatment, as well as functioning of men with e.g. prostatic hyperplasia or bladder carcinoma.

1612 /Evaluation of SEIQoL-DW as a Measure of Quality of Life in Patients Undergoing Cystectomy for Bladder Cancer: Patients’ Perspective Bhaskar K. Somani, Urology, Deborah Gimlin, Sociology, Peter Fayers, Public Health, James N'Dow, Urology, University of Aberdeen, Aberdeen, United Kingdom AIMS: Patients with advanced bladder cancer need major surgery with potential serious morbidity and a small risk of mortality. Only 30%-50% patients survive 5 years and half of these deaths is within the first year of surgery. Standard QoL measuring tools do not take into account patients view on what is important to them in their lives. We wanted to find out how important health or operation was in the lives of these patients just before surgery. METHODS: A single researcher using SEIQoL-DW interviewed 27 patients undergoing cystectomy 1-5 days before their surgery in four different hospitals across England and Scotland. SEIQoL-DW consists of a semistructured interview eliciting the five most important areas to respondents quality of life. All patients will be interviewed again after one year after surgery to look the changes in their QoL after cystectomy. Patients also had a Visual Analogue Scale assessment of QoL. RESULTS: Twenty-two of the 27 patients were retired at the time of interview( 8 females and 19 males). All patients nominated five domains each. Health, operation or cancer did not feature in 7 of the 27 (25.9%) patients in the list of five important domains/ cues. The mean SEIQoL-DW score was 75 (range: 53 - 94). The mean time taken to interview a patient was 21 minutes (range: 12-32 minutes). Apart from health or operation, the other domains most commonly nominated were family, friends, finance and hobbies. The mean Visual Analogue Scale score was 66.2 (range: 27-95). CONCLUSIONS: Contrary to the popular belief in the surgical community, cancer or operation is not always the most important determinant of QoL despite undergoing major life altering and potentially life threatening surgery. This should be borne in mind while counselling patients and helping them making decisions before surgery.

1521 /Experiences and Decision Processes During PSA Testing and Prostate Biopsy: A Qualitative Study Kerry N. Avery, Social Medicine, University of Bristol, Bristol, UK, Jane M. Blazeby, Surgery, United Bristol Healthcare Trust, Bristol, UK, Chris Metcalfe, J Athene Lane, Social Medicine, University of Bristol, Bristol, UK, David E. Neal, Oncology Centre, Addenbrookes Hospital, Cambridge, UK, Freddie C. Hamdy, Clinical Sciences (South), Royal Hallamshire Hospital, Sheffield, UK, Jenny L. Donovan, Social Medicine, University of Bristol, Bristol, UK AIMS: National trials evaluating prostate specific antigen (PSA) testing for prostate cancer (PCa) are in progress. Information about mens experiences of testing will supplement standard outcomes. The aim of this study was to understand experiences and decision processes during PCa testing. METHODS: A qualitative study nested within a cluster-randomised trial of case-finding and treatment for PCa. Semi-structured interviews with a purposive sample of 63 men aged 52 to 75 (mean 64) years responding to a PSA invitation with normal and raised PSAs, men not responding, men undergoing biopsy with positive and negative outcomes and men with a raised PSA refusing biopsy. Interviews were audio-recorded, transcribed, coded and analysed using constant comparison techniques. RESULTS: Overall, men accepting PSA testing and biopsy report positively on their experiences, regardless of final outcome. PSA testing is deemed an acceptable, routine procedure and biopsy is well-tolerated, with reports of pain, compromised privacy or significant side effects rare. Minimal decision-making was reported but those refusing biopsy often engage in complex analyses of benefits and consequences in light of their perceived susceptibility to PCa. A lack of knowledge and misunderstandings of PCa was identified. Men not responding to PSA testing often perceive themselves at low risk of PCa. CONCLUSIONS: Most men had positive experiences of PCa testing. Lack of knowledge and misunderstandings may influence decisions to be tested. Eliciting levels of understanding and providing appropriate information may improve uptake.

Urology and QOL 1204 /Erectile, Urinary and Bowel Function in Elderly Men Without Prostate Cancer Ida J. Korfage, Harry J. de Koning, Public Health, Monique Roobol, Fritz H. Schroder, Urology, Marie-Louise Essink-Bot, Public Health, Erasmus MC, University Medical Center Rotterdam, Rotterdam, the Netherlands AIMS: Since the introduction of prostate specific antigen (PSA) testing, prostate cancer is frequently diagnosed in an early stage. Surgery and radiotherapy are the most often applied primary therapies. Prostate cancer specific questionnaires, e.g. the UCLA EPIC, are available to assess post-treatment sexual, urinary and

A-51

and exploratory factor analyses regrouped the OAB-SAT-q items into three 3-item subscales (Satisfaction, Side Effects, Endorsement) and two individual items (Convenience, Preference). There was very little missing data across items (range 0-2.1%). Subscale Cronbach's alphas were 0.84-0.95. Correlations among the OAB-SAT-q Satisfaction and Endorsement subscales and OAB-q subscale change scores were moderate and significant (range 0.17-0.44; all p<0.01). The Side Effects subscale was significantly correlated with the number of treatment-related adverse events (r=0.25; p<0.01) and discriminated between patients with and without dry mouth and also patients with and without constipation. The three subscales and the Convenience item significantly discriminated between completers and non-completers. CONCLUSIONS: In a sample of 375 patients with OAB, the OAB-SAT-q demonstrated good psychometric properties, including internal consistency reliability and concurrent and discriminant validity.

1588 /Quality of Life Parameters of Urinary Function, International Prostate Symptom Scores, and Time to Continence are Affected by Degree of Neurovascular Bundle Preservation Following Radical Prostatectomy Christopher B. Dechet, Urology, University of Utah, Salt lake City, Utah, Karen J. Valentine, Oncology, Intermountain Healthcare, Salt lake City, Utah AIMS: To prospectively study the effects of neurovascular bundle preservation on time to continence, international prostate symptom scores and urinary function following radical prostatectomy using validated Quality of Life questionnaires. METHODS: Ninety-one patients (age range: 47 to 79 years) undergoing radical prostatectomy were prospectively studied using the UCLA Prostate Cancer Index and the International Prostate Symptom Score (IPSS). Questionnaires were completed pre-operatively, 3, 6, 9, 12 and 18 months following radical prostatectomy. Patients were divided in 3 groups (bilateral, unilateral and non-nerve sparing) based on degree of neurovascular bundle preservation. Mean urinary function and bother scores, IPSS scores and pad use were analyzed for each group and for differences between the three groups. RESULTS: All groups experienced an improvement in mean lower urinary tract symptoms as measured by IPSS scores. The magnitude of change was greatest for the nervesparing group. The bilateral nerve-sparing group achieved the lowest (best) scores overall (p=0.023). Mean urinary function and mean urinary bother scores over time were significantly different between the groups, p<0.001 and p=0.004. The bilateral nerve-sparing patients achieved final continence earlier than the others. In order to ensure that nerve-sparing was appropriately performed, sexual function scores were analyzed showing significant differences in mean sexual function with good preservation of sexual function in the bilateral nerve-sparing group. CONCLUSIONS: Despite earlier detection of prostate cancer and higher rates of organ confined disease, nervesparing prostatectomy is performed in the minority of cases due to its technical difficulty. Aside from the benefits in sexual function, our results suggest that nerve-sparing prostatectomy improves urinary function and time to continence and should be performed whenever possible.

1784 /Quality of Life and Patient Satisfaction Among Women with Urinary Incontinence: A Prospective Study John Wei, Rodney L. Dunn, Urology, John O. DeLancey, Gynecology, Edward J. McGuire, Urology, University of Michigan, Ann Arbor, MI AIMS: Patient satisfaction is the ultimate confirmation of quality care. The objective of this study was to determine if changes in urinary incontinence severity are reflected in patient satisfaction with physician manner and skill, outcome, access and information. METHODS: The Incontinence Symptom Index (ISI) and the Service Satisfaction Scale (SSS-16) are validated measures. 77 women with a chief complaint of urinary incontinence were enrolled during their initial visit and followed prospectively. The 10-item ISI, measuring 3 dimensions of urinary incontinence, was completed at baseline and again 3 months later along with the SSS-16. Changes in ISI scores were correlated to the satisfaction scores. RESULTS: With a mean age of 57.1 years, 48%, 20%, and 32% of women presented with stress, urge and mixed urinary incontinence, respectively. 47.3% were treated either pharmacologically, surgically or with behavioral therapy. A decrease in ISI domain scores (better health status) was associated with higher overall satisfaction as well as satisfaction with physician manner-skill, outcome and information (table). CONCLUSIONS: An improvement in urinary health status as measured by the ISI is associated with better satisfaction in multiple dimensions.

1352 /Psychometric Validation of the Overactive Bladder Satisfaction with Treatment Questionnaire Mary Kay Margolis, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Annamaria Cerulli, Rinat Ariely, Kristijan H. Kahler, US CD&MA / Health Economics & Outcomes Research, Novartis Pharmaceuticals Corporation, East Hanover, NJ, Kathleen M. Fox, ., Strategic Healthcare Solutions, LLC, Monkton, MD, Ray Hsieh, Karin Coyne, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD

Correlation of urinary outcome with satisfaction (*= p<0.05) MannerDifferen- Total SSS Skill ces in: Score score ISI Total -0.31* -0.32* score Stress -0.27* -0.27* score Urge -0.22 -0.29* score Pad Use -0.28* -0.23 score Bother -0.29* -0.27* score

AIMS: Evaluate the reliability and validity of the Overactive Bladder Satisfaction with Treatment Questionnaire (OAB-SAT-q). METHODS: A secondary validation analysis for the OAB-SAT-q, a new 11-item disease-specific treatment satisfaction questionnaire, was planned using data from a multicenter study of darifenacin alone or in combination with a Behavioral Modification Program for overactive bladder (OAB). Participants completed the OAB-q and a 3-day bladder diary for efficacy measurement at baseline and Week 12, and the OAB-SAT-q once at Week 12. Internal consistency reliability was assessed by Cronbach's alpha. Spearman correlations between the OAB-SAT-q and OAB-q change scores were used to assess concurrent validity. Discriminant validity was assessed among various patient subgroups using general linear models. RESULTS: Of the 395 patients in the intent-to-treat population, 375 patients completed the OAB-q at both assessments and OAB-SAT-q at Week 12 and were included in these analyses. 89% were female; mean age was 57.8+13.7 years; and the majority (89%) were Caucasian. Item

A-52

Outcome Access Informatio score score n score -0.30*

-0.10

-0.29*

-0.31*

-0.07

-0.24

-0.15

-0.03

-0.27*

-0.29*

-0.17

-0.21

-0.32*

-0.16

-0.19

unstable. Validity was considered by examining other HRQL outcomes for similar patterns, and by comparison with results from a smaller data set (n=130) which included the more traditional thentest. Validation results revealed that among the 107 persons (83%) with stable residuals over time, the magnitude of the recalibration using the then-test was -11 (95% CI: -14 to -8). For the 14 people belonging to the group where the rating of HRQL with respect to predicted value dropped, this value was -25 (range: -60 to 0); 7 people belonged to the group with a rise in HRQL with respect to predicted, and their then-test recalibration was +22 (range -10 to 40). CONCLUSIONS: This type of analysis could be used to stratify data, where persons with small residual fluctuations over time would form one strata, those with a different pattern of residuals would form other strata. An examination of mean change and variability of change within strata would provide information for the purposes of interpretation of change for the whole group.

Item Response Theory 2 1619 /Rasch Analysis of The VFQ-25 Maaike Langelaan, Ruth M. van Nispen, Ger H. van Rens, Ophthalmology, VU University Medical Center, Amsterdam, The Netherlands AIMS: The Visual Functioning Questionnaire (VFQ-25) is one of the most widely used measures of vision-related quality of life (VRQoL). To qualify as a measurement, the instrument must produce an interval. The objectives of this study are to obtain the interval scaling of the VFQ-25 by Rasch analysis, and to analyze VFQ-25 items and response categories to determine how well they measure the construct of VRQoL. METHODS: The study population consisted of 135 patients of a follow up study on the outcome of rehabilitation for visually impaired adults between 18 and 65 years old. They represent the whole range of visually impaired patients: from near normal vision to no light perception. Six patients were not able to participate. The VFQ-25 was administered by interview and consists of 25 items that are scored on an ordinal rating scale. Rasch analysis was used to convert ordinal data to interval scales. We estimated the visual ability for each person (person measure) and the required visual ability for each item (item difficulty). Visual acuity and visual fields were assessed for all patients. The relationships between visual acuity, visual field and visual ability were analyzed by linear regression analyses. RESULTS: Three items on driving were eliminated from the analyses, because only three patients were able to drive. Response categories were not being used with similar frequencies. Therefore, the response categories were collapsed. The change in rating scale resulted in slight improvement in person separation (from 4.77 to 5.68) and slight deterioration in item separation (from 2.30 to 2.17). Item reading normal print was too difficult for even the most able patients. This item could be removed, as it did not add any information. Regression analyses showed that persons are distributed from most to least disabled as would be expected from severity of visual impairment. CONCLUSIONS: Rasch analysis of the VFQ-25 confirms the unidimensionality of the instrument. However, the VFQ-25 could be improved through the removal of items with poor fit statistics. Until this is addressed, VFQ25-scores must be interpreted with caution.

1349 /Validation of a Computer-Adaptive Test for the Assessment of Depression Herbert Fliege, Psychosomatic Medicine and Psychotherapy, Charité Universitätsmedizin Berlin, Berlin, Germany, Otto B. Walter, Institute of Psychology IV Statistics and Methods, University of Münster, Münster, Germany, Janine Becker, Quality Metric, Waltham, MA, Rüya-Daniela Kocalevent, Psychosomatic Medicine and Psychotherapy, Charité Universitätsmedizin Berlin, Berlin, Germany, Matthias Rose, Quality Metric, Waltham, MA, Burghard F. Klapp, Psychosomatic Medicine and Psychotherapy, Charité Universitätsmedizin Berlin, Berlin, Germany AIMS: In earlier studies a German Computerized Adaptive Test to assess depression (D-CAT) has been developed and simulated on the basis of Item Response Theory (IRT) using real patients' data from a Psychosomatic Medicine setting. The present study aims at testing the feasibility of the D-CAT, validating the tool, and investigating screening cutoff-scores for depression across samples. METHODS: The D-CAT, supplied by an item bank of 64 items, was administered on pocket PCs to n=126 consecutive patients suffering from psychosomatic and other medical conditions, n=78 with depression, and n=200 healthy subjects. Items were adaptively presented to the patients until a predefined stopping rule (standard error less/equal 0.32) was reached. For validation purposes, the HADS, CES-D, and BDI were administered. RESULTS: The D-CAT was fast to complete (mean 74 seconds) and well accepted by the test takers. The precise IRT-based theta score could be estimated after an average administration of 6 items (SD=2.5, range: 4-18 items). In 95% of the cases 10 items or less were needed for a reliable theta score estimate. Overall, only 27 of the total item bank were used, four items accounted for 50% of all item administrations. Correlations between the D-CAT and the HADS, CES-D, and BDI were moderate (r=0.68 to 0.77). Exploring different scores for depression screening, we developed a cut-off score of theta=-0.5 with diagnostic sensitivity (86%) and specificity (also 86%) comparable to the HADS (89% / 81%). CONCLUSIONS: Overall, the Depression-CAT proved an economic, well accepted, reliable and valid tool performing well across different healthy and non-healthy samples. Diagnostic sensitivity/specificity is comparable to other depression measures, while the CAT is shorter and more precise. Item usage, however, raises questions of a future balancing of the item selection for content.

1545 /Using Latent Trajectory Analysis to Identify Response Shift at the Individual Level Nancy E. Mayo, Susan Scott, Clinical Epidemiology, McGill University Health Center, Montreal, QC, Canada, Nandini Dendukuri, Clinical Epidemiology, McGill University Health Center, Montreal, Quebec, QC, Canada, Sara Ahmed, Clinical and Health Information Research Group, McGill University Health Center, Montreal, QC, Canada AIMS: A statistical method for identifying response shift at the individual level would aid in interpreting change scores in situations where response shift is likely and design options either have not been used or are impractical. We explored whether a longitudinal comparison between predicted and reported health-related quality of life (HRQL) could be used as a method of identifying subjects who experienced response shift. METHODS: Data from a 1-year study of 678 persons with stroke were used to develop a predictive growth model of HRQL from physical, mental and cognitive measures. We hypothesized that large fluctuations between reported and predicted HRQL over time would identify individuals with a response shift. Subject-centered residuals were created and latent trajectory analysis was used to identify homogeneous groups of persons. RESULTS: Seven groups based on direction and timing of changes in selfreported HRQL with respect to predicted were identified: 67% with little fluctuation; 13% with deterioration; 14% with rises; and 4%

1656 /An Item Response Theory Model Analysis to Evaluate the Dimensionality of the EQ-5D Across Six Countries Oliver H. Günther, Health Economics Research Unit, Herbert Matschinger, Psychiatry, Hans-Helmut König, Health Economics Research Unit, University of Leipzig, Leipzig, Germany

A-53

AIMS: Within the framework of economic evaluation, health economists are interested in constructing a meaningful health index consistent with individual or societal preferences. One way to derive such an index bases on the EQ-5D descriptive system of health related quality of life (HRQOL). This study analyzed how well the EQ-5D reflects one latent construct of HRQOL and how large the potential impact of measurement variance is with respect to different countries. METHODS: Data came from the European Study of the Epidemiology of Mental Disorders (ESEMeD), a cross-sectional survey of a representative random sample (N=21,425) in Belgium, France, Germany, Italy, the Netherlands and Spain. The dimensionality of the construct was evaluated by means of oneparameter and two-parameter Item Response Theory (IRT) models. Models were estimated by a generalized random intercept model using the STATA module 'gllamm'. The potential impact of parameter variance was evaluated by Differential Item Functioning with respect to the six countries. RESULTS: Results showed remarkable item-bias for the item-difficulties and for the itemdiscrimination. A two-parameter IRT model that is unidimensional with respect to the countries fits considerably better for all countries, if the item 'anxiety/depression' is omitted. CONCLUSIONS: If the mental component represented by the item 'anxiety/depression' is included, the EQ-5D becomes a two-dimensional construct. Consequently, more items are needed to make the mental component accessible to psychometric analysis.

assessing depressed patients, the WHOQOL-Bref may need to be analysed differently.

Neurology and its Relationship to QOL: Recent Studies 1716 /Assessment of Health-Related Quality of Life in Persons After Traumatic Brain Injury - Development of the Qolibri, A Specific Measure Nicole von Steinbuechel, Dept. of Medical Psychology and Medical Sociology, Georg-August-University Goettingen, Goettingen, Germany, Stefan Hoefer, Dept. of Medical Psychology and Psychotherapy, University Clinic Innsbruck, Innsbruck, Tirol, Austria, Corinna Petersen, Institute of Medical Psychology and Psychotherapy, University Clinic Hamburg - Eppendorf, Hamburg, Germany, Tanja Lischetzke, Faculty of Psychology and Educational Science, University of Geneva, Geneva, Switzerland, Monika Bullinger, Institute of Medical Psychology and Psychotherapy, University Clinic Hamburg - Eppendorf, Hamburg, Germany, QOLIBRI Group, Dept. of Medical Psychology and Medical Sociology, Georg-August-University Clinic, Goettingen, Germany AIMS: Health-related quality of life (HRQOL) is a relatively new outcome variable in the field of traumatic brain injury (TBI). Validation of generic, cross-culturally (cc) administered HRQOL measures in persons after TBI is not yet well established. Diseasespecific HRQOL instruments do not exist in an international context. The objective here is to present the QOLIBRI [QOLIBRI (Quality of Live after Brain Injury] approach in cc development of a specific HRQOL measure, i.e. cc questionnaire construction, development, translation und psychometric testing. METHODS: 1568 patients after TBI from 10 countries and 8 languages filled out a preliminary version of the QOLIBRI assessing HRQOL within six domains (physical condition, thinking activity, feelings and emotions, functioning in daily life, relationships and social/leisure activities, current situation and future prospects). In the sample, psychometric testing was performed in order to obtain a psychometrically stable reduced item version. RESULTS: After testing the final version of the QOLIBRI now integrates disease-specific issues of TBI patients on four satisfaction scales (29 items) i.e. cognition, emotion and selfperception, daily life and mobility, and the social dimension and with four overall items. Additionally to the satisfaction scale also the symptom scale (nine items) with two subscales (negative feelings, and mobility) shows good to excellent psychometric qualities for item and scale level. CONCLUSIONS: In TBI patients, generic and disease-specific aspects of HRQOL can now be assessed with a new measure of adequate psychometric quality, applicable across different populations and cultural conditions. The QOLIBRI is a promising instrument for valid patient-centered specific outcome evaluation after TBI.

1747 /WHOQOL-BREF Validity in Depressed Adults: What we Can Learn by Integrating Classical Test Theory and Item Response Theory Neusa S. Rocha, Marcelo Fleck, Psychiatry,Universidade Federal do Rio Grande do Sul, Hospital de Clinicas de Porto Alegre, Porto Alegre, RS, Brazil, Michael J. Power, Clinical and Health Psychology, University of Edinburgh, Edinburgh, United Kingdom, Donald Bushnell, Health Research Associates, Seattle, WA, LIDO Group, HRA, Seattle, WA AIMS: The WHOQOL-Bref is a general quality of life measure consisting of 26 items or facets grouped into four domains related to quality of life (Physical Health, Psychological, Social Relationships, and Environment). While we know the WHOQOL-Bref is a good measure for discriminating depressed patients, further research on the applicability of its original structure is of interest for this population. This study was designed to search for a valid WHOQOL-Bref structure for depressed adults using LIDO (Longitudinal Investigation of Depression Outcomes) baseline data (n=1193) from 6 countries worldwide. METHODS: To test our hypothesis that for depressed people the WHOQOL-Bref may have a different structure, we conducted an analysis using: 1) Confirmatory Factorial Analysis (CFA) for each domain, 2) Rasch analysis, and 3) multidimensional item approach. RESULTS: The CFA provided a solution of 4 domains similar to the original version but including different composition of facets,Physical Domain (CFI=.98;GFI=.99;RMSEA=.06), Psychological Domain (CFI=.96;GFI=.99;RMSEA=.05), Social Domain(CFI=.99; GFI=.996; RMSEA=.03), finally, Environment Domain (CFI=.92; GFI=.98; RMSEA=.06). Searching for a better solution based on a Rasch model resulted in a 4-domain solution, as well, Psychological Domain (x2=58.99; P=0.62), Physical Domain (x2=70.53; P=0.06), Environment Domain (x2=45.78; P=0.95) and Social Domain (x2=42.19; P=0.22). All four domains generated by Rasch analysis could be validated by a CFA model, so this procedure identified a common structure by integrating both methods. Once using Rasch analysis, the question about multi-dimensionality of the construct is not accounted, we will also present results using multidimensional item approach. CONCLUSIONS: Our findings point that when

1328 /The Influence of Treatment Complications on Patient Reported Outcomes in PD Claire L. Shenton, Clare Dowding, Sam Salek, Pharmacy, Cardiff University, Cardiff, UK, Sandip Raha, Integrated Medicine, Louise Ebenezer, Princess of Wales Hospital, Bridgend, UK, Elizabeth Morgan, Rookwood Hospital, Cardiff, UK, Pandanda Pooviah, Zahed Ikram, Geriatrics and General Medicine, Dwarak Sastry, University Hospital of Wales, Cardiff, UK AIMS: The aim of this study was to assess the impact of complications of therapy on health-related quality of life (HRQoL) of Parkinson's disease (PD) patients using self-administered generic and PD-specific measures, United Kingdom Sickness Impact Profile (UKSIP) and Parkinson's Disease Questionnaire-39 (PDQ-39). METHODS: Patients (n=159) were recruited into the study from 2

A-54

outpatient movement disorder clinics in South Wales. They were asked to complete 2 HRQoL questionnaires and clinicians evaluated them on the Hoehn and Yahr (H&Y) scale and part IV of the Unified Parkinson's Disease Rating Scale (UPDRS), Complications of Therapy. Analyses of results were performed using non-parametric tests; Mann-Whitney U or Kruskal-Wallis tests and Spearman's Rank correlation. RESULTS: The mean patient age was 72 years (median 74, range 41-92), with 98 (61.6%) patients being male and 133 (83.8%) patients between H&Y stage 1.5 (mild) and 3 (moderate). Mean PDQ-39 and UKSIP scores were 28.9 (median 27.6) and 20.5 (median 17.7) respectively and 80 (50.3%) patients had no complications of therapy. Patients experiencing complications of therapy had significantly higher PDQ-39 (p=0.001) and UKSIP (p=0.003) scores (indicating poorer health) compared to patients without complications. PDQ-39 scores were significantly higher for patients with dyskinesias (p=0.012), motor fluctuations (p<0.001) or sleep disturbances (p=0.008). Similarly, UKSIP scores were significantly worse for patients experiencing dyskinesias (p=0.004) and motor fluctuations (p=0.001) compared to patients only displaying other types of complications. HRQoL scores significantly increased as disease severity (H&Y) increased. There was no significant difference between HRQoL scores for males and females and correlation was poor between age and HRQoL scores. CONCLUSIONS: Patients with treatment complications have significantly poorer HRQoL compared to those with no complications as measured by the PDQ-39 and UKSIP. In particular, there is significant negative impact on HRQoL for patients experiencing motor fluctuations and dyskinesias.

1269 /Understanding HRQL After Stroke: Help from Structural Equation Modeling (SEM) Sharon L. Wood-Dauphinee, Physical and Occupational Therapy; Epidemiology, Nancy E. Mayo, Medicine; Physical and Occupational Therapy, McGill University, Montreal, Quebec, Canada, Susan C. Scott, Clinical Epidemiology, McGill University Health Center, Montreal, Quebec, Canada AIMS: Studies have examined HRQL after stroke but none has used SEM to better understand relationships among observed and latent variables. We set out to test the Wilson and Cleary model of HRQL with personal characteristics, physiologic variables, clinical symptoms, function and HRQL. METHODS: 678 Canadians were recruited in acute care and interviewed by trained interviewers over the phone at 1, 3, 6 and 12 months post-stroke. SEM was used to test the theoretical model against observed data at 3 months. RESULTS: The best fitting model was obtained before adding HRQL (SatorraBentlerX2=256, 189 df; RMSEA=0.023, CFI=1.0). Strong association was found between stroke severity and strength, with paths to incontinence, communication and memory also significant. Comorbidity influenced strength and pain. Incontinence, vitality, vision, memory and comorbidity worsened with age, as did physical, cognitive and social functions. Strong association was found between strength and physical function, also influenced by incontinence, pain, memory and vision. Cognitive function was influenced by memory and communication. Social function increased most with strength and mental health but also with memory, vitality and vision. Adding HRQL, with paths from function but only indirect paths from symptoms produced a reasonable model (X2=457, 233 df; RMSEA=0.038 CFI=0.99) that explained 47% of the variation in HRQL. HRQL increased with higher function, particularly social. Better fit was found by allowing symptoms to impact directly on HRQL, with the proportion of HRQL explained increasing to 65% (X2=397, 230 df; RMSEA=0.033). There were direct paths to HRQL via pain, vitality and memory and indirect paths from strength and mental health through role function, the only function which remained significantly associated with HRQL. CONCLUSIONS: While relationships confirmed clinical intuition, SEM allowed us to better quantify and compare complex relationships. Associations between physiology, symptoms and function may be structured as by Wilson and Cleary but this may not be the best model to describe HRQL.

1536 /Relationship Between Satisfaction with Restless Legs Syndrome (RLS) Treatment and Time Since Diagnosis Martin C. O'Leary, GlaxoSmithKline, Greenford, Middlesex, UK, Michael O. Calloway, USP Health Outcomes, GlaxoSmithKline, Research Triangle Park, North Carolina AIMS: RLS is a chronic disorder characterized by an irresistible urge to move the legs. Requip® (ropinirole)is FDA-approved for the treatment of moderate-to-severe primary RLS. This US survey investigated patient satisfaction with ropinirole compared with other treatments for RLS and whether this was affected by length of time since diagnosis. METHODS: Adults from the National Health and Wellness US studies (2005/6) and Late Speed Research ailment panel were invited to complete a web-based survey (Apr 2006) investigating quality of life and health outcomes related to RLS treatments. Patients with a diagnosis of RLS and taking at least one prescription medication for RLS for 2+ months were eligible for inclusion (n=348). In this analysis patients were categorized as receiving ropinirole only (ropinirole-only group, n=58, 17%), or prescription medication(s) other than ropinirole (other Rx group, n=65, 19%). Demographic data and time since diagnosis of RLS were collected. Treatment satisfaction was measured on a 7-point Likert scale (1=extremely satisfied, 7=extremely dissatisfied). Ordered logistic regression was used to determine whether time since diagnosis affected satisfaction with treatment. RESULTS: The mean (SD) age of respondents was 50.3 (11.5) years and the mean (SD) duration since RLS diagnosis was 3.6 (4.1) years. Patients in the ropiniroleonly group were significantly more satisfied with their treatment than those in the other Rx group (odds ratio [OR]: 0.426; 95% CI: 0.192, 0.944; p=0.036); more patients in the ropinirole-only group were extremely satisfied or very satisfied with their treatment (48.3%) than in the other Rx group (35.4%). Satisfaction with ropinirole treatment was not affected by time since diagnosis of RLS (OR: 0.053; 95% CI: -0.156, 0.037; p=0.227). CONCLUSIONS: Patients with RLS taking ropinirole were significantly more satisfied with their treatment than those taking other prescription medications. Patient satisfaction with ropinirole treatment was not affected by time since diagnosis. This survey was supported by GSK R&D.

1568 /Psychometric Properties of the Multiple Sclerosis Quality of Life Inventory in a Minimally Affected Group: Results from the Champions Trial Deborah M. Miller, Medicine, Cleveland Clinic Foundation, Cleveland, OH, Craig Kollman, Biostatistics, Jaeb Center for Health Research, Tampa, FL, R. Phillip Kinkel, Neurology, Harvard Medical School, Boston, MA AIMS: The negative impact of multiple sclerosis (MS) on HRQoL is well documented in persons with moderate to severe clinically definite MS (CDMS). HRQoL consequences for those at risk for or recently diagnosed with MS are not understood. We investigated the psychometric property of Multiple Sclerosis Quality of Life Inventory (MSQLI), which includes the SF-36 and 9 disease specific scales in a group of patients 5 years after they are determined at risk for CDMS. METHODS: CHAMPIONS is an ongoing open label study for participants in the earlier, 2-year RCT CHAMPS study of the benefit of treatment for individuals at risk for CDMS. The MSQLI was administered to CHAMPIONS subjects 5 years after enrolling in CHAMPS. Internal consistency of the scales was determined using Chronbach's a. Scores >70% were considered acceptable. Floor and ceiling effects <15% were considered acceptable. RESULTS: 196 of 203 patients enrolled in CHAMPIONS

A-55

(97%) completed the MSQLI at their 5 year visit. 8 of 9 disease specific scales showed acceptable internal consistency, from .97 on the Modified Fatigue Impact Scale and the Modified Social Support Survey to .78 on the Bladder Control Scale. The Bowel Control Scale (a=.55) was not acceptable. Scales related to pain, sexual functioning, bowel and bladder control and vision had scores >15% at the top. Only the Mental Health Inventory had floor effects >15%. The Modified Fatigue Impact Scale, the Modified Social Support Scale and the Perceived Cognitive Deficits Questionnaire reached all three criteria for psychometric acceptability (internal consistency, floor effects, ceiling effects). The scales for Pain Effects, Sexual Satisfaction, Bladder Control, and Impact of Visual Functioning had adequate internal consistency and floor effects but unacceptable ceiling effects. The Bowel Control Scale and the Mental Health Inventory were acceptable only in terms of ceiling effects. CONCLUSIONS: These data provide preliminary evidence that the majority of the MSQLI scales are acceptable in this population.

AIMS: The use of preference-based generic instruments to measure Health Related Quality of Life of a general population or of individuals suffering from a specific disease has been increasing. However, there are several discrepancies in terms of utility results between preference measures. This study seeks to compare SF-6D and EQ-5D across subjects with asthma, chronic obstructive pulmonary disease, rheumatoid arthritis and cataracts and to investigate the differences in agreement between them. The main objective is to explore the eventual reasons for divergences found and to explore their implications. METHODS: Agreement between EQ5D and SF-6D health state classifications is assessed by correlation coefficients. The distribution of individuals reporting the lowest or highest health state in each instrument is analysed in terms of their report in the alternative instrument. Simple Correspondence Analysis was used to assess the agreement among the instrument s descriptive systems and to investigate similarities between dimensions modalities. Factor Analysis was used to identify the dimensions of health measured. A regression analysis was run in order to determine the relative contribution of each factor score and of the demographic characteristics to the global utility score. RESULTS: There is evidence for floor effects in SF-6D and ceiling effects in EQ-5D. Comparisons of mean indices found that SF-6D values exceed EQ-5D values. Agreement between both instruments was found to be higher, especially between similar dimensions: mobility, physical functioning and usual activities, pain and pain/discomfort, mental health and anxiety/depression, mental health and pain/discomfort, vitality and pain/discomfort and vitality and anxiety/depression. Differences in valuation methods and in scoring algorithms contribute to the main differences. CONCLUSIONS: To overcome their weaknesses, revisions of one or both instruments in their descriptive systems or on their scoring algorithm are necessary.

Preferences Research 2 1659 /Exploring the Relationship Between Health and Happiness: A Comparison Across Studies of Different Conditions Using the SF-36 and EQ-5D Clara W. Mukuria, John E. Brazier, Aki Tsuchiya, School of Health and Related Research, University of Sheffield, Sheffield, United Kingdom AIMS: Evidence suggests that health can have a considerable impact on happiness. The main objective of this study is to investigate the impact of different dimensions of health, as measured by the EQ-5D and SF-6D, on happiness, as measured by the happiness question in the SF-36. METHODS: The data used in this study pools the results from 15 studies undertaken in ScHARR and elsewhere. It consists of 12685 cases covering: asthma, chest pain, healthy older women, COPD, menopausal women, IBS, ITU patients, leg reconstruction, leg ulcers, lower back pain, osteoarthritis, trauma, sleep clinic, varicose veins and non-patients. Multivariate ordered probit analysis is carried out with the happiness item in the SF-36 as the dependent variable. Several models are investigated with demographic factors, health dimensions of EQ-5D and SF-6D, their preference-based single indices, and the medical conditions as independent variables. RESULTS: The results indicate that the most important dimensions of health negatively associated with happiness are the mental health (probability of reporting high happiness is lowered by -4.6% to 35.7% in SF-6D) and vitality (-4.3% to -49%). Social functioning limitations also have negative associations (-5.7% to -8.1%). Problems in physical functioning and role limitation due to physical health have an unexpected result as they indicate positive association with happiness (3% to 8.5% and 1.9% respectively) after controlling for other variables. Pain dimensions are not significant indicating there is no association between happiness and pain when controlling for other dimensions. Actual medical condition had an independent effect that was negative for some conditions such as osteoarthritis (6.1% and -46%). CONCLUSIONS: These findings rely on a comparatively crude measure of happiness, but they suggest that relative impact of the dimensions of EQ-5D and SF-6D on happiness differs from that suggested by general population valuations (e.g. Dolan, 1997). This may have important implications for policy.

1664 /Examing the Discrimination and Responsiveness of a Conditon-Specific Preference-Based Instrument: The Asthma Quality of Life Utility Index(AQL-5D) Yaling Yang, John Brazier, Aki Tsuchiya, HEDS, School of Health and Related Research, University of Sheffield, Sheffield, South Yorkshire, UK AIMS: There has been potential demand for an asthma-specific preference-based measure for calculating Quality Adjusted Life Years (QALYs), given that the generic measures such as EQ-5D, HUI and SF-6D may lack sensitivity in capturing changes in quality of life in patients with asthma. Based on the Asthma Quality of Life Questionnaire (AQLQ), a condition-specific preference-based utility index AQL-5D has been derived and a valuation survey carried out. The resulting utility scoring algorithm makes it possible to calculate asthma specific quality of life indices using existing and future AQLQ data sets. The purpose of this piece of analysis is to examine the psychometric properties of the AQL-5D indices in terms of discrimination and responsiveness. METHODS: The AQL-5D was applied to two datasets which cover a wide range of asthma patients. One dataset contains two asthma measures (AQLQ and the Newcastle Symptom Scores), and two generic measures (SF-36 and EQ-5D). This dataset was used to examine the discriminative ability of the AQLQ scores, EQ-5D indices, and AQL-5D indices by dividing patients into groups using the SF-36 generic health question and the Newcastle Asthma Symptom Score. Mean, standard deviation and effect sizes by patient groups were calculated for each instrument and compared. The second dataset contains AQLQ and mapped EQ5D indices at baseline and 12 weeks follow-up. Effect sizes by AQLQ scores, AQL-5D indices and mapped EQ-5D indices were calculated to test responsiveness over time. RESULTS: In the first dataset, AQL-5D indices, AQLQ scores, and EQ-5D indices all showed a decreasing trend following the decline of general health and asthma specific index. Standard deviations for AQL-5D values had

1562 /How Consistent is Preference Measurement? Pedro L. Ferreira, Centro de Estudos e Investigação em Saúde, Universidade de Coimbra, Coimbra, Portugal, Lara N. Ferreira, Department of Quantitative Methods, University of Algarve, Faro, Portugal

A-56

are consistently smaller than those for EQ-5D, and showed better discrimination ability between groups based on the asthma specific index grouping. Performance of AQL-5D indices was comparable to AQLQ overall scores. In the second dataset, effect size of AQL-5D indices over time were slightly smaller (0.51) compared with AQLQ (0.62), but larger than for mapped EQ-5D (0.52). CONCLUSIONS: AQL-5D showed good discriminative and responsive ability, although more studies are needed to reconfirm this conclusion.

prescribing clinicians. Patient preferences are revealed in the way patients make trade-offs across symptoms, side-effects of treatment, ability to do everyday activities, and the convenience and cost of taking regular medication. The aim of this study was to investigate patient preferences across these various attributes of treatment using a discrete choice experiment (DCE). METHODS: A DCE was embedded in a randomized clinical trial involving patients with mildmoderate persistent asthma. The present data were collected after patients had received 6 weeks of treatment with one of two drugs. Three choice options were presented: to continue with the current drug; to change to an alternative, hypothetical drug; or to take no preventive medication. Analysis used random parameter multinomial logit. RESULTS: Complete surveys were returned by 52 of the 58 patients randomized to the clinical trial (90% response rate). Each respondent completed 16 scenarios, giving a total of 832 choice observations. Choice analysis revealed that respondents valued their ability to participate in usual daily activities and sport, preferred minimal symptoms, and were less likely to choose drugs with sideeffects. Cost was also significant, but other convenience attributes were not. Demographic characteristics did not improve the model fit. CONCLUSIONS: This study illustrates how discrete choice experiments may be embedded in a clinical trial to provide insights into patient preferences.

1534 /Treatment Satisfaction with Ropinirole in Patients Who Have Discontinued Prior Restless Legs Syndrome (RLS) Therapies Michael O. Calloway, USP Health Outcomes, GlaxoSmithKline, Research Triangle Park, North Carolina, Angela Williams, Global Health Outcomes, GlaxoSmithKline, Martin C. O'Leary, Global Health Outcomes, GlaxoSmithKline, Greenford, Middlesex, UK AIMS: A Phase III, randomized, placebo-controlled trial TREAT RLS US found that Requip® (ropinirole) significantly reduced RLS symptoms compared with placebo. Ropinirole is the first FDAapproved treatment for moderate-to-severe primary RLS and has been on the market for 1 year. As health is a state of complete physical mental and social well-being and not merely the absence of disease or infirmity (Official Records of the WHO, no. 2:100),the aim of this US national survey was to assess the overall satisfaction with ropinirole in patients with RLS who had stopped a prior RLS therapy. METHODS: The RLS SELECT Study evaluated treatment satisfaction among US adults diagnosed with RLS using an extensive web-based survey. Patients were categorized as receiving ropinirole only, any prescription medication(s)other than ropinirole, or receiving over-the-counter (OTC) medication(s)only. Satisfaction with ropinirole and previous therapies (prescription and OTC) was assessed using a 7-point scale (1=extremely satisfied/convenient to 7=extremely dissatisfied/inconvenient). A difference of means t-test was used to compare satisfaction scores. RESULTS: In total 348 patients were surveyed. Of these 58 were, at the time of the survey, only taking ropinirole for RLS. Of these 23 (40%) had previously taken a prescription medication for RLS, 1 had been taking an OTC medication, and 34 (59%) were treatment naïve. The most reported reason for discontinuing prior RLS treatment was perceived lack of efficacy (31%). Of those patients currently receiving ropinirole only, and who had received a prior prescription treatment, mean satisfaction scores significantly favored ropinirole over their prior prescription treatment (mean [SD] score, 2.72 [1.39] and 4.10 [1.63] for ropinirole and prior treatment, respectively; t=4.20; df=55; p=0.0005). CONCLUSIONS: Patients currently taking ropinirole only for the treatment of RLS were significantly more satisfied with their treatment than they had been with their last prescription medication. This survey was supported by GSK R&D.

Surgery and QOL: New Developments and Insights 1316 Reliability of a Patient-Clinical Communication Form Maria Jose Santana, Public Health Sciences, University of Alberta, Edmonton, Alberta, Canada, David Feeny, Health Utilities, Inc., Dundas, ON, Canada AIMS: To assess the reliability of the communication form use to evaluate communication between patients and clinicians. METHODS: Data were collected from a study assessing the effects of including health-related quality of life measures in routine clinical care. To measure communication, the encounters between patient and clinician were audio tape recorded. The content of the recordings was analyzed with the aid of the communication evaluation form. This form is completed independently by two blinded raters for every patient-clinician encounter. It lists the health attributes included in the Health Utilities Index score card (emotion, self-care, pain, vision, hearing, speech, ambulation, dexterity and cognition). Raters listen to the conversations and count the number of issues discussed at every clinical encounter. The numbers of issues discussed are summed to produce a score. Inter-rater reliability is assessed by comparing the results obtained by each independent rater of the same encounter. Intra-rater or test-retest reliability is assessed by comparing the scores of the same rater for the same encounter at two different times. The interval of time between assessments was at least two weeks. Reliability is assessed with the intra-class correlation coefficient (ICC) based on a two-way random, parallel with absolute agreement model. Standards for interpreting agreement were taken from Nunnally JC (1978): poor agreement (ICC<0.40), moderate to good agreement (ICC=0.40-0.75) and excellent agreement (ICC>0.75). RESULTS: 227 recordings were analyzed. Inter-rater reliability: ICC for single measures was 0.70 (95%CI 0.61-0.76), for average measures among raters was 0.82 (95% CI 0.76-0.86). Agreement is moderate to good for single measures and excellent for average measures. Test-retest reliability: ICC for single measures was 0.85 (95% CI 0.79-0.90); for average measures was 0.92 (95% CI 0.880.95). Test-retest agreement was excellent. CONCLUSIONS: The inter-rater and intra-rater reliabilities of the communication form for average measures were excellent.

1539 /Managing Asthma: Patient Preferences and Trade-Offs for Symptoms, Side-Effects, Daily Activities, Convenience and Cost of Medication Madeleine T. King, CHERE, University of Technology, Sydney, NSW, Australia, Jane Hall, CHERE, UTS, Sydney, NSW, Australia, Emily Lancsar, Business School (Economics), University of Newcastle upon Tyne, Newcastle, England, UK, Denzil Fiebig, School of Economics, University of New South Wales, Sydney, NSW, Australia, Ishrat Hossain, CHERE, Jordan Louviere, Centre for the Study of Choice, UTS, Sydney, NSW, Australia, Helen Reddel, Cooperative Research Centre for Asthma, Christine Jenkins, Woolcock Institute of Medical Research, University of Sydney, Sydney, NSW, Australia AIMS: Effective control of asthma requires regular preventive medication. Poor medication adherence suggests that patient preferences for medications may differ from the concerns of

A-57

versions of the questionnaire were piloted to clarify ambiguities in item wording and confirm appropriateness, acceptability, and completion time (n=45). RESULTS: A conceptual model for the impact of breast surgery was developed that included the following: (1) body image, (2) psychological functioning, (3) social functioning, (4) physical functioning (5) sexuality and (6) process of care. The process of item generation led to a separate module for each surgery type: 1) breast augmentation, 2) reconstruction, and 3) reduction. Pretesting and expert review of the three modules resulted in minor wording changes. We named this modular PROM the Memorial Sloan-Kettering Cancer Center (MSKCC) BREAST-Q. CONCLUSIONS: The MSKCC BREAST-Q will provide essential information about the impact and effectiveness of breast surgery from the patients' perspective. It will complement current clinical outcome measures and provide individual surgeons with an important tool to support clinical audits and ongoing practice improvement. The MSKCC BREAST-Q is currently being tested in large heterogeneous patient populations in the US and Canada to determine reliability, validity, and responsiveness.

1498 Health Related Quality of Life Following Elective Hip or Knee Replacement in the Elderly - A One-Year Prospective Cohort Study Anita E. Zahlten-Hinguranage, Orthopeady, University of Heidelberg, Heidelberg, Germany, Desiderius Sabo, Orthopeady and Traumatology, Klinikum Karlsbad-Langensteinbach, Karlsbad, Germany, Dominik Parsch, Felix Zeifang, Orthopeady, University of Heidelberg, Heidelberg, Germany AIMS: To assess changes in patient perceived health-related quality of life (QoL)and functional status after elective hip or knee replacement in elderly. METHODS: During a one year period 87 consecutive enrolled subjects aged >75 were prospectively studied. 51 total hip replacements and 36 knee replacement. Baseline measures included demographic factors, comorbidities (ASA Score). The Nottingham Health Profile (NHP) and the FFbH scores were completed by the patients preoperatively (T0) and one year (T1) after surgery. RESULTS: The mean age of the patients was 79.0 years: 59 patients were aged 75-80 and 27 were over 80. The female-to-male ratio was 2.6:1. More than 75 % of the patients were classified as ASA II or III. In 5 cases minor complications occurred. The FFbH scores improved from 43% to 71% at one year. In comparing baseline and one year results of the NHP, significant improvements can be seen in the dimensions of energy, pain, sleep and physical mobility. Based on the mean changes of NHP scores health improvements for patients aged 75-80 were large for pain and physical mobility with a Cohen s d of 1.7 and 1.8, respectively and moderate for energy with 0.6. Likewise, effect sizes of 1.6 and 1.1 indicated great health improvements for pain and physical mobility for patients >80, but no improvement for energy (d < 0.1). While for those patients with good baseline scores no differences occurred for emotional reaction and social isolation between T0 and T1 (d=0.1) for those with low preoperative scores a clear positive effect could be revealed (d=1.5)after arthroplasty. CONCLUSIONS: Despite functional restriction elective arthroplasty for osteoarthritis leads to substantially improved quality of life in elderly. This is mainly due to an increase in mobility and pain relief after surgery. It could be shown, that QoL life was positively affected for those patients with unsatisfactory social contacts. The decision to surgically repair osteoarthritis by arthroplasty in the older population is supported by this study.

1515 /The Impact of Hormone Replacement Therapy on Menopausal Symptoms in Younger High-Risk Women Following Prophylactic Salpingo-Oophorectomy Joanna B. Madalinska, Psychosocial Research and Epidemiology, Marc van Beurden, Gynecology, Eveline M. Bleiker, Psychosocial Research and Epidemiology, The Netherlands Cancer Institute, Amsterdam, The Netherlands, Heiddis B. Valdimarsdottir, Oncological Sciences, Mount Sinai School of Medicine, New York, New York, Neil K. Aaronson, Psychosocial Research and Epidemiology, The Netherlands Cancer Institute, Amsterdam, North Holland, The Netherlands AIMS: Preventive health strategies for women at increased hereditary risk of ovarian cancer include gynecologic screening (GS) and/or prophylactic oophorectomy (PBSO). Hormone replacement therapy (HRT) is often prescribed to compensate for post-surgical endocrine deficiencies. This study examined the impact of HRT use on levels of endocrine symptoms and sexual functioning among premenopausal women who have undergone PBSO. Comparisons were made with similar women undergoing GS. METHODS: Questionnaire data on endocrine symptoms and sexual functioning were obtained from 450 premenopausal, high-risk women who had participated in this nationwide, cross-sectional, observational study. RESULTS: Thirty-six percent of women had undergone PBSO and 64% had opted for GS. In the PBSO group, 47% of the women were current HRT users. They reported significantly fewer vasomotor symptoms than non-users (p < .05). However, compared to premenopausal women undergoing GS, oophorectomized HRT users were more likely to report vasomotor symptoms (p < .01). HRT users and non-users reported comparable levels of sexual functioning. Compared to women in the GS group, oophorectomized HRT users reported significantly more sexual discomfort due to vaginal dryness and dyspareunia (p < .01). CONCLUSIONS: Although HRT has a positive impact on surgically-induced vasomotor symptoms, it may be less effective than is often assumed. Symptom levels remain well above those of premenopausal women undergoing screening, and sexual discomfort is not alleviated by HRT. Physicians need to provide younger high-risk women considering PBSO with realistic information about both benefits and drawbacks of this preventive strategy, including information about premature menopause and HRT.

1226 /Measuring Quality of Life in Breast Surgery: Content Development of a New Modular System to Capture PatientReported Outcomes (The MSKCC Breast-Q) Andrea L. Pusic, Plastic and Reconstructive Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, Anne Klassen, Pediatrics, University of British Columbia, Vancouver, British Columbia, Canada, Stefan Cano, Neurology, University College London, London, England, United Kingdom, E D. Collins, Carolyn Kerrigan, Plastic Surgery, Dartmouth-Hitchcock Medical Center, Lebanon, NH, Peter G. Cordeiro, Plastic and Reconstructive Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY AIMS: Patient-reported outcomes (PROs) in breast surgery are increasingly important to clinical research endeavors. While many measures (PROMs) have been used in cosmetic and reconstructive breast surgery, few are conceptually grounded in patient perceptions of the impact of surgery or have been validated against appropriate psychometric benchmarks. The objective of this study was to develop a new measure of the impact of breast surgery that can be easily incorporated into routine practice. METHODS: We conducted a qualitative study with three components: 1) extensive literature review; 2) consultation with expert panels (i.e., oncologic surgeons, breast clinic nurses and psychologists); 3) in-depth, semi-structured patient interviews with 48 breast surgery patients (n=20 breast reconstruction, n=15 reduction and n=13 augmentation). Preliminary

1561 /Quality of Postoperative Recovery Pereira J. Maria, Anestesiology, Centro Hospitalar de Lisboa, Zona Centro, Lisboa, Portugal, Pedro L. Ferreira, Centro de Estudos e

A-58

Investigação em Saúde, Universidade de Coimbra, Coimbra, Portugal

with worse renal function was consistently worse along most of the HRQOL domains (e.g.: median (interquartiles range, IQR) group 1: 81 (31); group 2: 75 (38); group 3: 69 (42) for burden of kidney disease, p<0.001, ANOVA). In multivariate analysis eGFR was a significant and independent predictor of different HRQOL domains after adjusting for serum albumin, age, gender and the number of comorbidities. CONCLUSIONS: Renal function is an independent predictor of HRQOL in kidney transplanted patients.

AIMS: The main objectives of this study were: 1.The validation of the Portuguese translation of the QoR-40 questionnaire on the quality of the recovery of neurosurgical patients submitted to general anesthesia; 2.Evaluation of QoR-40 total scores during hospital stay and its relationship with VAS pain and VAS general health status scales. 3.Determination of factors that are predictive of good or bad anesthesia recovery. METHODS: The Australian questionnaire QoR-40 was translated by using the forward-backward translation method. After a pilot test and a face validation of the questionnaire, we planned a prospective and longitudinal study in which QoR-40 was applied to neurosurgical patients. The study population included 200 patients recruited during 6 months. One hundred and fifty patients were included in the study, being 150 excluded and fifty missed. Only 5 patients refused to participate (3%) with a high compliance (87%). Of the 150 patients included in the study, 131 answered all three questionnaires. The statistical analyses performed included factor analysis, parametric tests such as t-Student tests for paired samples and independent samples, Pearson and Spearman correlations, Chronbach s a, and multiple regressions. RESULTS: The Portuguese version of the QoR-40 questionnaire obtained very good results when submitted to psychometric testing. Total QoR-40 scores were lower at 24 hours and increased afterwards. They were lower in column patients when compared to craniotomy patients. The most important factors showed influence (58.5%) on the total QoR40 scores, namely VAS general health at 72 hours, the VAS pain at 72 hours and the post-operative complications, as well as the female gender. CONCLUSIONS: The obtained Portuguese version of the QoR-40 questionnaire is a simple, inexpensive ands sensitive tool to assess the quality of the recovery after a general anesthesia. It can improve clinical anesthesiology, by monitoring the efficacy of anesthesiologists’ interventions in areas like pain, morbidity and post-operative complications.

1174 /Prevalence of Depression and Association with Quality of Life in Predialysis Patients Agnes Z. Kovacs, Andras Szentkiralyi, Miklos Z. Molnar, Maria Czira, Eszter Vamos, Agnes Koczy, Lilla Szeifert, Rezso Zoller, Andrea Dunai, Istvan Mucsi, Marta Novak, Institute of Behavioural Sciences, Semmelweis University, Budapest, Hungary AIMS: In this study we assess the prevalence of depression in chronic kidney disease (CKD) patients and analyze the association between the presence of significant psychological distress and healthrelated quality of life (QoL). METHODS: Ninety four patients at a single predialysis clinic in Toronto, Ontario received a battery of widely validated tools including the Center for Epidemiologic Studies-Depression (CESD) questionnaire, the Illness Intrusiveness Rating Scale (IIRS) and the Kidney Disease Quality of Life Questionnaire (KDQoL) kidney disease specific QoL questionnaire. Laboratory data was also recorded. RESULTS: Mean age was 70±13 years (28-91 years), 60% male, 46% diabetics. Mean calculated GFR (abbreviated MDRD formula) was 24±8 ml/min. The prevalence of clinically significant psychological distress (cutoff of 16 on the CESD scale) was 46%. The proportion of depressed individuals was similar in males versus females. The mean age and most of the laboratory parameters were similar between depressed versus nondepressed groups. Patients with CESD scores greater than 16 had significantly worse self reported health status, increased Fatigue Severity Scale scores (5,2±1,2 vs 3,2±1,7 for depressed vs nondepressed patients, respectively; p<0,001) and much increased illness intrusiveness (IIRS score 49±19 vs 26±14 for depressed vs nondepressed patients, respectively; p<0,001). In multivariate logistic regression analysis CESD score 16 or greater was independently associated with increased illness intrusiveness and impaired quality of life along multiple domains even after controlling for clinical and socio-demographic variables. CONCLUSIONS: In summary we found a high prevalence of clinically significant psychological distress in CKD patients that was independently associated with increased illness intrusiveness and impaired QoL.

Urology/Kidney Diseases and the Impact on QOL 1170 /Renal Function is Associated with Health Related Quality of Life in Kidney Transplanted Patients Miklos Z. Molnar, Zsofia A. Kovacs, Institute of Behavioural Sciences, Csaba Ambrus, Fresenius MC, Andras Szentkiralyi, Eszter Vamos, Lilla Szeifert, Agnes Koczy, Maria Czira, Anett Lindner, Marta Novak, Istvan Mucsi, Institution of Behavioural Sciences, Semmelweis University, Budapest, Hungary

1150 /Quality of Life Assessment of Genitourinary and Sexual Function After Radiation Therapy for Prostate Cancer Maria Hardy, John Robertson, Radiation Oncology, Mamtha Balasubramaniam, Research Institute, William Beaumont Hospital, Royal Oak, MI

AIMS: Most kidney transplanted (Tx) patients have some degree of renal impairment. It is not known if health related quality of life (HRQOL) is associated with the residual renal function in this population. In this cross-sectional study we asses the association between HRQOL and residual renal function in kidney transplanted patients. METHODS: All patients 18 years or older (n=1067) who were followed at our university outpatient transplant clinic were approached to participate in the study. Baseline demographic characteristics and laboratory results were collected at enrollment. The patients completed a battery of validated, self-administered questionnaires including the Kidney Disease Quality of Life Questionnaire (KDQOL). Estimated glomerular filtration rate (eGFR) was calculated using the abbreviated MDRD formula. Statistical analysis was carried out using the SPSS software. RESULTS: Mean age of patients was 49±12 years, 60% were male, the prevalence of diabetes was 17%. Mean eGFR was 49±12 ml/min. Based on the eGFR, patients were classified into groups corresponding to chronic kidney disease (CKD) stages: group 1: eGFR³60ml/min; group 2: eGFR 30-59 ml/min; group 3: eGFR <30 ml/min. HRQOL of patients

AIMS: To compare whether age and radiation modality affects prostate cancer patients' quality of genitourinary and sexual functions. METHODS: Radiotherapy treatments were external beam (EBRT), High Dose Rate brachytherapy (HDR Alone), Low Dose Rate brachytherapy (Pd103), HDR+EBRT and EBRT+Pre-treatment hormones (PTH). 316 pts were enrolled: 186 EBRT, 61 HDR Alone, 28 Pd103, 23 HDR+EBRT and 18 EBRT+PTH. 93 pts were <65 yrs, 162 were 65-75 yrs and 61 were >75 yrs. The International Prostate Symptom Score and International Index of Erectile Functioning Patient Sexual Function Questionnaire were completed pretreatment, at treatment completion, and 6, 12 and 24 months post treatment. GU questions addressed incomplete emptying, frequency, intermittency, urgency, weak stream, straining, nocturia and GU quality of life (QOLGU). Sexual function included sexual activity/intercourse,

A-59

stimulation and ejaculation. Responses were analyzed for significance and trends in response due to age, treatment modality, their combined effect and time using the Generalized Estimating Equation approach. RESULTS: Treatment modality was significantly associated with incomplete emptying, frequency, intermittency, weak stream and QOLGU (p<0.01 except QOLGU combined with time points which was p=0.02). Urgency, straining, nocturia and sexual function were independent of treatment modality. Age had no association with any GU questions except nocturia (p<0.0001) but was strongly associated with all sexual function questions (p<0.0001). Time points were significantly associated with urinary urgency, straining, weak stream, intermittency, erectile confidence (all p<0.0001) and QOLGU (p=0.02). A significant downward trend was observed for sexual function; the lowest score was at 24 months (p<0.0001). CONCLUSIONS: Pts treated with EBRT+PTH and HDR Alone reported superior GU function to those treated with Pd103 regardless of age or time to measurement. Sexual function was unaffected by treatment modality but was dominated by age and time to measurement, with the worst function at two years.

Patrick Français, Nephrologie, Clinique du Sud, Thiais, France, Emmanuel Ecosse, Biostatistique et Informatique Médicale, hop Cochin, Université de Paris 5, Paris, France, Serge Briançon, Epidemiologie & Evaluation Cliniques, CEC CIE6, CHU de NancyInserm,Université Henri Poincaré Nancy1, Nancy, France AIMS: End-stage renal disease has an important impact on the patients' daily life, which can be measured by quality of life questionnaires. The objective of this work was to adapt the Kidney Disease Quality of Life questionnaire (KDQoL) into French and to determine its basic psychometric properties, i.e. validity and reliability. METHODS: The KDQoL consisted of 8 generic dimensions and 11 specific dimensions. The questionnaire was translated several times independently, and then submitted to a committee of professionals. The study of the measurement properties was carried out near 68 dialysis patients. RESULTS: KDQoL is valid and reproducible, and has properties comparable to the original instrument: missing items proportion of 5.5%, limited floor and ceiling effects (except for 4 dimensions), Cronbach alpha coefficient varying from 0.64 to 0.92 (except for 2 dimensions), test-retest coefficient greater than 0.67 (except for 3 dimensions), and the items of KDQoL were better correlated with their dimension than with other dimensions (except for 2 dimensions). Correlations between the generic and the specific scores showed the absence of redundancies between specific and generic dimensions. CONCLUSIONS: Thus the French version has comparable properties to the original KDQoL. This questionnaire can be used to measure the quality of life of the dialysis patients. It constitutes a good tool in clinical research, allowing international comparisons.

1520 /Quality of Life Following Renal Failure Eugénia J. Anes, Escola Superior de Saúde, Instituto Politécnico de Bragança, Bragança, Portugal, Pedro L. Ferreira, Centro de Estudos e Investigação em Saúde, Universidade de Coimbra, Coimbra, Portugal AIMS: This paper aims to describe and compare the quality of life of individuals suffering from chronic kidney failure, in dialysis, according to the different types of admission. METHODS: A nonexperimental, transversal, descriptive/comparative study was developed using a sample of 231 individuals (76.3% of the population) with chronic kidney failure who were receiving dialysis treatment in the northeast of Portugal. We used an evaluating instrument (KDQOL-SFTM1.3) made up of a generic measurement of health (SF-36) and a specific one used for this clinical condition (ESRD). RESULTS: The majority of the participants studied were male (56.3%), married (68.4%), retired (84.9%), of rural origin (67.4%) and with, at most, a basic education or less (89.1%). The age of the respondents varied between 18 and 88 years old (average = 61.6; median = 65). The time of dialysis of these patients varied between 15 days and 24 years. The type of treatment used more frequently was hemodialysis (94.8%). The majority of the patients had other associated illnesses (56.3%) and complications (91%). The most prevalent associated illness was diabetes (26.4%) and the most prevalent complications were tiredness (69.7%), sleeping sensation in hands and feet (58.9%), cramps (54.5%), muscular pain (52.4%) and itching (51.5%). Those patients who had a programmed admission to dialysis showed a more satisfactory quality of life than those admitted from the emergency unit, with statistically significant differences (p<0,05) in their professional activity, physical performance, emotional performance, social function, and vitality. The results showed the negative impact of some socio-demographic and clinical variables. CONCLUSIONS: In this study we recognized the importance of the evaluation of the health-related quality of life as an indicator of excellence in health care provided for patients with renal failure.

Psychometrics: New Results and Insights 2 1572 /Ferrans & Powers Quality of Life Index: Development and Validation of the Wound Version Beatriz F. Yamada, Nursing, Enfmedic Saúde, São Paulo, São Paulo, Brazil, Vera L. Santos, Medical Surgical Nursing, Nursing College/ University of São Paulo, São Paulo, São Paulo, Brazil AIMS: to develop and validate a wound version of the Ferrans & Powers QOL Index (QLI-WV). METHODS: after accomplishment of all ethical issues, the study was conducted in three poles. The theoretical pole included the selection and analysis of the specific items. The empirical pole was developed through the application of the resulted instrument to test (n=362), test-retest (n=63) and convergent validity (n=179) samples, composed of outpatients from 16 health facilities. The analytical pole included the statistical strategies required to analyze the instrument´s psychometric properties: internal consistency (IC) and stability (ICC); the content validity, through the concordance level among judges; the concurrent validity, through the correlation between the item your satisfaction and the domains and overall QOL mean scores (Pearson or Spearman); the convergent validity, through the correlation between the domains and overall QOL of QLI-WV and WHOQOL-bref scores; and the discriminant validity through the comparison among the domains and overall QOL mean scores with age, wound number and durability, and level of pain (Mann Whitney, Kruskal-Wallis, tStudent and ANOVA tests). The confirmatory factor analysis (CFA) was tested by the correlation between the QLI-WV items and original domains (health/functioning HF; socioeconomic-SE, psychological/spiritual PS and family-Fa) and the adjustment measures model (LISREL). RESULTS: Cronbach´s alpha were 0.90; 0.88; 0.65; 0.81; 0.55 respectively for overall QOL and domains (HF, SE, PS and Fa) for IC; r> or = 0.83 (p<0.000) for stability; r=0.28 to 0.69 for concurrent validity; r=0.17 to 0.60 for convergent validity. Significant correlation between: overall QOL and number of wounds

1605 /Measuring Quality of Life in End-Stage Renal Disease. Transcultural Adaptation and Validation of the Specific Kidney Disease Quality of Life Questionnaire Stephanie Boini, Epidemiologie & Evaluation Cliniques, CEC CIE6, CHU de Nancy-Inserm,Université Henri Poincaré Nancy1, Nancy, France, Alain Leplege, Histoire et Philosophie des Sciences, Université de Paris 7 Denis Diderot, Paris, FRANCE, Carole LoosAyav, Epidemiologie & Evaluation Cliniques, CEC CIE6, CHU de Nancy-Inserm,Université Henri Poincaré Nancy1, Nancy, France,

A-60

(p=0.047); PS and wound durability (p=0.017); HF (p=0.043), SE (p=0.008), PS (p=0.000) and overall QOL (p=0.003) for elderly patients; and SE and worst last week pain (p=0.015) were also obtained. CFA suggested a reasonable fit for four original subscales. CONCLUSIONS: QLI-WV can be considered reliable and valid in wound care.

methods. RESULTS: There was a high correlation between the two ISI items (frequency of leakage and amount) (r = 0.83). Cronbach's alpha for the UDI6 was 0.78. However, item analysis indicated that the UDI6 could be improved by deleting poor items. The prevalence estimates for urinary incontinence derived from UDI6 (47%) were far higher than for the ISI (24%) but dropped to 36% when poor items were removed. There is a prima facie case for the revision of the UDI6. The Cronbach's alpha for the WCGS was 0.57, which was poor. The WCGS has been criticized for including flatus when the current definition of faecal incontinence from the International Continence Society excludes it. Based on the standard WCGS faecal incontinence prevalence was 35% but with flatus excluded this dropped to 8%, which is consistent with other prevalence estimates. Other concerns with the WCGS are noted. The impact of incontinence on quality of life is reported, showing a significant loss in utility value on the AQoL measure by incontinence severity. The findings suggest that severe urinary incontinence has a greater impact on quality of life than does faecal incontinence. CONCLUSIONS: There is a case for the revision of the UDI6 and the WCGS and/or the development of more refined instruments to assess continence status. Although incontinence is common, it has a moderate impact on quality of life. The results of this study can be used as population prevalences and estimates of the impact of the condition on quality of life in continence research.

1574 /Development and Validation of the Bladder Cancer Index (BCI) Rodney L. Dunn, David Wood, Scott Gilbert, James E. Montie, John T. Wei, Urology, University of Michigan, Ann Arbor, MI AIMS: To develop and validate a comprehensive, disease-specific health-related quality of life measure for localized bladder cancer treatment, with urinary, bowel, and sexual domains. METHODS: Survey items were drafted based on literature review and expert endorsement. An initial pilot study among patients with various stages of bladder cancer (N=62) was used to obtain feedback on content. The revised questionnaire was then used in a second pilot study (N=72) to evaluate item performance. In this study, factor analyses, test-retest correlation, item scale correlations, and response distributions were employed for item selection. Formal reliability and validity evaluation was then conducted for the final form of the questionnaire in a third sample of 315 patients receiving 1 of 4 treaments: cystectomy with either neobladder or ileal conduit, or endoscopic resection with or without intravesical therapy. A subset of 50 patients was randomly selected from across the 4 treatments to assess the test-retest reliability. RESULTS: Cronbach's alpha and test-retest coefficients demonstrated strong evidence for reliability (Table). Reliability was maintained within all 4 treatment groups and both genders (Cronbach's alpha >0.79 for all domains). The urinary domains of the BCI had moderate correlations with the total FACTBl domain (0.40-0.52) and lower correlations with the SF-12 domains (0.20-0.30). CONCLUSIONS: The newly validated BCI is a valuable complement to the clinical evaluation of health outcomes for localized bladder cancer patients, suitable for both genders and all commonly performed therapies.

1227 /Validation of a Questionnaire to Measure Perceptions About Medications For Diabetes(PAM-D) Patrick O. Monahan, Kathleen A. Lane, David G. Marrero, Medicine, Indiana University School of Medicine, Indianapolis, IN, Risa P. Hayes, Global Health Outcomes, Eli Lilly and Company, Indianapolis, IN AIMS: To facilitate an understanding of patient medication preferences, a questionnaire was developed to measure patients' Perceptions About Medications for Diabetes (PAM-D). METHODS: Questions (Likert scale) were developed from 18 focus groups consisting of individuals with diabetes. The questionnaire was mailed to 1000 patients with diabetes (primarily low-income). RESULTS: Mailings yielded 267 respondents; age ranged 20-95 (mean = 58); 74% female; 54% African American, 43% Caucasian; 47% < high school diploma; 77% at poverty level; 6% earned < $30,000. Exploratory factor analyses with varimax rotation identified 3 factors from 17 items related to administration: Flexibility of Timing (F1); Convenience (F2); and Hassle (F3); Cronbach's alphas were .88, .86, and .74, respectively. A single 10-item factor was identified as Medication Effectiveness (alpha = .97). A single 6-item factor was identified as Trouble Remembering to Take Medications (alpha = .91). A single 15-item factor was identified as Physical Side Effects (alpha = .88). A single 13-item factor was identified as Emotional Side Effects (alpha = .94). For known-groups construct validity, we examined the relationship between the 7 factors and medication type (insulin alone, oral alone, insulin plus oral). Medication type differed significantly on most factors, and in anticipated directions. Further construct validity was demonstrated in that correlations between factor scores and the SF-36 general health item were all significant (+/- 0.20 to 0.36), and in anticipated directions. Cronbach's alphas for all factors were very similar for African Americans and Caucasians, indicating appropriateness for ethnic subgroups. CONCLUSIONS: The PAM-D is reliable, factorially valid, and consistent with clinical observation. The revised questionnaire can be easily adapted to measure perceptions of medications for other diseases. Further research will examine applications in other disease areas.

Reliability statistics for individual domains of the BCI BCI Domains Urinary Bowel Sexual

TestNumber retest of Items reliability 12 0.92 10 0.87 12 0.92

Cronbach s alpha 0.85 0.82 0.91

1592 /The Psychometric Properties of Measures Assessing Continence Conditions Jan Sansoni, Nicholas Marosszeky, Emily Sansoni, Australian Health Outcomes Collaboration, CHSD, University of Wollongong, Wollongong, NSW, Australia, Graeme Hawthorne, Psychiatry, University of Melbourne, Melbourne, Victoria, Australia AIMS: This paper presents psychometric data from the field testing of measures for continence conditions using an Australian population sample. Testing included: two measures of urinary incontinence, the Incontinence Severity Index (ISI) and the Urogenital Distress Inventory-6 (UDI6); and the Wexner Continence Grading System (WCGS) for faecal incontinence; as well as the Assessment of Quality of Life (AQoL) utility measure. METHODS: Data was collected from the 2004 South Australian Health Omnibus Survey (N = 2924). Analysis was conducted using classical psychometric

A-61

(p=0.03; ES=0.20). On the LVQOL improvement was found at T1 for the multidisciplinary group (p=0.04; ES=0.15). Correlations between person measures and logMAR visual acuities varied between 0.22-0.40 for the VCM1 and 0.35-0.55 for the LVQOL. CONCLUSIONS: Although, sometimes significant improvement in vision related quality of life was shown over time, the effect sizes appeared to be small. We conclude that the sensitivity of both instruments is poor. There was no strong relation between outcome and visual acuity. Whether rehabilitation contributed to vision related quality of life or whether the two vision related quality of life measurement instruments should be used in the future to evaluate rehabilitation outcome remains a matter of further discussion. Improvement of both instruments, but also rehabilitation might be necessary.

1442 /Psychometric Comparison of the Standard EQ-5D to a 5 Level Version in Cancer Patients Maria C. De Leon, A. Simon Pickard, Center for Pharmacoeconomic Research, University of Illinois, Chicago, IL, Thomas Kohlmann, Community Medicine, University of Greifswald, Greifswald, Germany, David Cella, Sarah Rosenbloom, CORE, Evanston Northwestern Healthcare, Northwestern University, Chicago, IL AIMS: There is a perception that the 3 level structure of the standard EQ-5D may hinder limit its discriminative and evaluative ability compared to a version with more levels. The aim of this study is to compare an experimental 5 level US-English version of the EQ-5D to the standard 3 level based on self-reported HRQL assessments by cancer patients. METHODS: Cancer patients participating in a USbased multicenter symptom scale validation study completed the standard EQ-5D and a 5 level prototype in US English. Responses to the 3L and 5L were examined in terms of proportion of logical inconsistencies, convergent validity based on correlations with performance status (i.e. ECOG level). The Shannon Evenness Index (EH') was used to determine informational richness of each version, adjusted by number of levels. RESULTS: From 215 respondents, 39 of 1075 dimensional responses were inconsistent (3.6%) between the 3L and 5L versions, with the highest rate of inconsistency on the usual activities dimension (8.4%). Stronger correlations between performance status and the 5L version were consistently observed across all dimensions compared to the 3L, i.e. rs (5L/3L): MO 0.38/0.37; SC 0.32/0.29;UA 0.39/0.29; PD 0.27/0.20; AD 0.14/0.08; crude summary score 0.41/0.34. Informational richness of the 3L and 5L was similar (EH'5L=0.74; EH'3L=0.72). CONCLUSIONS: Results suggest a 5-level version of the EQ-5D will improve upon some of the psychometric properties of the standard 3 level version. This study broadens the generalizability of initial evidence by Janssen and Bonsel supporting a 5-level version through (i) selfreported HRQL from patients with a wide range of health; (ii) an English language version in a US-based setting.

1459 /Linking Systems: A Valuation Crosswalk Between the Standard EQ-5D and a 5 Level Version Using Item Response Theory A. Simon Pickard, Center for Pharmacoeconomic Research, Pharmacy, University of Illinois, Chicago, IL, Thomas Kohlmann, Community Medicine, University of Greifswald, Greifswald, Germany, M.G. Janssen, Gouke Bonsel, Public Health Epidemiology, Academic Medical Center, Amsterdam, Netherlands, David Cella, Sarah Rosenbloom, Evanston Northwestern Healthcare, CORE, Northwestern University, Chicago, IL AIMS: The objectives are to examine the equivalency of meaning between the standard 3 level (3L) EQ-5D and a 5 level prototype (5L) across dimensions, and to psychometrically derive a system of weights that facilitate the conversion of standard EQ-5D preferencebased algorithms to a 5L system. METHODS: Data from 2 studies that included EQ-5D 3L and 5L versions were analyzed from: (1) a Dutch study of primarily hypothetical health state assessments by 82 respondents; (2) a multicenter study of 215 cancer patients in the USA. Item Response Theory (IRT) models were used to examine the equivalency of the 3L and 5L systems using the ConQuest program. The 3L5L crosswalk was based upon model results for each dimension. Category-specific mean values of latent person parameters for the levels of the 3L and 5L systems were estimated without restrictions imposed on their values, using maximum likelihood estimates. RESULTS: Preliminary results indicate means on the latent continuum pertaining to level 3 in the 5L system and level 2 in the 3L (some problems) were similar for both datasets, suggesting equivalence of these levels. In general, intermediate levels inserted into the 5L were midway between responses on the 3L version across the 5 dimensions. In examining equidistance between levels, ratios of differences between mean values of the person parameters were relatively invariant (close to 1.00). Extremes of the 5L response structure consistently broadened the measurement continuum. 5L utility weights are proposed based upon calibrated distances on the interval level logit scale between 5L response options relative to response options on the 3 level version. CONCLUSIONS: Through IRT models, a crosswalk was proposed that facilitates estimation of a set of societal preference weights for a 5L EQ-5D from any existing 3L algorithm until 5L direct valuation studies are completed. This study illustrates the advantage of combining economic and psychometric approaches to measurement system development.

Item Response Theory 3 1493 /Measuring Vision Related Quality of Life of Dutch Visually Impaired Elderly Ruth M. van Nispen, Maaike Langelaan, Ophthalmology, VU University Medical Center, Amsterdam, The Netherlands, Michiel R. de Boer, Faculty of Earth and Life Sciences, VU University, Amsterdam, The Netherlands, Ger H. van Rens, Ophthalmology, VU University Medical Center, Amsterdam, The Netherlands AIMS: To evaluate the sensitivity of the Vision quality of life Core Measure (VCM1) and the Low Vision Quality Of Life questionnaire (LVQOL) to measure change, using the polytomous rating scale measurement model of Wright and Masters. METHODS: Visually impaired elderly patients (N=296) from ophthalmology departments of four Dutch hospitals entered our study from July 2000 until January 2003. Patients were referred either to monodisciplinary visual rehabilitation (optometrist: 53%) or to a multidisciplinary rehabilitation centre (47%): non-randomised design. After 4 months 83% and after a year 73% still participated in our study. All patients administered the questionnaires by self-report, at baseline, three months (T1) and one year (T2). Mean person measures (Winsteps) and effect sizes (ES) at T1 and T2 were calculated for both instruments, follow up times and rehabilitation groups. Pearson correlations were calculated for person measures and logMAR visual acuities. RESULTS: Paired samples t-tests for the VCM1 showed improvement at T1 (p=0.02; ES=0.12 and p<0.01; ES=0.24) for the monodisciplinary and the multidisciplinary groups resp. At T2 only the multidisciplinary group showed improvement on the VCM1

1813 /Development of a Computer Adaptive Version of the Interpersonal Competence Questionnaire (ICQ-CAT) Otto B. Walter, Heinz Holling, Uwe Kanning, Psychology, University of Munster, Munster, Germany AIMS: There has been little research on how social skills determine the social dimension of quality life. To keep item burden placed on

A-62

respondents low we developed a computer adaptive version of the Interpersonal Competence Questionnaire (ICQ). METHODS: This adaptive version was based on a representative sample (N=1934) of respondents who answered 30 items of the German translation of the ICQ. A random half (N=967) of the sample was used to evaluate test dimensionality, to calibrate the items, and to capture the relation between person parameters and total raw scores in a regression model. The other random half of the sample was employed to assess the deviation between scores predicted by this regression model and real raw scores. RESULTS: After confirmatory factor analysis 28 items remained in the pool and were calibrated using a two-parameter IRT model for polytomous items (Generalized Partial Credit Model). Correlation between total raw scores and estimated scores using all 28 items and expected a posteriori estimation was very high (R2=.98). In the simulation of a computer adaptive test algorithm, 17.1±3.4 items (mean±SD) sufficed to estimate the construct with high precision (SE<.32) in the second half of the sample. The regression model allowed for a very accurate prediction of total raw scores from estimated IRT scores (deviation±SD: -0.38±3.3). CONCLUSIONS: These results indicate that the computer adaptive version of this ICQ version can measure the underlying construct precisely and efficiently and appears to be a valuable tool for the investigation of the relation between social skills and the social dimension of quality of life.

1348 /Real World Applications of Dynamic Health Assessment: Lessons from the Development and Clinical Field Testing of Five New Prototype Tools Diane M. Turner-Bowker, Anastasia E. Raczek, Jakob B. Bjorner, Renee N. Saris-Baglama, Michael DeRosa, Milena Anatchkova, Janine Becker, John E. Ware, Jr., Research & Development, QualityMetric Incorporated, Lincoln, Rhode Island AIMS: PRO measures can be improved using item response theory (IRT), and competing objectives of more practical and more precise tools can be achieved over a wide range of severity levels using computerized adaptive testing (CAT). We share lessons learned from the development and clinical field testing of five new prototype CATs. METHODS: Five Small Business Innovation Research (SBIR) studies explored the feasibility of IRT and CAT in clinical practice, for pediatric health (N=80), chronic kidney disease (N=55), diabetes (N=200), headache (N=100), and asthma (N=114). CAT generic and disease-specific impact item banks were programmed with screening and user evaluation scales for Internet or tablet PC administration, generating IRT-based scores and patient/provider feedback reports. CATs were piloted in clinics with trained staff to evaluate feasibility; obtain preliminary estimates of item usage, response burden, range, and score accuracy compared with fulllength versions; patient/provider acceptance; and usefulness of reports. RESULTS: CATs were successfully integrated into the daily clinical routine; CAT scores using an average of 5 items were highly comparable (r=0.93 to 0.96) to full length surveys through the scale range; significant reductions in response burden were achieved while maintaining desired precision levels; CATs were as efficient as full length surveys in discriminating known groups; feedback reports were useful for patients/providers; and most patients found the CAT length appropriate, easy to complete, and relevant. Potential challenges include population-specific equipment adaptation needs, computer experience, and literacy level. CONCLUSIONS: CAT tools were well accepted by patients, providing high measurement precision with low response burden. IRT/CAT offer promising measurement improvements that may facilitate outcomes monitoring and care management. As with any survey, interface and cultural concerns must be considered. Forthcoming results from seven more SBIR CAT studies may address these.

1695 /Calibration of Item Banks for Use In Improving Estimates of Eight SF-36 Health Constructs Jakob B. Bjorner, Mark Kosinski, Xiaowu Sun, John E. Ware Jr., QualityMetric Inc., Lincoln, RI AIMS: Short-form questionnaires for patient reported outcomes have proven useful in outcomes research, but more precise measures are needed for optimal study power and for realiable assessment of individuals. We aimed to develop item banks for improved assessment of the eight SF-36 health constructs. METHODS: Items from 52 questionnaires hypothesized to measure the eight SF-36 generic health domains were administered to large general population samples using the Internet. Tests of unidimensionality, multitrait and confirmatory factor analysis, and item response theory methods were used to identify homogeneous item banks and to calibrate them for use in computerized dynamic survey administrations. Item banks were evaluated for differential item functioning with regards to age, gender, marital status, race, education, and disease group. All banks were normed to achieve a mean of 50 and a standard deviation of 10 in the US general population. RESULTS: Out of 354 items evaluated for inclusion in the item banks, 255 items fulfilled the measurement requirements. Final item bank size varied from 12 items (General Health Perceptions) to 61 items (Mental Health). Most excluded items showed departure from the unidimensionality assumption or conceptual overlap with other items in the bank. For items fulfilling these measurement requirements, only a few cases of IRT misfit or DIF were found. The final item banks substantially increased measurement precision and reduced floor and ceiling effects compared to the original SF-36 scales. Computerized adaptive administration allowed these gains to be achieved without increasing response burden. Procedures were developed to ensure backwards comparability with the SF-36 scales. CONCLUSIONS: Substantial measurement improvements can be achieved by building large item banks for patient reported outcomes. Such item banks enable computerized adaptive assessment and selection of improved fixed short forms while maintaining backwards comparability with existing short-form measures.

Cancer Studies and Approaches in QOL 3 1752 /Tu-Be or Not Tu-Be? The QOL-EF Tool for Measuring the Impact of Enteral Feeding on QOL Jolie Ringash, Radiation Oncology, The Princess Margaret Hospital/University of Toronto, Toronto, ON, Canada, Betty Lemon, Clinical Nutrition, Gina Lockwood, Biostatistics, The Princess Margaret Hospital, Toronto, ON, Canada, Andrea Bezjak, John Waldron, Radiation Oncology, The Princess Margaret Hospital/University of Toronto, Toronto, ON, Canada AIMS: To assess the reliability, validity and acceptability of a new therapy-specific QOL instrument for head and neck cancer (HNC) patients receiving enteral feeding. METHODS: At our centre, HNC patients receiving accelerated fractionation radiotherapy or concurrent chemoradiotherapy undergo prophylactic enteral feeding tube placement. Through focus groups, patient pilot and item reduction studies, we have developed a 20-item (score range: 0-80) instrument to measure the impact of enteral feeding on QOL called the QOL-EF. We accrued 72 patients with locoregional disease on treatment or < 3 months post-treatment to complete the FACT-H&N, UWQOL, QOL-EF and a preference questionnaire. A subgroup of 31 who reported no change in well-being repeated the QOL-EF 1 week later. RESULTS: Mean patient age was 56; 78% were male, 67% had

A-63

pharyngeal primaries and 88% had stage IV disease. Mean BMI (95% CI) was 23 (21-24) and time from treatment start was 2.2 (1.72.8) months. Mean performance status (PS) and QOL scores were: KPS 73 (71-75), ECOG 1.4 (1.2-1.5), FACT-H&N 84.2 (80-88), UWQOL 52.7 (50-56), QOL-EF 48.9 (46-51). Pearson correlation for test-retest reliability was 0.83 (0.66-0.91) and Crohnbach's alpha for internal consistency was 0.73. We confirmed a hypothesized moderate correlation between QOL-EF with FACT-H&N (0.61) and UW-QOL (0.47). QOL-EF score also correlated significantly with ECOG PS and weight loss, but not with chemotherapy (yes = 33, no = 36), or time since tube or treatment start. Significant item-specific correlations (> 0.45) were observed for the following: QOL-EF "eat by mouth" and UWQOL "swallow"; QOL-EF "vomiting" and UWQOL "feel ill"; QOL-EF "dependant on others" and FACT-H&N "losing hope". The questionnaire which "best reflects my QOL" was: FACT-H&N 42%, QOL-EF 34%, UWQOL 24%. CONCLUSIONS: The QOL-EF has shown acceptable reliability and validity for discriminative use. Its content dominates the overall QOL experience of 1/3 of participants. This instrument may have future applicability to enteral feeding in other diseases (eg. stroke).

Central Hospital, Helsinki, Finland, Pirjo Räsänen, Helsinki and Uusimaa Hospital Group and, Finnish Office for Health Technology Assessment, Helsinki, Finland, Risto P. Roine, Group Administration, Helsinki and Uusimaa Hospital Group, Helsinki, Finland, Tiina Jahkola, Plastic Surgery, Helsinki University Central Hospital, Helsinki, Finland AIMS: The objective of this study was to describe the quality of life (QoL) of patients treated for melanoma. METHODS: Retrospective QoL (EORTC QLQ C-30) data were obtained from 1019 patients treated for melanoma between 1988 and 2004 at the Department of Plastic Surgery, Helsinki University Central Hospital. The copliance (percentage recieved of expected forms) of the study was 72%. Functional and symptom QoL scores were compared with published age- and gender-adjusted reference data from the general population. Results were compared and interpreted in a descriptive way in line with the findings that differences of more than 10 points are clinically meaningful. T-tests were used to examine differences between agegroups. RESULTS: The mean age of the patients was 61 years (range 18-94). Compared with the general population, melanoma patients QoL differed in some respects. In the age group 40-49 years; women with melanoma reported higher scores for emotional functioning and less pain than the reference population. Within the melanoma group more severe limitations were observed for the physical, cognitive and role functioning scales and global QoL with increasing age. The oldest respondents (+80 years) had a greater degree of impaired QoL than the other age groups, with two exceptions i.e emotional and social functioning in which there was no deterioration. The older cancer patients scored also significantly higher for the symptom subscales of fatigue, pain, appetite loss, constipation, diarrhea, and financial difficulties. CONCLUSIONS: Results indicate that long-term survivorship following diagnosis and treatment of melanoma does not have a negative impact on long term quality of life. More severe limitations of QoL follows with increasing age, apart from emotional and social functioning.

1655 /Fatigue Severity in Different Risk Groups of NonHodgkin's Lymphoma (NHL) Patients Stratified by International Prognostic Index (IPI) Andrei A. Novik, Hematology Department, National Pirogov Medical Surgical Center, Moscow, Russia, Tatyana I. Ionova, Svetlana A. Kalyadina, Anton V. Kishtovich, Quality of Life Department, National Cancer Research and Treatment Center, St. Petersburg, Russia AIMS: Fatigue is the most distressing symptom associated with cancer. Specific aims of this study were to describe fatigue prevalence in the new NHL patients and to identify its severity level in the risk groups stratified by IPI. METHODS: 119 patients with newly diagnosed NHL - 57 aggressive (stage I-IV, mean age 60.0 SD=17.2, males/females 30/27) and 62 indolent (mean age 59.8 SD=16.0, males/females 35/27) completed Brief Fatigue Inventory before treatment. To find out the relationship between fatigue and IPI we examined fatigue severity and fatigue interference in patients' groups stratified by IPI. Differences between groups was tested by Mann-Whitney U-test. RESULTS: Fatigue was reported by 77.3% NHL patients with its prevalence in aggressive lymphoma (93%). Almost two thirds (60.5%) of patients experienced fatigue at the moderate-to-severe level. Aggressive NHL patients were significantly more interfered with fatigue than indolent ones: mean BFI interference score - 5.98 (SD=2.50) vs 2.16 (SD=2.55) (p<0,05). Fatigue severity differed significantly in the IPI groups (p=0.001). Patients of low risk according to the IPI both in aggressive and indolent NHL groups had no fatigue. Patients of intermediate risk experienced mild fatigue (mean 3.2, SD=2.3 for aggressive NHL; mean 2.1, SD=1.8 for indolent NHL). Patients in high-intermediate group were characterized by moderate fatigue (mean 5.7, SD=1.5 for aggressive NHL; mean 4.1, SD=2.4 for indolent NHL). Patients of high risk had severe fatigue (mean 7.5, SD=1.4 for aggressive NHL; mean 7.1, SD=0.9 for indolent NHL). Significant differences in fatigue interference with patients' daily life were found across IPI groups (p=0.001). CONCLUSIONS: Certain IPI group is strongly distinguished by fatigue severity and its impact on QoL. The findings support the suggestion that fatigue severity level might be a right factor to be included in new prognostic models for NHL and in doing so to select the most appropriate treatment in individual patients.

1682 /Combined Scaling of the Responses of Patients with Pancreatic Disease to the EORTC QLQ C30 and QLQ PAN26 Colin D. Johnson, Surgery, Pantalia Roussou, Scott Harris, Ruth M. Pickering, Health Care Research Unit, University of Southampton, Southampton, UK AIMS: The EORTC QLQ C30 is used with site specific modules (SSM). Some QL domains are covered by items in both the C30 and the SSM. This may lead to duplication. Scales derived from the combined C30 +SSM would maximise the information included in scales, and reduce their number. We have developed scales for evaluation of quality of life (QL) in patients with pancreatic cancer (PC) or chronic pancreatitis (CP), from combined responses to the EORTC QLQ C30 and the QLQ PAN26 METHODS: One complete record per patient was randomly selected from 6 PC trials (n=372), and 3 CP datasets (n=125). Data were analysed using SAS. After exclusion of 6 items (sexual function 2, support, information, and global scale 2), responses to 50 items were analysed by factor analysis using principal axis factoring with direct oblimin rotation and ´=0. Scree plots and derived models were inspected to determine the optimal number of factors RESULTS: For PC, 37% of variation was explained by a single factor model (eigenvalue 17), but this was not useful for separating QL domains. Up to 9 factors had an eigenvalue >1. An 8 factor model (activities and weakness, pain, body image, bowel function, food and appetite, discomfort, and emotional and social function) incorporated 41 items and explained 56% of the total variation. Cronbach s ± for these scales ranged from 0.72-0.95. Between scales, correlation was highest (0.48) between pain and emotional function. Similar scales arose for CP: a 7 factor model included 46 items and explained 61% of variation;

1335 /Quality of Life in Patients with Melanoma Compared with the General Population Minna-Liisa Luoma, Psychology, University of Helsinki, Helsinki, Finland, Minna Ristiniemi, Plastic Surgery, Helsinki University

A-64

Cronbach s ± ranged from 0.79 to 0.95, with moderate correlation between pain and activities (0.49) and emotional function (0.44) CONCLUSIONS: The combination of items from both the QLQ C30 and PAN26 has produced putative scales covering important aspects of QL, reducing overlap of scales compared with separate analysis of the two components. The scales have good face validity and few items are not included in the scale. QL assessment in patients with pancreatic disease will be simplified by combined scaling in the EORTC system

at two and four months using descriptive analysis. RESULTS: To date, 19 of the 24 (79%) patients that have been recruited and allocated randomly to the intervention group have completed the study. Diary completion rates were high with 74% of patients completing 85-100% of the diary each week over the four month period. Missing data was generally entire pages due to error or entire weeks due to deteriorating health. Only two patients (8%) showed consistently poor compliance (approximately 30% completion rate). There were no individual questions consistently omitted. Patients reported that the diary posed appropriate questions (90%), that the time taken to complete the diary was acceptable (95%) and that it was appropriate to include the diary in routine care (80%). Although 56% of patients reported that the diary could be useful in consultations with members of the health care team, only 8% stated that they used the diary in this way. CONCLUSIONS: These early interim results show high completion rates of a QOL diary in patients during routine care but few patients actively share the QOL information with health professionals. Further recruitment and follow-up of patients will be used to test the extent to which the completion of a QoL diary by lung cancer patients benefits their quality of care.

1496 /Symptoms Related to Endocrine Therapy in Women with Breast Cancer and Their Impact on Quality of Life Karin Ribi, Jürg Bernhard, Quality of Life Office, Internat. Breast Cancer Study Group Coord. Center, Bern, Switzerland, Kaspar Rufibach, Statistics, Swiss Group for Clinical Cancer Research, Bern, Switzerland, Beat Thürlimann, Roger von Moos, St.Gallen, Switzerland, Thomas Ruhstaller, St. Gallen, Switzerland, Agnes Glaus, Centre for Tumour Prevention, Detection and Treatmen, St.Gallen, Switzerland, Christel Böhme, St.Gallen, Switzerland AIMS: To investigate the perception of symptoms related to endocrine therapy in women with early or advanced breast cancer receiving endocrine treatment, and to examine the impact of these symptoms on quality of life (QL)indicators. METHODS: Perception of symptoms was assessed by the Checklist for Patients on Endocrine Therapy (C-PET, yes-no answer format) and by linear analogue selfassessment (LASA) indicators in a consecutive sample of 373 patients with early or advanced breast cancer receiving endocrine therapy. In addition, patients responded to global LASA indicators for physical well-being, mood, coping effort and treatment burden. The association between endocrine symptoms and these indicators was analysed by linear regression models. RESULTS: The comparison of women indicating a symptom as present versus absent yielded large effect sizes on the symptom LASA indicators (range of Cohens's d=-1.48 to -2.06). Six of 13 symptoms, hot flushes/ sweats, nausea, tiredness, irritability, vaginal dryness, breathlessness, and patients' age were significantly associated with the global indicators: Overall, younger women and those experiencing tiredness and irritability reported worst scores. CONCLUSIONS: The occurrence of symptoms related to endocrine therapy is a considerable burden to breast cancer patients. Less frequently reported symptoms may also have a substantial impact on QL. Patients' perception of symptoms needs to be considered both, in patient care and research when interpreting the association between symptoms and quality of life.

1175 /Serum Calcium as an Independent Predicitor of QOL in Multiple Myeloma Finn Wisloff, Ann Kristin Kvam, Hematology, Ulleval University Hospital, University of Oslo, Oslo, Norway AIMS: Skeletal disease with development of osteolytic lesions is an important feature of multiple myeloma. At diagnosis, about 30% of the patients have an increased serum calcium level. We wanted to examine the impact of serum calcium on the QOL scores of multiple myeloma patients at the time of diagnosis. METHODS: Data from two Nordic studies with 686 patients who completed the EORTC QLQ-C30 questionnaire were analysed by means of linear regression analysis and a curve fitting program. Independent (predictor) variables in addition to serum calcium were age, gender, haemoglobin, serum creatinine, serum albumin, serum beta-2 microglobulin, disease stage and extent of skeletal disease. RESULTS: Serum calcium was independently related to appetite loss, nausea/vomiting and physical functioning (P < 0.001) and to cognitive functioning (P = 0.001), i.e. scores reflecting symptoms that are well known in non-malignant hypercalcaemia (primary hyperparathyroidism). Also, we found a highly significant independent relationship between serum calcium and the scores for fatigue and pain (P < 0.001). Serum calcium appeared to be about as strong a predictor for fatigue as haemoglobin. A cubic model (y = a + bx3) fitted the data slightly better than the simple linear model (y = a + bx) and suggested worsening QOL scores at levels of serum calcium above 2.5 to 3.0 mmol/l. The R-squares of the models were low, ranging from 0.04 to 0.19, indicating that the ability of the models to predict the QOL scores of a patient from the values of the objective disease variables was low. A large proportion of the variability of QOL scores is likely to be due to individual psychological factors. CONCLUSIONS: Hypercalcaemia in patients with multiple myeloma seems to be associated with the same symptoms as in non-malignant hypercalcaemia. In addition, an increased level of serum calcium may aggravate the pain and fatigue caused by the multiple myeloma itself.

Cancer Studies and Approaches in QOL 4 1415 /Preliminary Evaluation of a Patient-Held Quality of Life Diary in Patients with Inoperable Lung Cancer Moyra E. Mills, Epidemiology & Public Health, Queens University Belfast, Belfast, County Antrim, Northern Ireland, Michael Donnelly, Epidemiology & Public Health, Queens University Belfast, Belfast, Co. Antrim, N. Ireland AIMS: To develop, implement and evaluate the use of a patient-held quality of life (QoL) diary in routine care. METHODS: Patients with recently diagnosed inoperable lung cancer, attending the oncology clinic for the first time, with a WHO performance status 0-2 and no cognitive impairment were recruited throughout Northern Ireland and randomly assigned to receive a patient held QoL diary or standard care. The QoL diary consists of a core questionnaire (EORTC C-30) and a lung cancer module (LC-13), in the form of a small A5 booklet, which patients complete weekly. Patients are encouraged to share their diary with health professionals involved in their care. Diary completion rates and patient utilisation and satisfaction were assessed

1584 /Relationship Between Quality of Life, Socio-Demographic Characteristics and Breast Cancer Surgery Cristina Pires, Medical Oncology, Hospital São Sebastião, Santa Maria da Feira, Portugal, Rute F. Meneses, FCHS, Universidade Fernando Pessoa, Porto, Portugal, Francisco L. Pimentel, Secção Autónoma Ciências da Saúde, Universidade de Aveiro, Aveiro, Portugal

A-65

AIMS: Breast cancer, a deadly and stigmatising disease, is the most common form of cancer in women, affecting physical, psychological and social domains. Both breast cancer and its treatment can influence the patient s Quality of Life (QoL). For optimal care, it is of the utmost importance to know which patients are at risk of having lower QoL and which treatment options are associated to better QoL. Consequently, the aims of this sudy are: a) to clarify the relationship between QoL and socio-demographic variables (civil status, age, and professional status); and b) to compare the QoL in women with cancer who have undergone mastectomy vs. conservative surgery, at least a year before assessment. METHODS: One hundred consecutive outpatient women, with an average age of 55.80 years (SD=13.03; 31-85), were evaluated. Of these, 73 were married, 52 were professionally inactive, 68 had undergone chemotherapy, 63 radiotherapy, 42 mastectomy and 58 conservative surgery. The instruments used, in the context of an individual interview, were a socio-demographic questionnaire designed for the present study and the Functional Assessment of Cancer Therapy - Breast cancer (FACT-B). Pearson correlations and t Student tests were used. RESULTS: Married participants reported statistically higher scores on the Social/Family Well-Being (p=.007), Additional Concerns (p=.02), and General QoL (p=.02); older subjects reported statistically higher scores on Physical Well-Being (p=.01), and lower scores on Additional Concerns (p=.001); and professionally inactive participants reported statistically lower scores on Additional Concerns (p=.001). No statistically significant differences were found between women submitted to mastectomy vs. conservative surgery. CONCLUSIONS: The results of the present study point: a) to the importance of socio-demographic variables in terms of QoL; and b) to a less negative effect of mastectomy than is currently believed. These results, if replicated, can guide intervention.

CONCLUSIONS: Roughly one half of the cancer patients studied indicated that they preferred to have a collaborative relationship with physicians. The US cohort seemed to want to be more assertive than their Canadian counterparts and women ended up playing a more passive role than they preferred. These gender and cross-county differences highlight the need for individualized patient communication styles to be incorporated into treatment plans 1178 /Differential Experience of Breast Cancer Treatment Related Symptoms: A Patient Cluster Approach Clement K. Gwede, Interdisciplinary Oncology, Moffitt Cancer Center and Univ. of South Florida, Tampa, FL, Brent J. Small, Interdisciplinary Oncology, Moffitt Cancer Center and Univ. of South Florida, Tampa, Florida, Michael A. Andrykowski, Behavioral Science, University of Kentucky College of Medicine, Lexington, Kentucky, Paul B. Jacobsen, Interdisciplinary Oncology, Moffitt Cancer Center and Univ. of South Florida, Tampa, Florida AIMS: Cancer patients experience multiple concurrent symptoms. This exploratory analysis assessed patterns of symptom burden among patients undergoing chemotherapy for breast cancer (BrCa) to identify clusters of patients who report similar or distinctly differing symptom burden, and assessed whether patient clusters were associated with deleterious quality of life (QOL) outcomes. METHODS: One hundred thirty three women with stage I or II BrCa undergoing adjuvant chemotherapy after primary surgery were evaluated at baseline and after initial chemotherapy for symptoms using the Memorial Symptom Assessment Scale (MSAS) and for QOL using the SF-36. Median patient age was 51.9 years; 90% received lumpectomy alone prior to chemotherapy. Post-treatment MSAS symptoms (presence or absence) were included in hierarchical cluster analysis. RESULTS: Two patient clusters were identified that corresponded to a high symptom prevalence group and a low symptom prevalence group. No statistically significant differences were found between groups for demographics, comorbidity, chemotherapy regimen, or body mass index. Patients in the high symptom cluster were more likely to have stage I disease (p<0.05). At baseline, no significant differences were present between groups for prevalence, severity and distress of symptoms or QOL. However, the two groups of patients showed significant differences in posttreatment SF-36 scores (p<0.05). In addition, post-treatment MSAS items were associated (p<0.05) with group membership, suggesting that the two groups of patients reported different prevalence and nature of symptoms. The high symptom burden group was more likely to report poorer QOL, and greater symptom prevalence, severity, and distress. CONCLUSIONS: Our preliminary findings suggest that a subgroup of patients experience a high symptom burden during chemotherapy for BrCa that has deleterious effects on QOL. Future research needs to examine why these differences occur despite similarities in treatment and how symptom burden can be reduced for the high symptom prevalence group.

1801 /What Role do Cancer Patients Want to Play in Treatment Decision Making: A Pooled-Analysis Jeff A. Sloan, Mashele Huschka, Pamela Atherton, Health Sciences Research, Mayo Clinic, Rochester, MN, Lesley F. Degner, Nursing, Tom Hack, Psychology, University of Manitoba, Winnipeg, MB, Canada, Tenbroeck Smith, American Cancer Society, Rochester, MN, Teresa Rummans, Psychiatry, Mayo Clinic, Rochester, MB, Matthew Clark, Psychiatry, Mayo Clinic, Rochester, MN AIMS: Patient involvement in the decision making process for cancer treatment can impact satisfaction with care. A pooled-analysis of clinical studies from the US and Canada incorporating the Control Preferences Scale (CPS) was conducted to produce normalized data regarding patient preferences and examine differences in role preference related to country, tumor type, gender and other demographics. METHODS: Patient data culled from six trials indicated the treatment decision making role preferred and the role actually experienced clinically. Fisher's Exact Tests were performed to compare role distribution concordance and association with clinical and demographic variables. RESULTS: Data available for 3491 patients indicated that 25% preferred an active role, 46% a collaborative role, and 29% a passive role in their medical treatment decision making. In terms of actual experience, 30% of patients reported taking on an active role, 34% collaborative, and 36% passive. Overall, 61% of patients reported playing the role they prefer. Differences between genders in the preferred role were slight, but males achieved their preferred role more often than females (66% vs. 60%, p=0.011). More women actually took a passive role than men (40% vs. 24%, p<0.0001) as did more patients in the US than Canada (84% vs. 54%, p<0.001). Canadian patients preferred more passive than active roles (33% vs 22.4%) and US patients preferred more active to passive roles (31.9% vs 14.2%) (p<0.001). Older patients preferred a more passive role and took on that role. Differences in role preference across tumor types were negligible.

Methodology Developments in QOL 3 1709 /The Composite Nature of Health-Related Quality of Life Measurement Benoit Arnould, Nadia Cadi-Soussi, Patient Scales for Clinical Practice, MAPI Values, Lyon, Rhone, France, Catherine Acquadro, MRT, Mapi Research Trust, Lyon, Rhone, France AIMS: The EMEA released a specific reflection paper about their expectations regarding the use of Health-Related Quality of Life (HRQoL) in clinical research. This guideline is applicable since the beginning of 2006. The FDA also released a draft guidance for industry with a broader scope, including HRQoL. Complexity is

A-66

acknowledged as a key characteristic of HRQoL. Our objective is to discuss how far HRQoL measures match the characteristics of composite endpoints and could be analysed as such, in order to address the regulators' expectations. METHODS: A systematic analysis of 32 EMEA Notes for Guidance (NfG) published before June 2005 and including explicit or implicit references to HRQoL and other PROs was conducted, focusing on the issues of complexity and multiplicity. The EMEA HRQoL reflection paper was then analysed and compared to Moye's definition of composite endpoint and their characteristics. A comparison was then performed with the FDA draft guidance. RESULTS: The analysis of EMEA NfGs and the FDA draft Guidance showed that expectations closely matched the 5 principles listed by Moye for composite endpoints: 1) Clinical Relevance, 2) Prospective Deployment, 3) Principle of Coherence, 4) Therapy homogeneity and 5) Full disclosure. CONCLUSIONS: Considering HRQoL as a composite endpoint seems to be a fair solution to address regulators' requests regarding the management of complexity in HRQoL measurement.

Family & Preventive Medicine, UCSD, La Jolla, CA, David Feeny, Health Utilities Incorporated, Dundas, Ontario, Canada, Paul Kind, Centre for Health Economics, University of York, York, UK AIMS: To introduce the National Health Measurement Study (NHMS) which establishes simultaneous norms for six different generic health-related quality of life (HRQOL) indices in noninstitutionalized older US adults. METHODS: Random digit dialed telephone interviews were used to obtain a population-based sample of 1262 non-institutionalized US adults aged 35-89 between June 2005 and Feb 2006; we over-sampled African Americans and persons aged 65-89. The SF-36v2, EQ-5D, QWB-SA, and HUI2/3 were administered in random order followed by the HALex, BRFSS QOL questions, 2 Psychology of Well-Being scales, 13 health conditions, smoking status, height, weight, 10 perceived discrimination items, socioeconomic, and demographic items. Individuals’data are linked to contextual US census data. We compute mean scores by age and sex, and also compute scores for body mass index (BMI) categories adjusted for age and sex. Regression analyses use post-stratified weights with SAS 8.02 . SF-36 responses were converted to SF-6D. RESULTS: The 6 indices show overall decreasing HRQOL means with approximately parallel slopes with increasing age (10y categories centered at 40, 50, 60, 70, 80, and sex-adjusted). QWB-SA (mean scores .67, .68, .64, .61, .59, respectively) had lowest means in each age category; EQ-5D (US weights; means .89, .88, .84, .84, .80) had highest means, except at ages 45-54 where EQ-5D, HUI2/3, and SF-6D give nearly identical means. While overall downward trend with age is significant (p<.05), the s.e. of each individually estimated mean is about 0.02 with the current sample (HUI3 had less precision). All but SF-6D show a slight inverted U-shaped relation with BMI, with lower values at BMI<20 and BMI>35. CONCLUSIONS: Widely used generic HRQOL indices show similar patterns when administered simultaneously and used to compute agespecific population means, although the means are offset approximately by a constant value when comparing one index to another. NHMS expects another 2000 interviews by Sept 2006, and data to be public in 2008, providing a rich resource for research.

1750 /Are SMPH Models Applicable for Clinical Studies of QOL? Michael T. Molla, Ageing and Chronic Diseases Study, National Center for Health Statistics, Hyattsville, Maryland, USA AIMS: According to the World Health Organization (WHO), for the older population, ‘independence’ is assumed to be ‘the ability to perform functions related to daily living’ or the capacity of living independently in the community with no and/or little help from others’. In population health studies, summary measures of population health (SMPH) is used to summarize the health of the population taking into account both mortality and morbidity. One such measure is the remaining years of independent life . For a cohort of a given age, the remaining years of independent life is defined as the part of remaining life to be spent with independence assuming that current mortality and morbidity conditions will continue into the future unchanged. In health studies, remaining years of independent life is estimated using data from cross-sectional or longitudinal health surveys and the life table technique. The model used to make the estimates depends on the source of the morbidity data. While models that are based on data from cross-sectional surveys use prevalence rates of dependence, models based on data from longitudinal surveys use incidence of dependence. METHODS: In this study we will introduce a method of estimating remaining years of independent life based on models that use incidence of dependence. We will use data from the Second Longitudinal Study of Aging (LSOA-II) and IMaCh to illustrate the method. IMaCh is a newly developed PC based computer program used for estimating remaining years of life in any health state. RESULTS: We will also discuss how this model in combination with this newly developed computer program can be used for estimating QOL of patients in clinical studies. CONCLUSIONS: In health studies, measuring the remaining years of independent life serves as an invaluable component of the health of older persons into the future, and provides the foundation for the efficient allocation of health care resources. Using the same model and software to estimate QOL of patients in clinical studies, will also serves the same purpose while at the same time saves a lot of resources.

1737 /To Establish a Need-Based Resources Allocation Formula According to The Quality-Of-Life and Mortality Health Outcomes: The Case of Taiwan Yu-Hung Chang, School of Health and Related Research, University of Sheffield, Sheffield, UK AIMS: Most resource allocation formulae adopt SMR as a health outcome to evaluate relative needs of health care across regions. QoL measures could be valuable in resource allocation since health care not only benefits in length of life but also quality of life. The study is to establish a need-based resource allocation formula, which integrates the QoL index into the length-of-life index, in order to ration health care resources in Taiwan. METHODS: The database of the 2001 Taiwan National Health Interview Survey (NHIS) and the 2001 vital statistics were employed to construct the formula. 1,860 age-sex classified subgroups of 128 townships in Taiwan are the allocating targets on which the formula is to distribute health care resources according to relative health gains attributed to health care utilisation. The health gains were measured by a combination index incorporating the average QoL index of age-sex groups, derived from the SF-36 outcomes in 2001 NHIS, with the expected-one-yearlength-of-life, derived from age-sex specific morality rates in vital statistics of the current year. The utilisation rates, from 2001 NHIS, were measured as the frequencies of visits or admissions of inpatient services, emergency and outpatients services. The 2SLS method was used to model the effects of the utilisation rates on the combination health index. Other factors such as morbidities, indices of socialeconomic status and ethnical composition were also adopted. With

1759 /The National Health Measurement Study: Us Data for Six Generic HRQOL Indices Dennis G. Fryback, Population Health Sciences, University of Wisconsin-Madison, Madison, WI, Mari Palta, Nancy C. Dunham, Janel Hanmer, Shani Herrington, Dasha Cherepanov, Jennifer Buechner, Population Health Sciences, University of WisconsinMadison, Madison, Wisconsin, Ron D. Hays, Medicine, Robert M. Kaplan, Health Services, UCLA, Los Angeles, CA, Theodore Ganiats,

A-67

applying the coefficients of the model, we calculated the relative need index for each age-sex group. RESULTS: The aggregate need index for each region could be obtained by population weighting of its age-sex composition.The final outcomes of the region-level need index will be presented in the study with detail. CONCLUSIONS: The QoL measures, integrated with mortality, can be used in resource allocation formula.

RESULTS: MD, a chronic, largely untreatable eye condition, is the most common cause of blindness in the Western world. Shock of diagnosis, coupled with lack of information and support are a common experience. Incidence of depression is twice that found in the community-dwelling elderly. Some people feel suicidal. MD threatens independence, especially when comorbidity exacerbates functional decline. Rehabilitation, including low vision aid (LVA) provision and training, peer support and education, can improve functional and psychological outcomes but many people do not receive services likely to benefit them. Medical treatments, available for only some types of MD, can improve function but most limit progress of MD, at least for a time, rather than cure. The White Paper considers difficulties associated with inappropriate use of health status measures and misinterpretation of utility values as QoL measures: evidence suggests they have poor validity in MD. CONCLUSIONS: There is considerable evidence for the major damage done to QoL by MD which is underestimated by health status and utility measures. Medical treatments are limited to a small proportion of people. However, much can be done to improve QoL by early diagnosis of MD with good communication of prognosis and continuing support. Support could include provision of LVAs, peer support including MD Society groups, education and effective help in adjusting to MD.The value of such provision is unlikely to be demonstrated by health status or utility measures. It is vital that appropriate measures of visual function and QoL be used in building a sound evidence base for the effectiveness of rehabilitation and treatment.

1774 /Quality of Life Before Intensive Care Admission is a Strong Predictor of Survival Jose G. Hofhuis, Dept of Intensive Care, Gelre Hospital, Apeldoorn, Netherlands, Henk F. van Stel, Julius Center for Health Sciences, UMC Utrecht, Utrecht, Netherlands, Jan Bakker, Dept of Intensive Care, Erasmus Medical Center, Rotterdam, Netherlands, Guus Schrijvers, Julius Center for Health Sciences, UMC Utrecht, Utrecht, Netherlands, Peter E. Spronk, Dept of Intensive Care, Gelre Hospital, Apeldoorn, Netherlands AIMS: Health related quality of life (HRQL) before admission to intensive care (IC) may have a relation to subsequent mortality. We assessed whether HRQL before IC can be used as a predictor of mortality, in addition to or as a substitute for other predictors. METHODS: Patients admitted to IC for >48 hours. Close relatives completed the SF-36 in the first 48 hours of admission to assess HRQL of the 4 weeks before admission to IC. Logistic regression was used to assess the influence of PCS, MCS and the general health item at mortality, corrected for age, gender and APACHE-II score. This is a severity of illness index that is a well-known predictor of mortality and widely used in IC. We report the percentage correct predictions of survival/death from classification tables and areaunder-the-curve (AUC) from ROC-curves. RESULTS: 451 patients were included at admission to the IC, with informed consent from the proxies and later the patients themselves. 170 patients got sepsis. At 6 months follow-up,159 patients had died, and 40 patients were lost to follow-up. Mean age was 68.6 (sd 12.5) years. AUC for models with HRQL-only was the same as APACHE-only in all patients (0.74), and better in sepsis patients (0.8 vs. 0.76). AUC and prediction of survival and death improved when HRQL was added to APACHE. Wrong predictions decreased: survival 37% to 34% (all), 35 to 27% (sepsis); death 30% to 25% (all), 27% to 22% (sepsis). Similar results were found when using only the general health item or PCS/MCS from SF-12. CONCLUSIONS: This study shows that pre-admission HRQL in IC patients, especially in sepsis patients, is at least as good in predicting survival/mortality as the APACHE-II score and improves prediction when combined. This is clinically relevant as the 1-item general health question is easily and quickly obtained, and the number of wrong predictions to relatives about survival/mortality decreases, especially in sepsis patients. The implication of this study is that IC clinicians and nurses should obtain HRQL information from close relatives as soon as a patient is admitted.

1072 /Ageing and Health-Related Issues Within the Iranian Families - A Sociological Study of Quality of Life Mohammad Taghi Sheykhi, Department of Social Science, Al-Zahra University, Tehran, Tehran, Iran AIMS: The paper explores the problem of social support to the senior citizens as a multiple issue in Iran. Problems of ageing people are usually considered as a consequence of gradual, but rapid urbanization and industrialization which need due attention and solutions. Similarly, change in social functions of families and generations, and the increasing need to provide long-term care for the aged is currently becoming problematic in Iran so far as the elderly’s quality of life is concerned. However, delving about the health status, and elderly’s satisfaction with life are among the aims of the research in Iran with special reference to Tehran. Children’s financial, emotional and physical supports towards their ageing parents are among the aspects investigated. METHODS: Methodology used in the survey based on 500 questionnaires, is a combined form of empirical-theoretical method of research. RESULTS: Finding show 30% of respondents had lost their partners, 56% not receiving any financial supports from their children, and 44% receiving emotional supports from them. Similarly, 45% party expressed their life satisfaction, and 40% were found healthy. CONCLUSIONS: Shortage of care and attention by the family, society, and lack of social security system are found as the main issues within the ageing people in Iran.

Quality of Life in the Elderly 2 1728 /Quality of Life in Macular Degeneration: Age-Related Macular Degeneration Alliance International (AMDAI) White Paper Jan Mitchell, Clare Bradley, Psychology, Royal Holloway, University of London, Egham, Surrey, United Kingdom

1822 /The Influence of Filial Norms and Intergenerational Exchange on Life Satisfaction of Elderly Parents in a Compartive Cross-Cultural Context Nurit Gur-Yaish, Ariela Lowenstein, Masters in Gerontology, Ruth Katz, Human Services, Haifa University, Haifa, Israel

AIMS: The AMDAI sponsored us to review the literature and prepare a White Paper on QoL in macular degeneration (MD) with a view to increasing awareness of MD, reducing its impact and improving services for people with MD worldwide. METHODS: A systematic review was conducted using electronic databases, conference proceedings and key journal hand search checks.

AIMS: The paper addresses the influence of filial norms and intergenerational exchange on life satisfaction of elderly parents. METHODS: These questions were explored under different family cultures and welfare state regimes with a sample of 1703 elder parents (75+) from the OASIS five-country study (Norway, England,

A-68

Germany, Spain and Israel). RESULTS: Descriptive analyses indicated differences in intergenerational exchange patterns between the five countries. At the same time, affectional and consensual solidarity with children significantly predicted life satisfaction even when country, personal and health variables were adjusted. Filial norms significantly predict satisfaction from life by itself, However, filial norms did not significantly predict life satisfaction when countries, personal and health variables were included in the model. CONCLUSIONS: Taking together, the data suggest that even though there are cultural differences in family relationships, some aspects of intergenerational solidarity effect satisfaction from life beyond these differences.

to complete an MVQOLI on admission to hospice. Failure to complete an MVQOLI was significantly associated with admission site (P<0.001), gender (P=0.002), marital status (P=0.037), 5 terminal diseases (P<0.05), 2 co-morbidities (P=0.013), KPS mean score (P<0.001), MADLS mean score (P<0.001), DSS mean score (P<0.001), and survival time (P<0.001). Failure to complete an MVQOLI was not associated with age, 6 terminal disease diagnoses, or 7 co-morbidity categories. CONCLUSIONS: Failure to complete MVQOLI on admission to hospice was mainly due to patient-related variables. Many patients were very ill, near the end-of-life, or experiencing significant dyscognition. However, hospice staff and system failures may have contributed to this omission. Non-response bias remains a major problem and longitudinal studies will be difficult to conduct in this population. New instruments and strategies are needed to overcome these barriers.

1233 /Assessment of the Psychometric Performance of the WHOQOL-BREF Instrument in a Sample of Brazilian Older Adults Eduardo Chachamovich, Dept of Psychiatry and Legal Medicine, Clarissa Trentini, Marcelo Fleck, Dept. of Psychiatry and Legal Medicine, Federal University of Rio Grande do Sul and HCPA, Porto Alegre, RS, Brazil AIMS: The population’s aging trend has been observed all over the world. Measurements of quality of life have been raising growing interest in cross-section investigations and the assessment of intervention outcomes in older adults. Several instruments used in the measurement of quality of life in older adults were not appropriately tested in this population. This study aims at describing the psychometrical properties of WHOQOL-BREF in a sample of Brazilian older adults. METHODS: 424 older adults were selected through convenience sampling and filled out the instruments WHOQOL-BREF, BDI, BHS and socio-demographic data form. Discriminant validity, concurrent validity, criterion validity and internal consistency were analyzed. RESULTS: The sample was predominantly comprised of women (64.2%), community-dwelling subjects (84.4%) and subjects who consider themselves healthy (67.5%). All domains in the instrument showed the ability to discriminate levels of depressive symptoms and hopelessness, as well as different perceptions of their health status. The correlation coefficients among the domains and the BDI and BHS scores were statistically significant. The reliability coefficients present scores ranging from 0.614 to 0.925. CONCLUSIONS: The WHOQOLBREF instrument shows suitable psychometrical performance in a sample of Brazilian older adults, becoming a useful alternative in the measurement of quality of life in this population. 1207 /Barriers to Measuring Quality of Life at the End of Life Thomas E. Elliott, Colleen M. Renier, Jeanette A. Palcher, Education and Research, Duluth Clinic, Duluth, Minnesota AIMS: Our study examined the barriers to obtaining health-related quality of life (HRQOL) data in patients admitted to hospice, using the Missoula-Vitas Quality of Life Index (MVQOLI). Assessing HRQOL in patients admitted to hospice may provide critical data about the patient's health status and needs, help establish treatment goals, and measure outcomes. METHODS: We included all patients admitted to a hospice located in the North-central U.S. June 1, 2002 December 31, 2003. A retrospective, cross-sectional cohort design was used that included a heterogeneous convenience sample of patients with any terminal disease. Independent variables included admission site, gender, age, marital status, terminal disease diagnosis(11 categories), co-morbidities (9 categories), Karnofsky performance status (KPS), modified activity of daily life scale (MADLS), descriptive symptom scale (DSS), and survival time. The dependent variable was completion of MVQOLI. Chi-square and ttest analyses were performed with 2-tailed tests of significance. RESULTS: The study included 1,047 subjects with 743 (71%) unable

A-69

Anxiety/ Depression and Mobility (34% each), Usual activities (23%) and Self care (8%). Concordance statistics showed excellent agreement between child and proxy ratings on the Mobility, Selfcare, Usual activities, Anxiety/ Depression and total scores (kappa = 0.642, 0.845, 0.574, 0.530, 0.539 respectively, p= 0.001) and very good agreement on the Pain/ Discomfort scores (kappa= 0.394, p= 0.001). On applying quartiles, a composite score of 1 emerged in the > 75th percentile (n = 12, 26%)i.e. 26% of the children had a good HRQOL despite illness. Thalassemic facies and the side effects of chelation accounted for 43.9% of variability in the child-rated HRQOL (R=0.343). The frequency of blood transfusions and sideeffects of chelation accounted for 41.3% of variability in Mobility scores (R= 0.308). The side effects of chelation and the duration of diagnosed illness accounted for 43.4% of variability in Anxiety/ Depression scores (R= -6.67). CONCLUSIONS: (1) The Clinical variables had a significant impact on HRQOL, suggesting that optimal medical management might improve HRQOL by reducing the complications of illness and its treatment.(2) The EQ5D merits further study as a tool for measuring HRQOL in Indian children.

POSTERS

1/Abstract 1443 Health Related Quality of Life In Children Following Stem Cell Transplantation Catharina M. Löf, Pediatrics CLINTEC, Karolinska Institute, Karolinska University Hospital, Stockholm, Sweden AIMS: The aim of this study was to explore aspects of health realted quality of life (HrQoL) in children and adolescents following allogeneic stem cell transplantation (SCT), both according to diagnosis and the use of a related or unrelated donor. METHODS: A total of 52 children, (age 9-22, m=15) and at least 3 years (median=8) beyond SCT for leukemia (n=31) or nonmalignant diseases, participated in a single center study of health related quality of life (HrQoL) at Karolinska University Hospital, Huddinge. 42 parents also participated. Descriptive statistics, Students T-test and standard multiple regression analayses were used to asses the effect of diagnosis, donor choice and subjective and objective health on HrQoL domains in "Child Health Questionnaire" (SCHQCF87/SCHQ-PF50). RESULTS: As a group the children had good self-rated HrQoL in comparison to a normgroup and other chronically ill children. Most children were also subjectively and objectively in good health. Analysis according to diagnosis, malignant or nonmalignant, showed that children with leukemia rated a lower psychosocial HrQoL and had a higher degree of late effects. Parents rated their children´s psychosocial and physical HrQoL lower than the normative group of parents, and lower if the child was diagnosed with leukemia. The severity of the child´s physician-rated late effects or of self-assessed subjective symptoms was associated with a lower parental rating of the child´s HrQoL in the psychosocial area, and of their own emotional situation. The psychosocial HrQoL area was most affected in recipients of an unrelated donor graft, according to child and parent. CONCLUSIONS: Children reason that they have a good HrQoL following SCT, but their parents do not agree with this view. Clarity with regard to who is responsible for assessing the child´s HrQoL is crucial when interpreting paediatric HrQoL studies. One important clinical conclusion from the study might be to focus health care resources on psychosocial support to both child and parent, especially on the group with more late-effectrelated problems.

3/Abstract 1802 Making Pediatric Health Assessments More Useful: A Focus Group Study of Parents and Clinicals Anastasia E. Raczek, QualityMetric Incorporated, Health Assessment Lab, Waltham, MA, Renee N. Saris-Baglama, Michael A. DeRosa, QualityMetric Incorporated, Lincoln, RI, Helene M. Dumas, Maria Fragala-Pinkham, Franciscan Hospital for Children, Brighton, MA, John E. Ware, Jr., QualityMetric Incorporated, Health Assessment Lab, Waltham, MA AIMS: To understand how parents and clinicians interpret and use information from pediatric health-related quality of life (HRQOL) and disability surveys and to identify perceived gaps in measurement and reporting. METHODS: Clinicians (Physical, Occupational, Speech Therapists, n=5) and parents (n=6) of children with a disability or chronic medical condition (Cerebral Palsy, Failure to Thrive/Hypotonia, Fetal Alcohol Syndrome/ADHD, Marfan Syndrome, Asthma/Allergy) participated in semi-structured focus groups. Participants described experiences with health surveys in clinical, rehabilitation, and school settings; defined key domains that tools should assess; and expressed preferences for reporting results. RESULTS: Clinicians reported familiarity with and use of several pediatric health tools (e.g., PEDI, Wee-FIM, Peabody). Although their children were frequently evaluated, most parents could not identify names or purposes of surveys and reported rarely receiving results; they said that the results they do receive are too technical to understand. Domains that clinicians want tools to assess included: mobility, personal care, emotional and social functioning, and sensory integration. Perceived gaps included: socialization, selfesteem, sensory integration, and pre-vocational skills. Parents focused on functional domains such as personal care, and identified social support and family dynamics as central. Clinicians and parents would like to see survey results compared with age- and conditionbased norms. Both groups emphasized a need for more interpretive forms of output than simple mean scores. In general, parents feel overwhelmed by information they receive about their children's diagnoses, and want surveys to provide information that helps them understand how their children are doing and can improve. CONCLUSIONS: Focus groups provide an opportunity for understanding clinician and parent perspectives when developing and improving pediatric assessments and reports. Improved reporting systems may improve communication with families of children with disabilities.

2/Abstract 1011 Thalassemia in India - QoL Issues Deepika Shaligram, Satish C. Girimaji, Santosh K. Chaturvedi, Psychiatry, National Institute of Mental Health & Neurosciences, Bangalore, Karnataka, India AIMS: (1)To assess the Health Related Quality Of Life (HRQOL) in children with thalassemia.(2)To compare child-rated HRQOL with caregiver (proxy)rated HRQOL. METHODS: Thirty nine children (8-16 years) with thalassemia major who attended the day care services at 2 general hospitals in India with their caregivers were recruited for the study after excluding those with other chronic illnesses and mental retardation. The EQ-5D rated by the child and the caregiver (proxy)was used to assess the HRQOL. The Chi square test and Multiple linear regression was used to characterize the relationship between demographic/ clinical details with the HRQOL. Concordance statistics was employed to assess the degree of concordance between the child and proxy rated HRQOL. RESULTS: The HRQOL was impaired in 29(74%) on child-rating and 32 (82%) on proxy rating. The dimensions affected - Pain/ Discomfort (64%),

A-70

and had a gap <=3 treatment days before entering the extension. RESULTS: 85 patients aged >=65 years who received DAR 7.5 or 15 mg od during the feeder studies with a gap <=3 treatment days entered the extension (65-89y; 84.7% female) and 52 (61.2%) completed. DAR 7.5/15 mg was associated with significant improvements in 7 of 9 KHQ domains (Incontinence Iimpact, Role Limitations, Physical Limitations, Social Limitations, Emotions, Sleep/Energy and Severity Measures) from baseline to the end of the 12-week, feeder studies (all p<0.01 vs feeder-study baseline). These benefits in HRQoL were sustained in the geriatric population during the open-label extension, in the same 7 KHQ domains (all p<0.001 vs feeder-study baseline) after a further 24 months of DAR 7.5/15 mg treatment. Furthermore a significant improvement was recorded in the General Health Perception domain (p<0.01 vs feeder-study baseline) at 24 months. DAR was well tolerated in the geriatric population and the long-term safety profile was consistent with that observed in the 12-week studies. CONCLUSIONS: HRQoL significantly improved with DAR treatment in geriatric patients with OAB during the 12-week feeder studies. These benefits were maintained throughout the 2-year extension study.

4/Abstract 1024 Support for Single Elderly People's Better QoL Living in Mountainous Areas in Japan Tada Toshiko, Tetsuya Tanioka, Ruriko Yamashita, Yasuko Matsushita, Chiemi Kawanishi, Shu-ichi Ueno, Nursing, University of Tokushima, Tokusima, Tokushima, Japan AIMS: We plan to identify the factors associated with the quality-of-life (QOL) for residents who need low nursing care and live alone in mountainous regions in Japan. METHODS: Of the 76 elderly subjecs, 58 were female and 8 male ,(mean age ± S.D.: 79.9 ± 5.0) with low nursing care needs (level 1 or 2 among 6 graded by the Japan elderly care insurance system, JECIS) living by themselves included in this study. The survey procedure was explained and informed consent was obtained from all the subjects. The health conditions were evaluated by the assessment in the JECIS during the previous 6-month period. A selfadministered questionnaire was performed to evaluate daily living including the attributes of the subjects, disorders, and the degree of satisfaction. The QOL was evaluated as the satisfaction rating using a scoring system from 0 to 10 (10 is the highest satisfaction level) by subjects themselves. Subjects were divided into three groups as improved, unchanged or aggravated, by how necessity for support states changed in six months. Analysis of variance and Pearson correlation coefficient were used for statistical method. RESULTS: Of the 76 elderly subjects, 14(18.4%)were improved,50(65.8%) unchanged or 12 (15.8%) aggravated. The percentage of males was higher in the aggravated group than that in any other groups. There were no significant differences in self- reported QOL scores among three groups. In all groups, the gonalgia was the most frequently observed, and next was vertebral disorders such as spondylosis deformans. In the improved group, the data showed a significant correlation between the age and the QOL (r=0.472, p<0.05). CONCLUSIONS: In order to improve elderly people's QOL, it was suggested that the influences of joint limitations on daily life can be reduced by care such as rehabilitation for range of activity and pain control for motor dysfunction.

6/Abstract 1322 Does New Comprehensive Balneotherapy Program Improve the Quality of Life in Elderly Persons Michiko Kobayashi, Yuukoh Agishi, Takashi Mandai, Japanese Society of Quality of Life Research, Kobe, Hyogo, Japan AIMS: The purpose of this study was to investigate the efficacy of new comprehensive balneotherapy program for elderly persons from the viewpoint of Quality of Life(QOL). METHODS: Eighteen elderly females participated in this 6 days and 5 nights balneotherapy program. This program consisted of balneotherapy, exercise therapy, dietary therapy, and cultural interchange. Our self-administered questionnaire including 40 questions divided into 15 categories with the Life Satisfaction Index(LSI), Profile of Mood States(POMS), the Revised UCLA Loneliness Scale were used. Physical examinations including blood pressure, pulse rate, body temperature, and body weight were also performed with above-mentioned questionnaire examinations threetimes(ie, at beginning, one and 6 months later). RESULTS: Pearson's correlation coefficients between our questionnaire and LSI was r=0.66(P<0.01). Cronbach's alpha coefficients of our questionnaire were high enough to accept for clinical use : 0.82 in dietary problems, 0.82 in economic problems, and 0.80 in medical service etc. respectively. Compared with before program, there were improvements in average QOL levels especially in the categories of happiness and psychological problems after program. In social participation, significant improvements of QOL(P<0.05) were demonstrated 6 months after especially in over 80 years old females. CONCLUSIONS: These findings indicate that this residential balneotherapy program may have the possibility to improve the QOLs in elderly persons. Comprehensive balneotherapy may play an important role to establish the ideal community for the elderly persons by the effects of stress reduction, activation for mental function, and mental stabilization in future.

5/Abstract 1265 Long-Term Darifenacin Treatment for OAB in the Geriatric Population: QoL Outcomes from a 2-Year, Open-label Extension Study Simon Hill, Obstetrics and Gynaecology, Queens Park Hospital, Blackburn, Lancashire, UK, Richard Millard, Urology, The Prince of Wales Hospital, Sydney, NSW, Australia, Jacques Corcos, Urology, Jewish General Hospital, McGill University, Montreal, Quebec, Canada, Claudio Teloken, Urology, Santa Casa Hospital & FFFCMPA Federal Medical School, Porto Alegre, RS, Brazil, Karine Lheritier, Biostatistics, Fernando Kawakami, Clinical Development & Medical Affairs, Novartis Pharma AG, Basel, Switzerland, Mark Slack, Urogynaecology, Addenbrooke's Hospital, Cambridge, Cambridgeshire, UK AIMS: To evaluate the effects of long-term treatment with oral darifenacin (DAR), a muscarinic M3 selective receptor antagonist, on health-related quality of life (HRQoL) in geriatric patients with overactive bladder (OAB). METHODS: Patients with OAB for >=6 months who participated in two 12-week, placebo-controlled, doubleblind feeder studies of DAR 3.75, 7.5 or 15 mg once daily (od) were enrolled into this 2-year, multicentre, non-comparative, open-label extension study. During the first 2 weeks of the extension all patients received DAR 7.5 mg od, after which titrations between 7.5 and 15 mg od were permitted. HRQoL was assessed using the King's Health Questionnaire (KHQ). This analysis involves geriatric patients (>=65y) who received DAR 7.5 or 15 mg od during feeder studies

7/Abstract 1334 Subjective Health Status Uging Visual Analogue Scale Among Elderly Japanese Tomoaki Kimura, Yoshimitsu Araki, Public Health & Health Policy, Hiroshima University, Hiroshima, Japan, Kenshi Hayashida, Healthcare Economics and Quality Management, Kyoto University,

A-71

significance adopted for the statistical tests was 5%. RESULTS: The mean age of the participants was 75.1 years and the age related macular degeneration was the most prevalent ocular disease. In this study occurred modifications from the original English NEI VFQ-25. The Cronbach_s a was 0.86 and 0.91 (without Optional Appendix and with Optional Appendix, respectively). Intraclass correlation coefficient for test-retest reliability was r= 0.98 in both cases. CONCLUSIONS: In conclusion, this study reveals the Portuguese version NEI VFQ-25 is reliable and suggests to be continued the analyzes the psychometric properties to determine the validity in this population.

Kyoto, Japan, Akira Eboshida, Public Health & Health Policy, Hiroshima University, Hiroshima, Japan AIMS: As the population of Japan has been rapidly aging and single elderly people or families of only old couple are increasing, the maintenance of their quality of life is an essential issue. For assessment of health-related quality of life, not only questionnaires but also visual analogue scales (VAS) are frequently used. However, the elderly are so unfamiliar with VAS that its validity is not clear. The objective of this study was to clarify the validity of VAS among elderly people. METHODS: We reevaluated results of a health survey for an aging community conducted from July to November 2003 at two rural villages in Japan. A total of 2328 subjects from 60 to 75 years of age consented to participate in the survey. A subset of the participants (n=1595, 68.5%) whose VAS scores of health status were effective was analyzed. We examined the associations between VAS scores and responses to each question related to health status or quality of life, and verified the construct validity of the VAS. RESULTS: The subjects included 706 men and 889 women, and their mean age was 67.5. The number of single elderly was 139 (8.7%), and that of people living only with their spouse was 785 (49.2%). Of the subjects, 72.5% earned an annual income below 3 million yen (about $25,000). The mean VAS score was 6.40 (range: 0-10), and negatively correlated with age, but no significant difference was found between genders and between the two villages. The VAS score was correlated with general health status (r=0.53), physical strength (r=0.51), satisfaction in daily life (r=0.30), pain and discomfort (r=-0.47), and anxiety and depression (r=-0.40). The outpatients' VAS scores were lower than that of people who did not visit a clinic or hospital. However, it was not associated with the number of family members, annual income, and lifestyle items such as diet and smoking. CONCLUSIONS: The VAS score was moderately correlated with the equivalent question about health status. It was considered relevant that data on health status are collected through VAS even among elderly people.

9/Abstract 1668 Linguistic Validation of the PAC-SYM and PAC-QOL for Use in International Studies Laurence Chevalle, Mapi Research Institute, Lyon, France, Dominique Dubois, Johnson & Johnson Pharmaceutical Services, Beerse, Belgiumthy Bell, Research Triangle Park AIMS: Prior to use in an international study involving individuals suffering from constipation, the Patient Assessment of Constipation Symptoms questionnaire (PAC-SYM, 12 items) and the Patient Assessment of Constipation QoL questionnaire (PAC-QoL, 28 items) underwent linguistic validation in respectively 13 and 12 languages. The original scales were developed in US- (PAC-SYM) and UKEnglish (PAC-QoL). A rigorous methodology was required to ensure conceptual equivalence and cultural relevance across different languages. METHODS: The translation process was conducted by a specialist in each target country using the following standardized methodology: (1) two forward translations by professional translators who were native speakers of the target language and fluent in English;(2) comparison and reconciliation of the translations by the specialist in the target country and the translators;(3) backward translation by a native English speaker;(4) comparison of source and backward version;(5) review by a clinician and (6) comprehension test on 5 individuals suffering from constipation. RESULTS: The translation process revealed challenges such as: a) the translation of medical terms which had no direct equivalent in certain languages and b) the translation of idiomatic expressions. Medical terms such as "abdomen" required the addition of an explanation to describe the anatomic area more clearly. The translations of idiomatic expressions often needed paraphrasing to convey the intended meaning. CONCLUSIONS: The language versions of the PAC-SYM and PAC-QoL were established according to a rigorous translation methodology. The process aims to ensure conceptual equivalence across different language versions to facilitate international comparison and pooling of data. The linguistic validation process as a whole supports the advantage of integrating international feedback on concepts and wording before a questionnaire is finalized.

8/Abstract 1435 Cross-Cultural Adaptation in the National Eye Institute Visual Functioning Questionnaire NEI VFQ-25 in Brazilian Low Vision Elderly Fernanda A. Cintra, Isabel Cristina M. Fonseca, Neusa M. C. Alexandre, Nursing, State University of Campinas, Campinas, São Paulo, Brazil AIMS: The purpose of this study was to translate and adapt a version of the _National Eye Institute Visual Function Questionnaire _ 25 items (NEI VFQ-25) into Brazilian Portuguese and evaluate its reliability (with and without Appendix Optional) in low vision elderly. METHODS: The cross-cultural adaptation was developed according to internationally recommended methodology, using the following guidelines: translation, back-translation, revision by an expert committee, test of the pre-final version and reliability. First, the questionnaire was independently translated into Portuguese by two bilingual translators. Second, two other professional translators whose mother tongue was English performed a back-translation independently from one another. A committee consisting of six specialists developed a final version. This version was pretested on 30 elderly people with low vision. The psychometric properties of the translated version were evaluated by administering the questionnaire to 65 elderly people with low vision in the philanthropic institution in Campinas, Sao Paulo, Brazil. Descriptive statistics were used to describe the sample.The internal consistency of the instrument was examined with Cronbach_s alpha. Test-retest reliability was tested by using intraclass correlation coefficient (ICC). The level of

10/Abstract 1148 Preferences for Endoscopic Surveillance: A Discrete Choice Experiment Marie-Louise Essink-Bot, Michelle E. Kruijshaar, Public Health, Erasmus MC / University Medical Center Rotterdam, Rotterdam, The Netherlands, Bas Donkers, Business Economics, Erasmus University Rotterdam, Rotterdam, The Netherlands, Ewout W. Steyerberg, Public Health, Peter D. Siersema, Gastroenterology and Hepatology, Erasmus MC / University Medical Center Rotterdam, Rotterdam, The Netherlands AIMS: Individuals with Barrett esophagus (BE) are recommended to

A-72

patients, which might help to attenuate the impact of preoperative psychosocial distress on postoperative QoL-scores. This program is part of the overall support program for surgically treated cancer patients, but focuses more on cognitive instead of emotional components and includes the relatives of male patients much more than in female patients. Preliminary QoL-results are encouraging, but data collection for this subsequent study is still on the way.

undergo regular endoscopic surveillance for early detection of esophageal cancer (EC). However, upper gastrointestinal (GI) endoscopy is burdensome, the absolute 10-year risk of developing EC is estimated as low as 4%, and the survival benefit of surveillance is as yet uncertain. We investigated patients_ preferences for endoscopic surveillance in a discrete choice experiment (DCE). METHODS: We constructed choice scenarios based on 3 attributes: type of test (endoscopy, videocapsule (less burdensome), saliva test (no burden)); frequency of surveillance (5 levels ranging from 4 times a year to once every 5 years); risk of dying from EC (specified decreases in 10-year risk). Fifteen choice scenarios, including 2 options for surveillance and an opt out were selected for the questionnaire. A total of 297 patients with endoscopy experience (155 BE patients and 142 patients with non-specific GI symptoms) agreed to participate in a telephone survey. We included items to help subjects understand the task and items to check if they did. RESULTS: In total, 247 patients completed the questionnaire. The standard scheme for endoscopic surveillance for BE (once every 1 or 2 years, resulting in an estimated decrease of the risk to die from EC from 4% to 1.0% or 1.2%, respectively) was preferred over most other scenarios. Preferences were dominated by the decrease in risk of dying from EC. Preferences shifted to other tests only when endoscopy frequency increased to more than once a year. Patients stated as main reasons for these preferences that they were used to follow their doctor_s advice and not used to make such deliberative choices in matters of health. CONCLUSIONS: Given realistic estimates of expected health gains, endoscopic surveillance once every one or two years is acceptable to patients. DCE is a feasible and promising technique if the task is associated with deliberative choices that subjects make in real life.

12/Abstract 1762 Design Considerations for a Computer Adaptaive Test (CAT) Based on the International Classification of Functioning Disability and Health (ICF) Craig A. Velozo, Reh Outcomes Research Center, Occupational Therapy, North Florida/South Georga VA, University of Florida, Gainesville, Florida, Bradford L. Barrett, No Department, No Institution, Newberry, Florida, Jia-Hwa Wang, Leigh Lehman, YingChih Wang, Rehabilitation Sciences, University of Florida, Gainesville, Florida AIMS: The purpose of this paper is to recommend specific CAT features based on our experiences in developing the ICF Activity Measure. METHODS: The measure was designed to be an efficient and precise self-report of physical function based on contemporary disability theory and classification. Initial design considerations included operation via standard browsers to allow for both web-based and computer-based administration. A separate administration page was designed for selection of: item response theory (IRT) estimation method, initial theta, and randomness in initial theta. Exit criteria included: standard error, maximum/minimum number of questions, maximum number of unexpected responses, and threshold for unexpected responses. Five movement-related constructs were identified within the ICF, each requiring a separate CAT: positioning/transfers, upper extremity function, walking/climbing, wheelchair/scooters, and self care. To selectively administer specific items to a respondent, branching questions included: gender, walk/wheelchair, and wheelchair type (manual/electric). Indication of functional limitation allowed for the administration of deficit-specific constructs. While the above features were pre-planned, field and alpha testing of 480 individuals receiving interventions for back pain, lower extremity (LE) impairment, upper extremity (UE) impairment or spinal cord injury (SCI) led to further design features. RESULTS: Multidimensionality of the UE function construct required designing separate CATs for fine and gross motor function. Differential item functioning (DIF) revealed that four constructs were disease specific, thus requiring separate CATs across several diagnostic groups. Finally, alpha testing revealed that many respondents inappropriately received incontinence items, therefore requiring branching of these items only to individuals with SCI. CONCLUSIONS: Planning of CAT features is necessary but not sufficient. Measurement properties and respondent input can lead to important CAT design features.

11/Abstract 1504 Psychosocial Distress During Gastrointestinal-Cancer Treatment Affects QoL Especially in Men Maria Berend, Regine Sagermann, Malchow Björn, Kirsten Teren, Thomas Küchler, Reference Center Quality of Life in Oncology, University Medical Center Schleswig-Holstein, Kiel, Kiel, SchleswigHolstein, Germany AIMS: Many studies have shown that diagnosis and treatment of GIcancer is associated with a substantial decrease in QoL. Here we looked at the specific impact of additional psychosocial distress at the first stage of treatment, i. e. at baseline on the postoperative course of QoL. METHODS: A total of N = 356 patients with GI-cancer (esophageal, gastric, pancreas and colorectal cancer) were included. These patients have been surgically treated between 4/1998 and 12/2001 at the Dep. of General and Thoracic Surgery at the University Hospital of Kiel (UK S-H). They all completed the EORTC QLQ C 30 (pre- and postoperatively, 3, 6, 12, 24 month postop.) and a validated (Berend, 2005) 9-item questionnaire assessing psychosocial distress. This questionnaire (Hornheider Fragebogen) was originally developed by Strittmatter 1997 and has been modified (and psychometrically tested) for the use with surgical cancer patients. The global statistical significance level was set to a=0.05 and corrected for multiple tests. The design of the study is that of a prospective cohort-study. RESULTS: Psychosocial distress at baseline does not only influence preoperative QoL (EF, GH), but has an (negative) impact on all QoL-assessments up to two years postop. This impact is highest in male patients, who show in some scales more than 20 QoL-points difference to those males who are not distressed at baseline. CONCLUSIONS: These clinically significant differences (Osoba et al., 1998) between psychosocially distressed vs. not distressed male patients with GI-cancer led us to develop a psychosocial intervention program especially for male

13/Abstract 1738 Examination About Scoring System of Japanese Version of HIT-6 Takafumi Wakita, Outcomes Research, iHope International, Kyoto, Japan, Yoshimi Suzukamo, Physical Medicine and Rehabilitation, Tohoku University, Sendai, Japan, Hiroyuki Noguchi, Department of Psychology and Human Development Scien, Nagoya University, Nagoya, Japan, Shunichi Fukuhara, Epidemiology and HealthCare Research, Kyoto University, Kyoto, Japan AIMS: Headache Impact Test(HIT-6) measuring the hindrance degree that what kind of influence the overall headache gives

A-73

everyday life has scoring system to assign 6, 8, 10, 11, 13 points for five rating alternatives in each question. It is necessary to examine whether the scoring system is valid in Japanese version the same as the original version. The purpose of this study is considering about the scoring system on HIT-6 of Japanese version. METHODS: 252 persons that have headache were made applicable to investigation. 1) Unidimensionality of HIT-6 was examined. 2) Each parameter that represents the item characteristics and the degree of each patient_s headache impact on the latent trait scale were estimated with Generalized Partial Credit Model. 3) New scoring system (NS) different from original version was proposed. 4) Three correlation coefficients were calculated to show association among latent trait of HIT-6(LT), score with HIT-6 original scoring system (OS) and score with new scoring system. RESULTS: 1) The result of scree test show that HIT-6 has unidimensionality. The estimates of item category parameters of item 2 to 5 were 1.675, 1.199, -0.918, -1.957. 2) When the evaluate methods by Wakita (2004) was applied for this result, the values to represent each category was -2.089, -1.410, 0.095, 1.317, 2.335. 3) These values were converted to integer scores (6, 7, 9, 11, 13 points) that had the same score range to HIT-6 original version. 4) Correlation coefficients between LT and OS, LT and NS, and OS and NS are .952, .981, and .979, respectively. CONCLUSIONS: It was shown that the NS scores are closer to the LT ones than the OS scores are. Both _never and _rarely were relatively cover narrow range and at near position, and _sometimes covers relatively wide range on the latent trait continuum. These results might be influenced by the characteristic of data in this study that were provided by patient with headache. So, it will be thought that further investigations are necessary in consideration of appropriate standard of the outside criterion to perform the effective measurement.

them higher than 0.4. Both methods showed homogeneous DIF for Australia, Germany, France, Great Britain, Switzerland, and Canada versions. The regression model detected DIF for at least 3 items of Hungary, Slovakia, Spain and Yugoslavia versions. IRT models_ results showed DIF for some more countries CONCLUSIONS: The pattern of DIF showed by most country versions, suggest an appropriate level of equivalence across them. This result is specially relevant for international projects assessing HRQL in patients with Heart Failure. (Funding:EC ALRI-CT-2000-00551; SAF2002-11009E; DURSI-GENCAT 2005-SGR-00491) 15/Abstract 1613 Constructing Shorter Versions of the WHOQOL-BREF for Six Disease Groups Ping-Chi Lee, Grace Yao, Chia-Huei Wu, Psychology, National Taiwan University, Taipei, Taiwan AIMS: Previous studies on the WHOQOL-BREF showed that the four domains in the WHOQOL-BREF are highly correlated and suggested that one general factor might be sufficient for the WHOQOL-BREF. Indeed, recently Rasch analysis also showed that after deleting three items, the remaining 23 items in the WHOQOLBREF is unidimensional. To constructing shorter versions of the WHOQOL-BREF for practical use, the aim of the present study is to selecting fewer items from the standard WHOQOL-BREF for six disease populations under the unidimensional construct. By use of Rasch analysis (rating scale model), we plan to choose about ten items for each disease group. METHODS: Thirteen thousand and eight participants from the 2001 National Health Interview Survey (NHIS) database are included in this study. The six diseases diagnosed by doctors are cardiac disease, lung disease, hypertension, peptic ulcer, sinusitis, and liver disease. The sample sizes are 181, 235, 486, 487, 376, and 500 respectively. The six samples are analyzed separately. Rasch analyses are performed using WINSTEPS. Item with the largest infit or outfit MNSQ is deleted and the estimation is repeated with the remaining items until all infit and outfit MNSQs are not larger than 1.40. Then the easiest and most difficult items are chosen as the lower and upper items alone an item difficulty continuum. The remaining ideal eight items should be evenly distributed alone the continuum. Items nearest the ideal item difficulty position are selected. However, it is possible that fewer items are selected than we expected if the items are far away from the ideal position. RESULTS: Seven to ten items are selected for the six disease groups to form the shorter versions of the WHOQOL-BREF respectively. The result shows that the distributions of item difficulties are different across disease groups and the selected items are also different. CONCLUSIONS: The current study constructs shorter versions of the WHOQOL-BREF for six disease groups. These shorter versions are more convenient to be used in practice, especially in clinical evaluation.

14/Abstract 1402 Differential Item Functioning (DIF) of the Minnesota Living with Heart Failure Questionnaire (MLHFQ) Across 21 Versions Olatz Garin, Ferrer Montse, Angels Pont, Josué Almansa, Health Services Research Unit, Institut Municipal d'Investigació Mèdica, Barcelona, Spain, Eric Van Ganse, Faculte de Medecine Lyon Sud, Lyon, France, Ingela Wiklund, Outcomes Research, AstraZeneca R&D, Moelndal, Sweden, Jordi Alonso, Health Services Research Unit, Institut Municipal d'Investigació Mèdica, Barcelona, Spain AIMS: The Minnesota MLHFQ has shown good metric properties in several language/country versions, but the equivalence across items has never been examined. The purpose of this study is to test DIF between USA original version and other 20 country versions: Australia, Brazil, Canada, Czech Republic, Denmark, Finland, France, Germany, Great Britain, Hungary, Israel, Italy, Netherlands, Norway, Poland, Slovakia, Spain, Sweden, Switzerland, USA, and Yugoslavia METHODS: The MLHFQ has 21 items and 2 dimensions (physical and emotional, with 8 and 5 items respectively). Data from 8 different studies was pooled in a common database of 3266 patients to evaluate DIF of the 13 items included in the dimensions. DIF was assessed by two different approaches: ordinal logistic regression, and polytomous Item Response Theory (IRT) models. In the first method models were constructed with the item response as dependent variable following a steps process to include independent variables: scale score, country, and their interaction. Comparison of the stepped models (likelihood ratio test) allowed the distinction of uniform or non-uniform DIF (per item per country). For estimating DIF with IRT the graded response model was chose; and analyses were conducted with MULTILOG software RESULTS: The percent of missing for each item was very low; correlations between items and their hypothesized scale were similar, and most of

16/Abstract 1524 Psychometric Properties of the Migraine-Specific Quality-of-Life Questionnaire (MSQ) V2.1 During Migraine Prophylaxis Treatment Jason C. Cole, Peggy Lin, QualityMetric, Lincoln, RI, Marcia F. Rupnow, Outcomes Research (Neurology), Ortho-McNeil Janssen Scientific Affairs, L.L.C., Titusville, NJ AIMS: MSQ V2.1 is a 14-item disease-specific measure of selfreported quality of life on three dimensions: role restrictive (RR), role preventive (RP), and emotional functioning (EF). As the MSQ was

A-74

developed and validated on studies evaluating the acute treatment of migraine, it was important to determine if the MSQ can be appropriately utilized in studies evaluating the preventive treatment of migraine. METHODS: 850 migraineurs completed the MSQ during two randomized clinical trials investigating a migraine prophylaxis. Patient population was 89% female, 89.3% Caucasian, with a mean age of 40.9 years (SD=10.7 years). Confirmatory factor analysis (CFA) tested the dimensionality of the 3 MSQ scales at once, including comparing baseline and end point latent structures. Next, item response theory (IRT) examined item properties and any bias for age or gender. Classical psychometrics were examined, including item- and scale-level reliability, convergent and discriminant validity. RESULTS: CFA confirmed the known MSQ factor structure is appropriate for the current sample (strong datamodel fit and all loadings>.70), even at study end point for those on prophylaxis or placebo. IRT estimates showed a good range of scores with appropriate standard errors. Two items showed differential item functioning (DIF) between men and women, but they appeared to show true gender differences rather than bias. Item- and scale-level psychometrics were all appropriate (e.g., alphas were .91, .83, and .83 for RR, RE, and EF, respectively; MSQ dimensions correlated with convergent measures.76-.92 and less with discriminant measures .27.29). CONCLUSIONS: The MSQ was a psychometrically strong measure of migraine-specific QoL among migraineurs receiving preventive treatment. Some interesting findings found that 2 items had gender DIF, and evidence that the EF frustration item was strongly related to RR; previous research has similar findings. Overall, the current findings suggest that the MSQ is appropriate for use in migraine prevention study.

correlations to existing tools (BAI, HADS-A, and STAI: r=0.61 to 0.71), discriminant validity was suggested by significant mean score differences between patients suffering from anxiety disorders and patients suffering from other medical conditions (p<0.001). CONCLUSIONS: We conclude that the German A-CAT is an efficient, reliable, and valid tool for assessing anxiety in patients suffering from anxiety disorders or other medical conditions. Future IRT- and CAT research is needed for norming the A-CAT, contentbalancing the item selection algorithm, and building a common metric for different anxiety measures to allow comparisons and crosscalibrations to existing tools.

18/Abstract 1756 Minimally Important Differene for the Incontinence Severity Index (ISI) Rodney L. Dunn, John T. Wei, Urology, Dan Morgan, Obstetrics/Gynecology, University of Michigan, Ann Arbor, MI AIMS: To determine the minimally important difference for each domain of the ISI. METHODS: The ISI is a 10-item diseasespecific, validated, HRQOL instrument for patients with urinary incontinence. It provides an overall incontinence score and domains for stress (SUI) and urge (UUI) incontinence, pad use, and incontinence-related bother. A cross-sectional cohort of 447 women who underwent surgery for stress urinary incontinence at a single institution participated in this IRB approved survey study. We applied both the distribution (0.5 standard deviations) and anchorbased approaches to estimate the minimally important difference (MID). The anchors chosen were 1-point difference in the ISI Bother domain and a 2-point difference in the Hunskaar measure. RESULTS: The MID for each domain were fairly consistent between the distribution- and anchor-based approaches (Table). Based on these two methods, a change of 4 points in the Total ISI, 2 points for the SUI and UUI, and 1 point for the Pad Use domain scores are considered to be relevant. CONCLUSIONS: The minimally important differences for the domains of the ISI have been established and will provide the necessary context for determining when changes in treatment are either beneficial or necessary.

17/Abstract 1789 Validation of a German Computerized Adaptive Test for Measuring Anxiety (A-CAT) Janine Becker, Science Team, QualityMetric Incorporated and Health Assessment Lab, Waltham, Massachusetts, Herbert Fliege, Rueya D. Kocalevent, Psychosomatic Medicine, Charite, Humboldt University Hospital, Berlin, Germany, Matthias Rose, Jakob B. Bjorner, Science Team, QualityMetric Incorporated and Health Assessment Lab, Waltham, Massachusetts, Burghard F. Klapp, Psychosomatic Medicine, Charite, Humboldt University Hospital, Berlin, Germany

MID by distribution and anchor based-methods

AIMS: In 2004 a German Computerized Adaptive Test to measure anxiety (A-CAT) has been developed on the basis of Item Response Theory (IRT). The item bank (50 items) serving the A-CAT had been built using items of a set of established questionnaires administered to 2,348 psychosomatic patients. Items were expert-rated for content and empirically tested for their psychometric properties including checks of unidimensionality, evaluation of item response curves (IRC), and item parameters. In previous studies the A-CAT had been tested for feasibility and validity in small patient samples and simulation studies. The present study aims at validating the A-CAT in a larger patients sample. METHODS: The A-CAT was administered on pocket PCs to n=348 consecutive patients suffering from anxiety disorders (n=72) or other medical conditions (n=276). For validation purposes, the BAI, STAI, HADS, and BSF were administered. Items were adaptively presented to the patients until a predefined stopping rule (reliability r>=.9) was reached. RESULTS: The A-CAT was fast to complete and well accepted. The precise IRT-based CAT score could be estimated after an average administration of 6 items (SD=4.2, range: 4-41 items). In 95% of the cases 12 items or less were needed for a reliable theta score estimate and four items accounted for 45% of all item administrations. Convergent validity of the A-CAT was supported by moderate

Domain ISI Total Score SUI UUI Pad Use

0.5 ISI Possible Hunskaar Standard Bother Range (2 pt inc) Deviation (1 pt inc) 0-32

4.53

3.43

3.88

0-12 0-12 0-8

1.79 2.04 1.19

1.28 1.36 0.78

1.30 1.63 0.94

19/Abstract 1788 Comparison of SF-36 Utility Estimates: The Impact of Comorbidity on Change in Utility Michael B. Nichol, Tara K. Knight, Tom Dow, Joanne Wu, Pharmaceutical Economics and Policy, University of Southern California, Los Angeles, CA, Femida Gwadry-Sridhar, Medicine, University of Western Ontario, London, Ontario, Canada AIMS: 1) To determine the impact of an increase in comorbidity on change in SF-36 based utility estimates from Baseline (T0) to Year 2

A-75

(T1); 2) To compare the impact across six utility estimates. METHODS: The study population consisted of patients from a large health maintenance organization dataset (n = 6,923). Six utility estimates were derived for both T0 and T1 SF-36 scores, including Nichol OLS, SF6DSG, SF6DVAS, Lundberg, Shmueli, and Visual Analogue Scale (VAS). Chronic Disease Scores (CDS) were computed for each patient for T0 and T1 based on the number of comorbidities in the year prior to assessment. Subjects were categorized into CDS Increase (CDS Inc) or No Increase (No Inc) groups based on whether or not they had an increase in CDS score from T0 to T1. Change in utility scores and effect sizes were calculated. T-tests were used to compare group differences in change in utility over time. Effect size (ES) was calculated as the difference in utility between T1 and T0 divided by standard deviation of utility at T0. RESULTS: For all estimates, the CDS Inc group had a greater decline in utility over time than the No Inc group (all p<.01). However, none of the effect sizes were large enough to indicate any meaningful change per Cohen_s definition. Utility scores at T0 were lowest for both groups using the SF6DVAS estimate (CDS Inc=0.63, No Inc=0.66) and highest using Nichol OLS (CDS Inc=0.80, No Inc=0.82). Change in utility score from T0 to T1 was largest for both groups using the Lundberg (CDS Inc: Change=-0.02, ES=-0.17; No Inc: Change=-0.01, ES=-0.09) (p<.001) and Shmueli estimates (CDS Inc: Change=-0.02, ES=-0.14; No Inc: Change=-0.007, ES=-0.05) (p<.0001). CONCLUSIONS: An increase in patient comorbidity has a statistical significance, but does not appear to have a meaningful impact on utility. As this population already suffers from multiple conditions, an increase in comorbidity may not pose the burden that it might in a healthier cohort.

21/Abstract 1622 The Impact of Huntington's Disease on the Quality of Life Carers Steve M. Smith, Health, University of East Anglia, Kings Lynn,, Norfolk, England, UK. AIMS: Huntington_s disease (HD) is an autosomal dominant familial neurodegenerative disorder with typical mid-life onset. It manifests with neurologic (movement disorder), cognitive and psychiatric disturbances. Effective treatment is lacking. Despite recognised family implications, studies on the perceived impact of HD on carers are scarce. The aim of this pilot study was to explore the impact of HD on carers, and identify specific factors associated with larger carer impact. METHODS: A postal survey was conducted among HD carers enrolled from a UK university HD clinic. The impact of HD on carers was assessed by the HD Quality of Life battery for carers (HDQoL_C), an HD carer specific questionnaire that yields scores between 0-100 (0 = better). Relationships between HDQoL_C scores and demographic and HD related variables were explored by Spearman correlations and by comparing HDQoL_C scores across subgroups RESULTS: Most participants were women (59%) and the spouse of a person with HD (79.3%). Mean (SD) age was 55 (12) years. Median (IQR) HDQoL_C scores were 50 (32-60). The number of years HD had been known in the family (Spearman r -0.008), duration of care-giving (0.060), carer age, income (0.031) or amount of time weekly spent caring (-0.246). did not correlate significantly with HDQoL_C scores. The amount of social contacts did (- 0.427). Those married or living as married had significantly (p = 0.0001) lower (better) HDQoL_C scores than non-spousal carers, and male carers scored better than female carers (p = 0.038). CONCLUSIONS: This study illustrates that there is a vast impact associated with caring for someone with HD, and suggests the importance of facilitating social contact. Clinical HD trials should consider not only patient- but also caregiver- reported outcomes. Additional studies are needed to gain a better understanding of the impact of caring for people with HD and how this impact compares with that experienced by carers of people with other chronic disorders.

20/Abstract 1162 The Extent and Breadth of Benefits From Participating in Chronic Disease Self-Management Courses: A National PatientReported Outcomes Survey Sandra Nolte, Gerald R. Elsworth, CIRCLE, RMIT University, Melbourne, VIC, Australia, Amanda L. Springer, Medicine, University of Melbourne, Parkville, VIC, Australia, Andrew J. Sinclair, Exercise and Nutrition Sciences, Deakin University, Burwood, VIC, Australia, Richard H. Osborne, Medicine, University of Melbourne, Parkville, VIC, Australia

22/Abstract 1194 The Relationship Between Co-Morbidity and Health-Related Quality fo Life in Those with Symptomatic Cerebral Small Vessel Disease Carole L. White, Medicine, University of Texas Health Sciences Center, San Antonio, Texas, Leslie A. McClure, Biostatistics, University of Alabama, Birmingham, Alabama, Donna J. Brauer, Neurology, University of Minneapolis, Minneapolis, Minnesota

AIMS: To quantify the benefits that people receive from participating in self-management courses and identify subgroups that benefit most. METHODS: People with a wide range of chronic conditions attending self-management courses (N=1,341 individuals) were administered the generic Health Education Impact Questionnaire (HEI-Q). Data were collected before the first session (baseline) and at the end of courses (follow-up) resulting in 842 complete responses. The median (interquartile range) age was 64 (54 to 73) years and most participants were female (75%). Outcomes were categorized as Substantial improvement (Effect Size, ES > 0.5), Minimal/No change (ES -0.49 to 0.49) and Substantial decline (ES < -0.5). RESULTS: On average, one third of participants reported substantial benefits after attending a self-management course. Proportions of participants reporting substantial benefits ranged from 49% in Skill and technique acquisition to 27% in Health service navigation. Stratification by gender, age and education showed that younger participants were more likely to benefit, particularly young women. No further subgroup differences were observed. CONCLUSIONS: Given that the health of people with chronic diseases tends to decline, this evaluation is reassuring in that about one third of participants coming from a wide range of backgrounds receive substantial improvements in their self-management skills.

AIMS: Persons with stroke often have co-existing health problems which influence recovery and increase risk of recurrent stroke. Comorbid conditions, in particular hypertension and diabetes, are prevalent in those with cerebral small vessel disease. Our objectives were to estimate the impact of co-morbid conditions on the stroke survivor's health-related quality of life (HRQL) and to identify domains most affected. METHODS: Data come from the Secondary Prevention of Small Subcortical Strokes Study, an ongoing NIHfunded international trial with the objective of defining efficacious therapies for prevention of stroke recurrence and cognitive decline in those with symptomatic subcortical stroke. HRQL is measured with the Stroke Specific Quality of Life Scale (SSQOL) at baseline and annually. Overall quality of life (SS-QOL) and domain scores range from 1 to 5 (best). Using multivariable linear regression, data from the baseline interview (n=757) were used to examine the relationship between the number of co-morbidities and HRQL. RESULTS: The

A-76

majority (93%) was diagnosed with at least one co-morbid condition, with hypertension most prevalent. The differences between those with stroke only and those with stroke and 3 co-morbidities were, on average, 0.5 SD lower on the sub-scales of the SSQOL with work, social, family, mobility, arm function, and overall SS-QOL most affected. Mean SS-QOL score for those with stroke only was 4.47 with a significant mean decrease of 0.17 (95% CI: 0.10, 0.23) for each additional condition. After adjusting for age, gender, race, functional status and depression, the relationship between SS-QOL and the number of co-morbid conditions remained significant but was attenuated. CONCLUSIONS: Co-morbid health problems adversely influence the stroke survivor's HRQL affecting both physical and social aspects. Future analyses will investigate severity of comorbidities and their management to better characterize the relationship between stroke, co-morbid conditions, and HRQL and to identify possible interventions.

AIMS: The aim of this study was to investigate to what extent the questions of the Parkinson_s Disease Questionnaire (PDQ-39) could be identified and quantified according to the International Classification of Functioning, Disability and Health (ICF) core concepts. METHODS: The PDQ-39, consists of 39 questions and was developed for assessment of health among patients with Parkinson_s disease. It has become the most used instrument in Parkinson research. The analyses were made by experienced researchers within nursing, occupational therapy and physiotherapy. Stepwise analyses, comprising individual analyses as well as group discussions on consensus was used. RESULTS: The result revealed that 23 of the questions of the PDQ-39 were concerned with health and well-being in relation to body functions, and of these 19 questions were concerned with mental functions. Sixteen questions were related to activity and participation, and of these 10 questions were concerned with mobility and 5 with ADL. Only one question was concerned with involvement in more social activities. Overall, 10 questions involved environmental factors and 3 questions were difficult to classify into specific categories and therefore only could be analysed in relation to the block the belonged. CONCLUSIONS: In order to get amore global view of health and well-being among patients with Parkinson_s disease a revision of the PDQ-39 is needed. More questions that considers involvement in life situations have to be developed.

23/Abstract 1701 Quality of Life in Parkinson's Disease Eva Dragomirecka, Social Psychiatry, Psychiatric Centre, Prague, Czech Republic, Anna Blabolova, College of Education, Charles University, Prague, Czech Republic, Jitka Bartonova, Social Psychiatry, Psychiatric Centre, Prague, Czech Republic AIMS: Parkinson's disease (PD) is a neurological disorder with great impact on quality of life. The aim of this study was to investigate which aspects of patients' lives were most affected and which symptoms contribute the most to the quality of life decrease. METHODS: Single postal survey using the WHOQOL-BREF instrument supplemented by items focused on symptoms and specific problems of PD. Sample: Questionnaires were sent to all members (N = 1051) of the Czech Parkinson Society, 46 % of whom responded. Analyzed were data from 457 persons with PD. 39 % of the sample were women, the mean age was 69 years, with range from 41 to 93 years. Most of the sample (78 %) has experienced PN longer than 6 years. Analysis: The WHOQOLBREF domain scores of persons with PD were compared with scores of the representative sample of 325 older adults living in Prague. Differences were tested using one-way ANOVA. To identify impact of selected symptoms on quality of life we ran regression analyses separately for each of four WHOQOL-BREF domains. RESULTS: As the most serious and frequent symptoms were considered bradykinesia, rigidity, fatigue, walking and writing problems. We found significant gender differences in evaluation of symptoms severity: while for men problems with writing and communication with others were more distressing, women evaluated pain as the more distressing item. The comparison with the control sample showed significantly lower quality of life in two domains (Physical health and Psychological health) both for men and women. In addition men had significantly lower quality of life in Social relationship domain. The strongest predictor for Social relationship domain was speech pathology that was experienced by more than half of respondents. CONCLUSIONS: Our results showed the negative impact of communication problems on quality of life in patients with PD, especially in men. It confirms the importance of complex rehabilitation care including speech therapy. Supported by research project MZCR MZ0PCP2005

25/Abstract 1432 Health Related Quality of Life in Elderly with Stroke Maria José D. Diogo, Keila C. T. Cruz, Nursing, State University of Campinas, Campinas, São Paulo, Brazil AIMS: To evaluate health-related quality of life in elderly subjects with history of stroke and to investigate the relationships between quality of life, social-demographic variables and health. METHODS: Exploratory and descriptive study with 44 elderly stroke survivors. Health related quality of life was evaluated by means of the questionnaire SF-36 - "The Medical Outcomes Study 36-item shorts-form health survey". Analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA) were used to verify the influence of age, presence of a caregiver and the stroke in the dimensions of SF-36. RESULTS: Functional capacity achieved the lowest mean score, whereas general state of health obtained the highest score among the dimensions of SF-36. Chronological age was negatively correlated to functional capacity and positively correlated to physical aspects and emotional aspects. The following variables affected different dimensions of SF-36: chronological age, parestesis, hemiparesis and visual deficiency. Results from MANOVA showed that only hemiparesis interfered in the dimension of functional capacity. CONCLUSIONS: According to our results, socialdemographic variables and current sequels of stroke affect health related quality of life of elderly stroke survivors, and functional capacity is the most impaired dimension of SF-36. Actions aimed at improvement and rehabilitation of functional capacity and addressing specific needs of elderly people can benefit that population.

26/Abstract 1692 24/Abstract 1758 An Assessment of the Structure of Self-Reported Health in Parkinson's Disease Using the SF-36 Peter Hagell, Health Sciences, Lund Uniersity, Lund, Sweden, Anna Lena Törnqvist, Neurosurgery, University Hospital, Lund, Sweden, Jeremy Hobart, Clinical Neuroscience, Peninsula Medical School, Plymouth, UK

Analyse of the PDQ-39 in Relation to the International Classification of Functioning, Disability and Health (ICF) Core Concepts Carita K. Nygren, Occupational Therapy, Health Sciences, Lund, Sweden, Peter Hagell, neurology, Caring sciences, Lund, Sweden

A-77

vs.23%,n.s.). Sensitivity analyses were performed using stricter cutoffs for critical BI levels (< 25 and 0 score points), but the effects of the various therapies on the rate of critically ill patients remained unchanged in most cases. CONCLUSIONS: Specific QoL dimensions rather than global QoL were affected by medical therapies. Body image turned out to be particularly sensitive dimension. When considering specific QoL therapies in clinical settings (e.g. physiotherapy, psychotherapy, social rehabilitation) ongoing medical therapy at the time of QoL assessment has to be taken into account.

AIMS: Factor analytic studies in general populations and several diagnostic groups have found a 2-dimensional pattern among the 8 health constructs (scales) of the SF-36, congruent with its proposed physical and mental component summary measures (PCS and MCS, respectively). However, studies in neurological disorders such as stroke, multiple sclerosis and motor neuron disease have failed to replicate this structure. This paper assesses the structure among the 8 SF-36 scales in people with Parkinson's disease (PD). METHODS: Postal survey data of 202 people diagnosed with PD at a Swedish university hospital were used for the main analyses. Mean (SD) age and disease duration were 69.8 (10.0) and 8.7 (6.6) years, respectively, median (IQR; range) Hoehn & Yahr stage of PD was 3 (2-4; 1-5) and 49% experienced dyskinesias. The 8 SF-36 scales were factor analyzed according to recommendations by the developers (principal component extraction, varimax rotation, eigenvalue >1 criterion). RESULTS: All 8 SF-36 scales clustered together in 1 component. Following re-analysis specifying that 2 factors should be extracted, Physical Functioning loaded with Role Physical and Role Emotional, whereas the other scales loaded together. Because PD is associated with motor complications that may be accompanied by fluctuating symptoms of depression and anxiety, separate post-hoc analyses of data from people with and without dyskinesias were performed. Clustering of scales was very similar to that of the mother sample among people with dyskinesias. No clear clusters were seen among those not reporting dyskinesias, even when pre-specifying 2 components to be extracted; the scales loading strongest on the 2 factors were Role Emotional and Mental Health, respectively. CONCLUSIONS: The physical and mental health constructs are disrupted in PD, which in part may relate to motor and/or mood fluctuations. Regardless of its cause, this study argues against using the PCS and MCS in PD. Further research is warranted to understand the health impact of PD.

28/Abstract 1304 Is There a Difference in the Methodological Quality of HealthRelated Quality of Life Assessment Between Academic and Industry Sponsered Studies? Evidence From Randomized Controlled Trials in Oncology Fabio Efficace, Corneel Coens, Andrew Bottomley, Quality of Life Unit, EORTC Data Center, Brussels, Belgium AIMS: Little evidence exists regarding possible differences in terms of the quality of health-related quality of life (HRQOL) evaluation of industry sponsored trials versus academic sponsored ones. METHODS: Based on a predefined set of criteria, 159 randomized controlled trials (RCTs) with a HRQOL endpoint published between 1990 and 2004 were systematically identified and analyzed. These were focused on four major disease sites (breast, colorectal, lung and prostate). The methodological quality of HRQOL assessment was evaluated by the "Minimum standard checklist for evaluating HRQOL outcomes in cancer clinical trials". This checklist consists of 11 key issues related to the methodology used to assess HRQOL. These items can be scored as 'yes' (scoring 1) or 'no' (scoring 0) or N/A, with higher scores indicating better quality. Descriptive statistics and multiple regression analysis were used to investigate possible relationships between industry sponsored trials and the adjusted checklist score (i.e. raw score divided by the number of applicable questions). RESULTS: Out of the 159 RCTs reviewed, 93 were industry sponsored studies and 66 were not. The median adjusted checklist score for the industry sponsored studies was 0.6 (range 0.3 to 1.0) while the median for academic studies was 0.7 (range 0.2 to 1.0). Linear regression analysis with industry sponsorship as a dependant variable and adjusted checklist score as independent revealed no significant relationships (P=0.07). This relationship remained non-significant when also including, in a multiple regression analysis, additional covariates: year of publication, endpoint (primary vs. secondary), cancer disease site (breast, colorectal, lung and prostate), sample size of the trial and HRQOL difference between treatment arms (yes vs. no). CONCLUSIONS: This study suggests that the there is no difference, in terms of methodological quality of HRQOL, between academic and industry sponsored RCTs in oncology.

27/Abstract 1309 Medical Therapy in Breast Cancer Patients is Associated With Distinct Deficits in Quality of Life:Univariate and Sensitivity Analysis Following a Prospective Implementation Study Monika Klinkhammer-Schalke, Tumorcenter Regensburg, Regensburg, Germany, Germany, Michael Koller, Regensburg, Germany, Christoph Ehret, Brunhilde Steinger, Brigitte Ernst, Ferdinand Hofstädter, Wilfried Lorenz, Tumorcenter Regensburg, Regensburg, Germany AIMS: In a prospective Qol study deficits in quality of life (QoL) are not only due to underlying illness, but may be consequences of various forms of cancer therapy. METHODS: 170 breast cancer patients, age 58 (34-86), UICC (0-4) 4,41,41,9,7%, were recruited. QoL diagnostics and -therapy were implemented along a clinical pathway including clinicians, the patients' self reports(EORTC QLQ C30+BR23), coordinating practitioners (n=38) and professional helpers (physiotherapists,psychologists, social workers etc.) for treatment of QoL deficits. QoL scores of <50 points on a 0=very bad to 100=very good scale were considered therapy-relevant. Sensitivity analyses also used more strict criteria of <25 and 0 score points. RESULTS: Concerning surgery 100/164 patients received breast conserving therapy (BCT), 64/164 mastectomy (ME). At the time of QoL assessment, 20/168 underwent chemotherapy, 7/168 radiotherapy, and 68/168 endocrine therapy (ET). Score points of <50 in global QoL were observed in 15% BCT versus 25.4% ME patients (n.s.), as well as in patients undergoing chemotherapy (15%), radiotherapy (19.5%) and ET (21.3%). Critical score points (< 50) regarding body image (BI) were observed in 7% BCT-patients, but 31% ME-patients, p<.001. Similar differences were not seen in emotional functioning (38% vs.34%,n.s.) or fatigue (37%

29/Abstract 1356 Validation of the Functional Assessment of Cancer TherapyBrain (FACT-BR) in Patients with Recurrent High-Grade Gliomas Jennifer L. Beaumont, Center for Outcomes, Research and Education (CORE), David Victorson, CORE, Evanston Northwestern Healthcare, Evanston, IL, Amy H. Peterman, Psychology, University of North Carolina, Charlotte, NC, David Cella, CORE, Evanston Northwestern Healthcare, Evanston, IL, Astra M. Liepa, Global Health Outcomes, Eli Lilly and Company, Indianapolis, IN, Howard A. Fine, Neuro Oncology, National Cancer Institute, Bethesda, MD

A-78

qualitatively by 2 researchers. Text fragments will be categorized inductively and subsequently compared and combined with aforementioned frameworks in an iterative process. RESULTS: The interview was pilot tested in 4 respondents. All respondents were capable of completing the think aloud interview. The probes provided additional information and the cognitive frameworks were applicable. Cognitive processes differ for answers based on an automatic level of thinking and answers requiring reasoning. Results from interviews conducted with cancer patients (May-August 2006) will be presented. CONCLUSIONS: This study will lead to a better understanding of self-reported change in QL and will provide insight into the usefulness of designs commonly used to measure change.

AIMS: The FACT-Br was developed to assess health-related quality of life in patients (pts) with primary brain tumors. Four additional items were added to the previously validated 19-item brain subscale (BrS), based on interviews with expert providers upon development of a symptom index. Recently, the FACT-Br (with 23-item BrS) was administered to pts with recurrent high-grade gliomas participating in a phase II trial of enzastaurin for the purpose of validating it in this specific population. METHODS: It was administered prior to start of therapy, at 3 weeks, at 6 weeks and then every 6 weeks thereafter while on therapy. Reliability was assessed using Cronbach_s alpha. Construct validity was assessed using criterion validity and responsiveness to change. Distribution- and anchor-based methods were used to estimate minimally important differences (MIDs). RESULTS: FACT-Br data were available from 115 of 120 pts. 72% had a diagnosis of glioblastoma multiforme and 68% had a Karnofsky performance status rating (PSR) of >80 at baseline. FACT-Br data were available for 104 pts at baseline, 87 at 3 weeks and 60 at 6 weeks. Cronbach_s alphas for the 23-item BrS and 37item Trial Outcome Index (TOI = BrS + Physical Well-Being + Functional Well-Being) at the three assessments were 0.86-0.90 and 0.92-0.95, respectively. Both scales differentiated patients by PSR at all assessments (all p<0.05), with effect sizes (mean difference / standard deviation [SD]) for group comparisons of 0.34 _ 0.92. The BrS and TOI both demonstrated responsiveness to change in PSR (p=0.004 and 0.002, respectively) at 6 weeks. Effect sizes (mean change / SD at baseline) were -0.65 to -0.75 for pts with worsened PSR and -0.03 to -0.14 for pts with unchanged or improved PSR. MIDs were estimated as 5-7 points for the BrS and 6-11 points for the TOI. CONCLUSIONS: The FACT-Br has demonstrated reliability and validity in pts with recurrent high-grade gliomas. The BrS and TOI may be useful endpoints in the assessment of therapies for recurrent high-grade glioma.

31/Abstract 1678 Question Struture and Translation Issues in EORTC QLQ System Colin D. Johnson, Surgery, University of Southampton, Southampton, UK, Michael Koller, Center for Clinical Studies, University of Regensberg, Regensberg, Germany, Karen West, Quality of Life Unit, EORTC, Bruxelles, Belgium AIMS: In the EORTC QLQ system there is no uniform wording for individual items; this may cause confusion in translation. For example two auxiliary verbs may require identical or different translations. Words with disparate meanings are problematic. The aim of this study was to simplify translation by identifying standard question formats, and problem words METHODS: One person reviewed the English versions of all 20 EORTC QLQ site-specific questionnaires. Items were grouped according to grammatical structure and were compared for differences in meaning. Problem words were noted that had produced difficulty in translation. Translations in French, German and Spanish were reviewed selectively to identify related variations in item structure and meaning. Findings were cross-checked, reviewed and discussed by the Translations Group RESULTS: Most items begin _Have you had.. either symptoms, or problems (physical or emotional difficulties or troublesome symptoms). Items beginning _Have you felt.. usually refer to an emotional state but could refer to physical symptoms. A minority of items begin Did you have.. (a symptom). Translation varied with target language, from a single rendering of both forms (French) to three variations of _Have you had.. (German). Translation of sexual items was sometimes difficult. Problem words included sore, trouble, problems, bother. A few items were complex in structure, with a conditional element defining the subsequent question and response. Such items pose greater difficulty for translation and for comprehension than a simple question CONCLUSIONS: In English versions, items relating to symptoms or _problems should begin Have you had.. In British English the phrase Did you have.. refers to a completed event or one experience whereas Have you had.. refers to repeated events or ongoing states. Items dealing with emotional experience should begin Have you felt. . This phrasing should be avoided for other types of item (physical state or symptom, physical experience). Problem words (in English) should be avoided or if essential, should be clearly defined

30/Abstract 1337 Understanding Changes in Quality of Life in Cancer Patients: A Cognitive Interview Approach Elsbeth F. Bloem, Medical Psychology, Mechteld R. Visser, Family Practice, Academic Medical Center, Amsterdam, The Netherlands, Florence J. van Zuuren, Clinical Psychology, University of Amsterdam, Amsterdam, The Netherlands, Caro C. Koning, Radiotherapy, Mirjam A. Sprangers, Medical Psychology, Academic Medical Center, Amsterdam, The Netherlands AIMS: To increase our understanding of how cancer patients evaluate their quality of life (QL) over time we will study the cognitive processes underlying these evaluations in 3 designs commonly used to assess change: (a) baseline follow-up, (b) transition and (c) thentest design. Specific aim is to examine the tenability of the assumptions inherent to these designs: stability of the evaluation process over time (design a); correct recall and accurate incorporation of baseline QL (design b); correct recall of baseline QL and stability of the evaluation process between follow-up and thentest (design c). METHODS: Cancer patients will be interviewed prior to and after completion of radiotherapy. We will enroll patients whose symptom level is expected to improve, deteriorate or remain stable. A purposive sampling strategy will be used until data saturation is reached; expected sample size is 90. Cognitive processes will be elicited by the Three Step Test Interview, in which patients think aloud while completing QL questions and are subsequently probed to elicit more information. Probes are based on frameworks of Tourangeau (2000) and Rapkin & Schwartz (2004). We will use 7 EORTC QLQ-C30 items reflecting both global and specific content and physical as well as psychosocial domains. Data will be analyzed

32/Abstract 1034 Health Related Quality of Life in Chronic Low Back Pain Patients Treated with Instrumented Fusion Signe Berit Bentsen, Nursing Education, Stord/Haugesund University College, Haugesund, Norway AIMS: The aim of this study was to examine the health related

A-79

quality of life in chronic low back pain patients treated with instrumented fusion, based on the patients` subjective responses after surgery. METHODS: The samples consisted of 101 chronic low back pain patients treated with instrumented fusion between 1993 and 2000 (responserate 80%). The patients filled in a questionnaire consisting of 1)Demographic data, 2)SF-36 Health Survey. 71% of the sample was female, the mean age was 46 years (SD8.9), age range 26-60. RESULTS: SF-36 scores; Physical Functioning=64(SD27.0), Role Physical Functioning=39(SD44.6), Bodily Pain=54(SD33.2), General Health=65(24.7), Vitality54=(SD23.1), Social Functioning=68(SD32.2), Role Emotional Functioning=77(SD40.1), and Mental health=74(SD20.9). The chronic low back pain patients had 10 points lower score than the general population on Physical Functioning, Role Physical Functioning, Bodily Pain, General Health and Social Functioning. CONCLUSIONS: The SF-36 scores show that the chronic low back pain patients had less Physical Functioning, Role Physical Functioning, General Health and Social Functioning, and more Bodily Pain compared with the general population.

34/Abstract 1416 Validation of A Priority Tool for Hip/Knee Replacement Antonio Escobar, Unidad de Investigacion, Hospital de Basurto, Bilbao, Bizkaia, Spain, Jose M. Quintana, Unidad de Investigacion, Hospital de Galdakao, Galdakao, Bizkaia, Spain, Berta Ibañez, Oficina de apoyo metodologico, Fundacion Vasca de Innovacion e investigaciones Sani, Sondika, Bizkaia, Spain, Amaia Bilbao, Oficina de apoyo metodologico, Fundación Vasca de Innovación e Investigación Sanita, Sondika, Bizkaia, Spain AIMS: One criteria for judging the effectiveness of a health system is that patients with urgent conditions should receive services before those with less urgent conditions. An option for making waiting lists more rational is the criteria for prioritization. The aim of this study was to study the validity of a developed explicit prioritization criteria for patients who are to be placed on waiting list for total hip or knee replacement due to primary osteoarthritis METHODS: Criteria were developed using a modified Rand Appropriateness Method (RAM). We convened a panel of nine members who scored the scenarios created by the research team and by patient focus groups. The variables of interest were: mechanical pain, walking functional limitation, pain on rest, other functional limitations (putting on shoessock-stockings, sitting to standing or managing stairs-ramps), abnormal range of motion, threatens to social role and existence of pathologies that could improve by surgical management of the affected joint. The priority weights of the variables and their levels were synthesized using the optimal scaling model to transform the RAM scores to a 0-100 score of prioritization. The modified RAM defines 3 categories, urgent preferent and ordinary. We used the WOMAC questionnaire to evaluate the validity of the tool. Statistical analysis was performed by means of ANOVA (dependent variable was the score of each WOMAC dimension and the independent was the 3 groups of RAM) and Spearman correlation coefficient among priority score and each dimension of WOMAC questionnaire RESULTS: We have studied 983 patients (446 hip and 537 knee). The mean priority score was 66.5 (SD 25.1). The means (SD) of WOMAC scores were: pain 57.0 (18.5); function 65.0 (17.4) and stiffness 59.7 (23.7). There were statistically significant differences in means of the 3 RAM categories in all three dimensions of the WOMAC. The Spearman coefficients among priority score and WOMAC dimensions were 0.69 (pain), 0.79 (function) and 0.50 (stiffness) CONCLUSIONS: These data support the validity of the prioritization tool for hip and knee replacement

33/Abstract 1339 Quality of Life Among Osteoarthritis Patients and Rheumatoid Arthritis Patients Grazyna Baczyk, Dorota Talarska, Monika Kmieciak, Iwona Walkowiak, Nursing, University of Medical Sciences, Poznan, Poland AIMS: Joint disorders,particularly osteoarthritis (OA) and rheumatoid arthritis(RA) are a leading cause of disability.The aim of this study was to assess the quality of life(QoL) of OA and RA patients treated in a Outpatient Clinic Rehabilitation in Poznan,Poland.The specific questions were:does the quality of life of OA and RA patients depends on demographic variables (gender,age)and duration of the disease? Does the quality of life among OA and RA patients depends on clinical symptoms and pain. METHODS: The study sample consisted of (OA) patients,including 86 women(88%)and 11 men(12%).The sample 123 patients with RA included of 102 women(83%)and 21 men(17%).The mean age of the treated patients with OA was 56.8 years,53.38 years for RA patients.The Polish version of the Arthritis Impact Measurement Scales-2(AIMS-2) was used to assess the quality of life.AIMS-2 scores range from 0-10,with 0(representing good QoL),10 (representing poor QoL).Clinical measurements:VAS Pain Scale,Grip Strength Measurement and Morning Stiffness. RESULTS: The results showed that the mean score on the AIMS -2 for OA patients was:physical-3.53,affect-4.42,symptom -6.74,social interaction 3.33,role -4.2.Mean score on the AIMS-2 for RA patients was:physical-3.73,affect-4.48,symptom -7.09,social interaction 3.45,role - 3.63.The quality of life depended on the sex of these patients. Women of OA and RA patients scored significantly higher in physical state and symptom than men (p<0.05).Also,younger patients and suffering shorter than 5 years demonstrated higher evaluation of QoL in the physical state (p<0.05). CONCLUSIONS: This study showed that quality of life of OA and RA patients depends on gender, age and clinical variables(the assessment in most of the subscales of AIMS-2 correlated significantly with Pain, Morning Stiffness and Grip Strength for OA and RA patients).Further research into the quality of life can lead to improvement in the quality of care nursing and treatment of people suffering from osteoarthritis and rheumatoid arthritis.

35/Abstract 1413 A Systematic Review of Instruments Assessing Participation: Challenges in Defining Participation and First Results Isaline C. Eyssen, Occupational Therapy, Caroline B. Terwee, EMGO Institute, Joost Dekker, Rehabilitation Medicine, VU University Medical Center, Amsterdam, the Netherlands AIMS: Improvement of participation is an important aim of occupational therapy and rehabilitation. We aim to identify available instruments for assessing participation and to review the extent to which the instruments indeed measure participation. METHODS: A Medline search was performed. Keywords were a.o. participation, handicap, client-centered, role, social disability, reintegration. Selected were generic or disease-specific instruments that primarily aim to measure participation. Participation was defined according to the International Classification of Functioning, Disability and Health (ICF) as "involvement in a life situation". Some information on the psychometric properties had to be available. Two authors independenlty selected the abstracts and evaluated the extent to

A-80

which the instruments indeed measured participation (number of participation items). RESULTS: 3614 abstracts and 270 full-text articles were scanned. It was often difficult to distinguish between 'activities' and 'participation', i.e. to define whether an action occurred at the societal level, e.g. transport and using telephone. Based on other literature, participation was further defined as performing socially-defined tasks and roles in the domains of household, caring, work/education, leisure activities, and interpersonal relations. A distinction was made between level of participation (i.e. the extent to which a person participates in social roles), and participation problems (i.e. the extent to which a person perceives limitations or problems in role performance). Instruments measuring related concepts such as work productivity, work limitations, job satisfaction, self efficacy, autonomy, and quality of life, were excluded. Only patient-reported measures were included. The instruments that include at least 50% participation items will be presented. CONCLUSIONS: The ICF definition gave insufficient understanding of the concept participation. Further definition is required. Most instruments that aim to measure participation do not entirely measure participation.

five single-item scales. Multi-trait scaling analyses revealed high internal consistencies for the subscales with Cronbach_s alpha coefficients ranging from 0.72 to 0.87 (Symptom Experience 0.72, Body Image 0.86, Sexual/Vaginal Functioning 0.87). Convergent and discriminant validity were fulfilled with scaling errors below 3%. The QLQ-CX24 is able to discriminate between clinical subgroups. All items exhibited good compliance with less than 3% missing values. Most patients completed the EORTC QLQ-C30 and the QLQ-CX24 in less than 15 minutes (86%) and many did not require any assistance to complete the questionnaires (65%). CONCLUSIONS: These psychometric analyses support the content and construct validity as well as the reliability of the EORTC QLQ-CX24 module. This newly developed module is a useful instrument for assessing the QoL of patients treated for cervical cancer in clinical trials as well as in clinical practice.

37/Abstract 1282 The World Health Organization Quality of Life (WHOQoLBREF) Questionnaire: Translation and Validation Studies of the Iranian Version Saharnaz Nedjat, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran, Ali Montazeri, Mental Health, Iranian Institute for Health Sciences Research, Tehran, Iran, Kourosh Holakouie, Kazem Mohammad, Reza Majdzadeh, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran

36/Abstract 1305 Development of the EORTC Quality of Life Cervical Cancer Module (EORTC QLQ-CX24) Elfriede Greimel, Obstetrics and Gynecology, Medical University, Graz, Austria, Luca Incrocci, Radiation Oncology, Erasmus MCDaniel den Hoed Cancer Center, AE Rotterdam, The Netherlands, Pernille T. Jensen, Gynecology and Obstetrics, Herlev University Hospital, Herlev, Denmark, Jacqueline A. Routledge, Clinical Oncology, Christie Hospital NHS Trust, Manchester, Lancashire, UK, Camila B. Souza, Health Psychology, Methodist University of São Paulo, São Paulo, São Paulo, Brazil, Andrew J. Nordin, Gynaecological Oncology, East Kent Gynaecological Oncology Centre, Margate, Kent, UK, Young Ho Yun, Quality of Cancer Care Branch, National Cancer Center, Gyeonggi, Korea, Susanne Singer, Social Medicine, University of Leipzig, Leipzig, Germany, Dariusz G. Wydra, Gynaecology, Medical University of Gdansk, Gdansk, Poland, Vesna Bjelic Radisic, Gynecology, Medical University, Graz, Austria, Carien L. Creutzberg, Clinical Oncology, Leiden University Medical Center, Leiden, The Netherlands, Eva M. Mautner, Obstetrics and Gynecology, Medical University, Graz, Austria, Fabian Fehlauer, Radiation Oncology, University of Hamburg, Hamburg, Germany, Vlatka M. Duric, Clinical Trials Centre and Gynaeoncology, University of Sydney and Royal Hospital for Women, Camperdown, Sydney, NSW, Australia, Ann-Charlotte K. Waldenström, Oncology, Sahlgrenska University Hospital, Göteborg, Sweden, Karin Kuljanic Vlasic, Gynecology and Obstetrics, Clinical Hospital Center Rijeka, Rijeka, Croatia, Karin Bergmark, Gynaecological Oncology, Karolinska Institutet, Stockholm, Sweden, Wei-Chu Chie, Preventive Medicine, National Taiwan University, Taipei, Taiwan on behalf of the EORTC Quality of Life Group

AIMS: The objective of the current study was to evaluate the WHOQOL-BREF in Iran. It is a 26-item instrument consisting of four domains: Physical Health, Psychological, Social Relationship, Environment and two overall QOL/health items. METHODS: Forward-backward translation procedure was followed to develop the Iranian version of the questionnaire. A stratified sample of individuals aged 15 and over were completed the questionnaire in Tehran, Iran. Psychometric properties of the instrument including reliability (internal consistency, and test-retest analysis), validity (discriminate validity, convergent validity) and factorial structure were carried out using the statistical analysis. To assess responsiveness of the WHOQOL-BREF a sample of childbearing women completed the instrument 1 month before and 2 months after childbirth. RESULTS: In all 1167 individuals were entered into the study. The mean age of the participants was 36.6 (SD = 13.2) and the median year for their formal education was 12. All domains met the minimum reliability standards (alpha > 0.7) with exception for the Social Relationship (alpha = 0.55). This study demonstrated a good to excellent reliability and acceptable discriminate validity across different groups after adjusting for confounders by performing regression analysis. Since the WHOQOL-BREF demonstrated statistically significant correlation with the Iranian version of the SF36 scales in the expected direction, the convergent validity of the questionnaire was found to be desirable (all correlation coefficients >= 0.4). The factorial structure of the questionnaire assessing by correlation matrix also showed satisfactory results. CONCLUSIONS: The Iranian version of the WHOQOL-BREF performed well and was found to be a reliable, valid, and sensitive measure of health related quality of life among the general population.

AIMS: This study reports the development and the validation of the European Organisation for Research and Treatment of Cancer (EORTC) quality of life (QoL) questionnaire module designed to assess disease- and treatment specific aspects of the QoL of patients with cervical cancer. METHODS: The cervical cancer module (EORTC QLQ-CX24) was developed in a multi-cultural, multidisciplinary setting to supplement the EORTC QLQ-C30 core questionnaire. The QLQ-C30 and the cervical cancer module were administered to 377 cervical cancer patients treated with radical hysterectomy, radiotherapy and chemotherapy. Psychometric analyses were performed using data of two independent samples. RESULTS: The QLQ-CX24 consists of three multi-item scales and

38/Abstract 1429 QoL of Homeless and Emancipated Youth Barbara A. Elliott, Deidre Quinlan, Caroline A. Woods, Family Medicine and Community Health, University of MN Medical School Duluth, Duluth, MN

A-81

AIMS: Quality of Life (QOL) of homeless and emancipated youth has not been investigated, although they were in the population used to develop the YQOL. This presentation reports QOL and YQOL scores for a sample of homeless and emancipated youth. METHODS: TeenLife Center (TLC) is program that provides innovative, research-based programming for adolescents in a midsized US community. TLC-Wellness Center (WC) is a health clinic that addresses emancipated youths urgent- and primary-care needs. One measured outcome is the WC youths reported quality of life. In its initial 18 months, 90 of 148 (61%) enrollees completed a baseline YQOL; 52% (N=47) were under 19 years of age; their results are reported here. 85% were female, and 46% self-reported as ethnic minority. There were no differences in gender or ethnicity between participants and non-participants. YQOL-R (perceptual) and 5 contextual items were analyzed using the scoring template. The quantitative results were interpreted in light of the youths lifestyles, using qualitative data from their charts. RESULTS: Published YQOL scores include school-based (with/without risk behaviors) and clinical group (ADHD, depression, etc) samples. As expected, WC youths YQOL-R total and subscale scores were lower than the scores of youth who were in school, and higher than the depression and ADHD samples scores. WC youths contextual item scores were also lower than those of in-school youth with multiple risk behaviors (use substances, take sexual risks). WC youths live a street-based subsistence lifestyle; many struggle with emerging mental illnesses. Their subscale, total, and contextual item scores reflect this: their self and environment subscale scores were lowest and markedly lower than school-based youth scores. Also, many WC youth were using drugs and taking sexual risks; their low contextual item scores reflect this experience. CONCLUSIONS: Homeless and emancipated youths QOL can be measured using the existing YQOL and interpreted through their reported life experience. The scores recorded to date in this sample provide baseline comparisons to follow the population over time.

40/Abstract 1507

39/Abstract 1234

41/Abstract 1593

Development and Validation of the Portuguese Versionof the WHOQOL-OLD Module Marcelo Fleck, Eduardo Chachamovich, Clarissa Trentini, Dept. of Psychiatry and Legal Medicine, Federal University of Rio Grande do Sul and HCPA, Porto Alegre, RS, Brazil

Development and Validation of the Korean Version of the Moral Sensitivity Questionnaire (MSQ) -A Pilot StudySung-Suk Han, Nursing, the Cathoilc University, Seoul, Korea, Kyeong-Uoon Kim, Nursing, University of Wisconsin(visiting professor), Madison, WI, Kwang-Ho Meng, Medicine, The Catholic University, Seoul, Korea, Ju-Hu Kim, Education, Ajou University, Suwon, Kyounggi, Korea, Sung-Hee Ahn, Nursing, The Catholic University, Seoul, Korea, Yong-Soon Kim, Nursing, Ajou University, Suwon, Kyoungggi, Korea

Construct Validity of the MOS Short-Form 36-Item Health Survey (SF-36) among Community-dwelling Elderly in Korea Chang-wan Han, Eun-joo Lee, Health and Welfare, Woosong University, Daejeon, Korea, Masahiro Kohzuki, Internal Medicine and Rehabilitation Science, Tohoku University Graduate School of Medicine, Sendai, Japan AIMS: To evaluate construct validity of the MOS Short-Form 36Item Health Survey (SF-36) in a sample of Korean communitydwelling elderly, using structural equation modeling(SEM). METHODS: Data were collected from 1,636 participants in a stratified sample of community-dwelling elderly in Korea, interviewed face-to-face by trained interviewers. The subjects completed individual characteristics (sex, age, with or without stroke), the MOS Short-Form 36-Item Health Survey (SF-36), and the Barthel Index (BI). The construct validity of the SF-36 was assessed using Structural Equation Modeling (confirmatory factor analysis) through an evaluation of the original eight-factor structure: Physical Functioning (PF)"Role Physical (RP)"Bodily Pain (BP)"General Health (GH)"Vitality (VT)"Social Functioning (SF)"Role Emotional (RE)"Mental Health (MH). RESULTS: Structural equation modeling(SEM) analyses supported the superiority of the eight first-order factor model of health. Higherorder analyses suggested that a model with three correlated secondorder factors (physical health: PF, RP, BP, general well-being: GH, VT, general mental health: SF, RE, MH) and one third-order factor (health) provided a satisfactory fit to the data. CONCLUSIONS: These results confirm the multidimensional structure of the SF-36 and underscore the feasibility of multinational comparisons of health status using this instrument. They also support the use of eight subscale scores in parallel with two second-order summary scores rather than one overall score.

AIMS: The rise in the proportion of older adults in the general population and the particularities presented by this age group show the need for the development of specific instruments to measure quality of life in older adults. The World Health Organization Quality of Life Group developed the WHOQOL-OLD module, which is comprised of 24 items divided into six facets: sensory abilities, autonomy, past present and future activities, social participation, death and dying and intimacy. This study aims at describing the development and application of the WHOQOL-OLD module Portuguese version. METHODS: The WHOQOL-Old module, the WHOQOL-BREF, BDI and BHS instruments were applied in a sample of 424 older adults in the city of Porto Alegre. RESULTS: The instrument showed suitable characteristics of internal consistency, discriminant validity, concurrent validity and test-retest reliability. Criterion validity presented findings that need further studies. CONCLUSIONS: The WHOQOL-Old module represents a useful alternative with good psychometrical performance in the investigation of quality of life in older adults.

AIMS: as health professionals, nurses need to be sensitive to moral issues related to their proessional responsibilities such as respect clients, advocate for their clients' rights in decision making, and management of moral dilemma in clinical practice. In order to improve nurses' moral sensitivity, a measurement tool is necessary for the assessment of moral sensitivity as well as development of moral education programs. The purpose of this pilot study was to develop and validate the korean version of k-msq. Fro the purpose, the original msq developed by lutzen(1997)was translated and validated using a sample of registered and student nurses. Methods: an expert discussion meeting and a pilot test(15 nurses & 15 student nurses)were utilized for the examination of the translated msq items. To validate the korean version of msq, psychometric properties were also examined: reliabilty was assessed by test-retest and cronbach's alpha. As an evidence of construct validity, correlation coefficients among msq 6 subscales were calculated. Results: there were no significant differences in test-retest reliabilty. Cronbach's alpha was estimated as .78. The correlation coefficients among the 6 subscales

A-82

were not high. The low correlation coefficients showed independent relationships among the 6 subconstructs. Conclusions: the korean version of msq can be used as an important tool to test moral sensitivity of korean nurses and student nurses.

nucleoside reverse transcriptase inhibitors [NRTI]). A multivariable analysis by a linear regression model was done to identify variables independently correlated to symptom score. RESULTS: 241 patients were enrolled. People on PI regimens had lower CD4 cell counts and had been on therapy for longer. Factors independently associated with the symptom score included non-adherence and PI-based regimens. The most frequently reported symptoms were anorexia (70%), vision troubles (53%), dyspnea (45%), pain (42%) and fever (40%). Overall, the symptoms differed significantly among groups: pain (higher in people taking nelfinavir compared to those taking lopinavir/ritonavir), fatigue, taste alterations, mental confusion, insomnia and cough (all higher for people taking PI-based versus PIsparing therapy). People taking lopinavir/ritonavir-containing regimens reported more fatigue (p<0.0001), dysphagia (p=0.06), confusion (p<0.0001), insomnia (p=0.05), and cough (p=0.01) compared to people taking NNRTI. CONCLUSIONS: PI-sparing regimens were better tolerated than PI-based ones according to a selfreported survey. Perceived-toxicity is important to take into account for the success of treatment (both virological and immunological) and for the quality of life, and should help determine the choice of antiretrovirals for individual patients.

42/Abstract 1426 The Translation and Linguistic Validation of the Imen's Sexual Health Questionnaire (MSHQ) Caroline Houchin, Translation and Linguistic Validation, Diane W. Wild, Patient Reported Outcomes, Oxford Outcomes, Oxford, United Kingdom AIMS: The objective of the study was to translate and linguistically validate the MSHQ for use in 27 languages. The questionnaire was developed in English for the USA by Raymond Rosen and SanofiSynthelabo in 2002/3 and was designed to gauge mens' perceptions of their own sexual functioning. METHODS: A standard methodology was employed: 2 forward translations and forward translation reconciliation or in-country review of existing translation, 2 back translations, back translation review, linguistic validation interviews with 5 men with erectile dysfunction in each country and/or review by clinicians in the country, and 2 proof readings. RESULTS: Translation issues were both linguistic and cultural. Some problems were highlighted after back translation while others, although regarding theoretically sound items, only came to light after pilot-testing. Issues included: - Confirmation being added to "_in the last month" to circumvent confusion between _"in the last 31 days" and_ "in the previous calendar month". - The use of "more than enough" to translate _"too long" in one version as _"too long" does not exist in Arabic. - "If it occurs" being added to items relating to sexual activity in the Arabic version, to avoid causing offence. Response options being altered in Italian for a closer grammatical fit with _proportion of time rather than _number of times. CONCLUSIONS: A rigorous process is imperative in translating Patient Reported Outcomes measures. In adhering to this we can successfully produce different language versions that are not only culturally and linguistically acceptable for use in their own countries but accurate to the original version and to each other. This enables results generated from the questionnaires to be compared to those obtained from other language versions worldwide. Furthermore, while translations may be linguistically accurate and stand up to theoretical tests of back translation and reviews, linguistic validation is an extremely valuable stage in the process as it may highlight previously apparently unproblematic wording.

44/Abstract 1371 Health-Ralated Quality of Life in HIV-Negative and Positive Homo- and Bisexual Men in Sweden Lars E. Eriksson, Neurobiology, Care Sciences and Society, Karolinska Institutet, Huddinge, Sweden, Torsten Berglund, Epidemiology, Swedish Institute for Infectious Disease Control, Solna, Sweden AIMS: To investigate HRQL in HIV-negative men who have sex with men (MSM) visiting a sexually tramsmissible disease (STD) clinic years 1996 and 2004 and to compare the results with the results of HIV-positive MSM and men from a general population sample. METHODS: Setting: A clinic at a Swedish urban hospital offering STD counselling, testing and treatment for MSM. Instrument: The Swedish Health-Related Quality of Life questionnaire (SWEDQUAL). Subjects: (a) a group of HIV-negative MSM investigated year 1996 (1996 cohort; n=164), (b) a group of HIV-negative MSM investigated year 2004 (2004 cohort; n=201), and (c) extracted data from a study of HIV-positive male patients who were investigated 1993-95, i.e., before the era of successful HIV treatment (HIVpositive group; n=48). The results are also compared with sex and age matched data from a published study of HRQL in a random sample of Swedish citizens (Swedish population group). Statistical methods: One-sample t-test (MSM 1996 cohort vs Swedish population group); Mann-Whitney U-test (MSM 1996 cohort vs HIVpositive group and MSM 2004 cohort respectively). RESULTS: The HIV-negative MSM had statistically significant worse emotional well-being, satisfaction with family life, relation to their partner and general health perception, and better physical HRQL compared with the general Swedish male population. There were no differences in the emotional and family dimensions between HIV-negative and HIV-positive MSM, but the HIV-negative group had less physical role limitations, better sexual functioning and better general health compared to the HIV-positive group. No apparent differences were seen in the comparison of the HIV-negative MSM 1996 and 2004 cohorts. CONCLUSIONS: The results show that psychosocial HRQL is lower for MSM than for a general male population sample. Since a relationship between mood and unsafe sexual practice in MSM has been suggested, it could, also from an STD prevention perspective, be of significance to give psychosocial aspects increased attention when dealing with patients from this group.

43/Abstract 1462 Tayloring Haart: A Pl-Sparing Antiretroviral Therapy Versus A Protease Inhibitor-Based Regimen Rita Murri, Infectious Diseases, Catholic University of Rome, Rome, Italy, Gabriella Orlando, Marcella Vandelli, Nicola Squillace, Giovanni Guaraldi, Infectious Diseases, University of Modena, Modena, Italy AIMS: To compare self-reported symptoms of HIV-infected patients taking a PI-based regimen to those of people taking a protease inhibitors (PI)-sparing combination. METHODS: Design: Crosssectional self-reported-survey. Between January 2004 and June 2005 consecutive patients attending a clinic for lipodystrophy were enrolled. Measures included a self-reported Symptom Index and selfreported adherence. People were grouped according to the regimen they were taking: PI-based and PI-sparing regimens (including nonnucleoside reverse transcriptase inhibitors [NNRTI] or three

A-83

Further, the standardize residuals are subjected to principle component analysis to examine the dimensionality of the remaining items. RESULTS: Item 3, 4, and 21 are eliminated because of poor fitting. These three items are two inverse items (_Pain and _Medication ) and the item for _Sexual Life . The infit and outfit MNSQ of the remaining 23 items are acceptable, ranging form 0.68 to 1.40. Principle component analysis on the standardize residuals reveals no dominant component. The first and the second components account for only 5.5% and 4.5% of the residual variances, respectively. Therefore, after eliminating the three misfit items, the WHOQOL-BREF can be regarded to be unidimensional. CONCLUSIONS: The present study shows that after deleting three misfit items, the WHOQOL-BREF fits a unidimensional Rasch model, suggesting that a single-factor structure is proper for the WHOQOL-BERF. In a summary, although there are four domains in the WHOQOL-BREF theoretically, however, statistical evidence from Rasch analysis reveals that there is an overall QOL factor in the WHOQOL-BREF.

45/Abstract 1823 Health-Related Quality of Life of Adult People with HIV/AIDS in Spain Lilisbeth Perestelo-Pérez, Pedro Serrano-Aguilar, Julio LópezBastida, Service of Evaluation and Planning, Canary Health Care Service, Santa Cruz de Tenerife, Spain AIMS: The use of highly active antirretroviral therapy (HAART) has reduced morbidity and mortality by HIV/AIDS. The reduction in HRQOL according to disease progression is also relevant for allocation resources and clinicians. The aim of the study was to determine the impact on HRQOL for people with HIV/AIDS in Spain. METHODS: Multicentre cross-sectional study with 572 people diagnosed with HIV/AIDS and receiving outpatient care. People were consecutively recruited according to disease severity from four hospitals in Canary Islands, Spain. In order to establish the severity of HIV infection, we used the 1993 classification scheme proposed by the CDC in the US, which divides the disease into HIVasymptomatic stage, HIV-symptomatic stage and AIDS. The information sources were medical records and self-completed questionnaires. HRQOL information was gathered by the SF-36. RESULTS: All SF-36 domain scores (except bodily pain) discriminated between patients asymptomatic, symptomatic and AIDS. Patients symptomatic had significantly lower scores than patients asymptomatic in the role physical and general health. Patients with AIDS had lower mean scores in all HRQOL domains than patients asymptomatic and had lower mean scores in the social function, mental health and mental component summary than patients symptomatic. Univariate analyses indicated that the CD4 + lymphocytes counts were significantly correlated with the physical functioning in the SF-36 and weak and negatively correlated was established between the viral load, general health and physical functioning. Patients who received HAART showed lower HRQOL scores than patients who no received HAART, although significant differences were found for physical function. CONCLUSIONS: QOL ratings on both physical functioning and psychological wellbeing aspects of the SF-36 were lower in people with AIDS and significant differences were obtained among patients according to disease severity of HIV disease. The dimensions of the SF-36 were able to discriminate between patients with HIV/AIDS, especially among patients asymptomatic and with AIDS.

47/Abstract 1667 Cultural Specificities of the WHOQOL-100 in Portugal: The Focus Groups and the Development of a Portuguese Facet Political Power Daniel B. Rijo, Faculdade de Psicologia e Ciências da Educação, Universidade de Coimbra, Portugal, Maria C. Canavarro, Faculdade de Psicologia e Ciências da Educação, Adriano S. VazSerra, Faculdade de Medicina, Marco D. Pereira, Marco D. Pereira, Faculdade de Psicologia e Ciências da Educação, Mário R. Simões, Faculdade de Psicologia e de Ciências da Educação, Manuel J. Quartilho, Faculdade de Medicina, Universidade de Coimbra, Carlos F. Carona, Núcleo Regional do Centro, Associação Portuguesa de Paralisia Cerebral, Carla S. Gameiro, Tiago F. Paredes, Faculdade de Psicologia e de Ciências da Educação, Universidade de Coimbra AIMS: Aims: This paper presents the qualitative pilot study conducted as part of the Portuguese validation of the WHOQOL-100. The goal was to study if there were any specific Portuguese facets besides the ones proposed by the WHOQOL construct. . METHODS: Methods: This study used the focus group approach; four different focus groups were organized: (1) people from the general population, including non professional caretakers, (2) people in contact with health facilities, (3) health professionals, and (4) students of a Masters Degree in psychological assessment. The groups discussed their own concepts of Quality of Life (QoL) as well as the construct proposed by the WHO. RESULTS: Results: A new issue emerged from the discussions: the influence of politics in people's QoL. From the analysis of the tape recordings of the focus groups discussions, four different aspects of the political power appeared: participation in political decisions, satisfaction with policies, impact of political decisions in everyday life and trust in politics. Eight different questions were developed in order to evaluate all of these. Using internal consistency degradation (Cronbach ±) of the facet as a criteria (n=604), the final four questions included in the new facet were obtained: To what point do you trust in politics?; To what point are you satisfied with decisions made by politics?; To what measure decisions made by politics respond to your everyday needs?; To what point do you think you can trust in decisions made by politics?. Quantitative results show that the Portuguese Facet Political Power has good psychometric properties: item-facet correlations ranging from .67 to .85, the Cronbach ± for the facet is .78, and test-retest coefficient is .83. The facet score discriminates between normal (n=315) and patient (n=289) samples (p=.003). CONCLUSIONS: Conclusions: The way people perceive politics

46/Abstract 1609 Is the WHOQOL-BREF Unidimensional or Multidimensional? Grace Yao, Ping-Chi Lee, Chia-Huei Wu, Psychology, National Taiwan University, Taipei, Taiwan AIMS: Although the WHOQOL-BREF was claimed to have four domains, recent studies seemed to suggest that the WHOQOL-BREF is unidimensional. Researches with diverse methods, such as correlation analyses, confirmatory factor analyses, multitraitmultimethod (MTMM) analyses, and content validity studies showed that the four domains are highly correlated and one general factor might be sufficient for the WHOQOL-BREF. With the preceding researches, the aim of present research is to re-examine the unidimensionality of WHOQOL-BREF by item response theory (IRT). METHODS: Thirteen thousand and eight participants from the 2001 National Health Interview Survey (NHIS) database are included in this study. Rasch analysis (Rating Scale Model) is applied using WINSTEPS software. Infit and outfit MNSQ are used to determine how well each item contributes to a common quality of life (QOL) factor. Item with the largest MNSQ is deleted, and the estimation is repeated until all MNSQs are not larger then 1.40.

A-84

and political decisions, trust and satisfaction they report concerning political decisions, as well as the perception that political decisions respond to people needs, revealed to be an important feature of the QoL in Portugal.

Sweden and compared with patient-data from KIMS (Pfizer International Metabolic Database) for 5-6 years follow-up. Both population samples and patient-data were classified into >60 and <60 years of age. The deficit in QoL-AGHDA scores in patients were defined by the decrement from general population values for respective age-group. Estimates were adjusted for gender. Confidence interval level was set to 95%. RESULTS: On average, before treatment, similar deficits in QoL were observed in both agegroups within each country. Older patients reported better QoL. The most profound deficit was found in patients from E&W (mean = -7.7 in older patients and -9.5 in younger patients), followed by the Dutch patients (-3.5 and -5.7, respectively) and the Swedish (-3.5 and -4.6, respectively). The largest improvement occurred during the first year of GH therapy in both age-groups and in all countries. The QoLAGHDA mean scores were close to "normal values" in the elderly patients from Netherlands and Sweden after the first year, and E &W after 5 years of GH replacement. Younger patients approached population levels after longer follow-up (3-5 years) but only in the Netherlands and Sweden. CONCLUSIONS: Regardless of age, patients with GHD benefited from GH replacement measured by QoL-AGHDA, and showed similar pattern of response however elderly patients reached population levels earlier. The more profound QoL impairment demonstrated by patients from E&W is most likely related to criteria required by the Health Service for consideration of adult GH replacement.

48/Abstract 1594 The Mediating Effect of Physical Activity on Chronic Disease and Heath Outcomes in Older Adults Richard G. Sawatzky, Nursing, Teresa Liu-Ambrose, Rehabilitation Sciences, Carlo A. Marra, Pharmaceutical Sciences, William C. Miller, Rehabilitation Sciences, University of British Columbia, Vancouver, British Columbia, Canada AIMS: The objective of our study was to test whether leisure-time physical activity mediates the relationship between having a chronic disease and self-reported health outcomes, such as physical mobility and pain, in older adults. METHODS: The data were taken from the Canadian Community Health Survey (cycle 1.1) which was completed by 130,000 respondents who were randomly selected from 136 health-regions across Canada. Only respondents who were 65 years and older were included in our study. We used the Health Utilities Index Mark 3 (HUI3) to measure eight health attributes that were specified as independent ordinal variables in our models. Leisure-time physical activity was assessed by calculating weekly energy expenditure (kcal per week) based on the metabolic equivalents of self-reported leisure activities. Structural equation modeling techniques were used to test the hypothesized relationships between leisure-time physical activity, chronic disease and each of the HUI3 attributes while controlling for demographic variables (age and sex), health-behaviours (tobacco use and alcohol consumption) and obesity. RESULTS: Having a chronic disease is predominantly associated with an increase in the HUI3 attributes of limitations associated with mobility, dexterity and pain. Leisure-time physical activity that resulted in a weekly energy expenditure of at least 1000 Kcal is associated with a decrease in mobility limitation, pain and emotional problems (i.e., increased happiness) when controlling for chronic disease. The results confirm that physical activity primarily mediates the relationships between various chronic diseases and mobility. The indirect relationships remain statistically significant when controlling for demographic variables, health-behaviours and obesity. CONCLUSIONS: Our study shows that chronic disease and physical activity are significantly associated with several selfreported health outcomes in older adults. The findings provide support for health promotion programs to reduce the impact of chronic disease on physical activity for older adults.

50/Abstract 1610 The Determinants of Quality of Life in Nine Chronic Diseases Lin W. Ling, Lin Y. Siou, Yao K. Ping, Psychology, National Taiwan University, Taipei, Taiwan AIMS: Many literatures indicate that chronic diseases have negative impact on quality of life. People suffering chronic diseases not only have higher health service utilization rates but also have their life style changed to adapt effects of those chronic diseases. The aim of this study is to determine the predictive variables of overall quality of life for patients with nine chronic diseases and health people in Taiwan. METHODS: This study used 13,008 participants who aged 20 to 65 years old from 2001 Taiwan National Health Survey Information System (NHISIS) database. Nine chronic diseases, including cardiac disease, pulmonary disease, hypertension, diabetes mellitus, hyperlipidemia, peptic ulcer, sinusitis, renal disease and liver disease, were selected. There are more than 200 people in each disease. Hierarchical regression analysis was conducted to find significant determinants of overall QOL measured by Visual Analogue Scale (VAS) from the demographics, the objective health status and the subjective quality of life domains (WHOQOL-BREF). RESULTS: Hierarchical regression analysis showed that after controlling the effect of demographics, the objective health status had an incremental effect in predicting overall QOL. Further, after controlling the effect of demographic and the objective health status, the four QOL domains had an incremental effect in predicting overall QOL. This finding was observed in all nine diseases. CONCLUSIONS: In general, demographics, the objective health status and the subjective QOL domains had significant effect on overall QOL, in which the effect of the subjective QOL domains was greater than the effect of the objective health status and demographics. In addition, this result was reliable among nine diseases.

49/Abstract 1340 The Impact of Age on the Beneficial Effects of Long-Term Growth Hormone (GH) Replacement in Patients with GH Deficiency (GHD) - In Relation to General Population Values Anders F. Mattsson, Maria Koltowska-Haggstrom, KIGS/KIMS/ACROSTUDY Medical Outcomes, Pfizer Endocrine Care, Pfizer Health AB, Sollentuna, Sweden, Stephen M. Shalet, Department of Endocrinology, Christie Hospital, Manchester, United Kingdom AIMS: To compare impairment in QoL in patients with GHD who were >60 with those <60 years of age before GH replacement and their response to long-term treatment in relation to general population values. METHODS: QoL was measured by QoL-AGHDA (Quality of Life Assessment for Growth Hormone Deficiency in Adults) in general population in England & Wales (E&W), the Netherlands and

A-85

51/Abstract 1629

53/Abstract 1489

Quantitative Review of Research into Diabetes Intervention Programmes Ana Andrés, Juana Gómez, Methodology for the Behavioural Sciences, University of Barcelona, Barcelona, Spain, Spain

Self- and Proxy-Rating of Quality of Life in Patients with Dementia and Mild Cognitive Impairment Using the EurohisQoL Index Monika Bullinger, Medical Psychology, Sönke Arlt, Juliane Hornung, Holger Jahn, Psychiatric University Hospital, University of Hamburg, Hamburg, Germany

AIMS: The aim of this study is to carry out a meta-analytic review of studies that, using the Transtheoretical Model of behaviour change, attempt to improve the quality of life in diabetic patients. METHODS: A bibliographic search was conducted in the databases of the ISI Web of Knowledge and Medline. The keywords used were transtheoretical model AND diabet*. A total of 33 documents were gathered, but only 18 were selected. The study sample was restricted to those studies where an intervention programme, based on the Transtheoretical Model, had been applied. Moreover, only those interventions which aimed to increase the quality of life in diabetic people were selected. Most of these were related to the promotion of physical exercise or healthy eating habits among these patients. For each one of the studies, methodological and substantive variables were coded, analysing the effect sizes from the same metrics. RESULTS: The information obtained from this meta-analysis is useful to determine the extent to which the interventions based on the Transtheoretical Model are effective in diabetic samples. The effect sizes provide us with a quantitative measure of whether the therapy has achieved its aim or not. However, it should be emphasised that in some studies the samples used were too small and this could be a problem in terms of generalizing the results. CONCLUSIONS: This meta-analysis provides an overall view of the efficacy of interventions which seek to improve quality of life in diabetic patients, whether through diet or physical exercise.

AIMS: Measuring and improving quality of life (QoL) of dementia patients is a major concern in clinical research and practice. METHODS: An eight-item version of the WHOQOL-BREF, the EUROHIS-QoL index, was completed by 66 outpatients with mild cognitive impairment (MCI) and mild to moderate dementia and their primary caregivers (proxy) in a cross-sectional study. Cronbach´s alpha (Ca) was determined as a measure of internal consistency of the EUROHIS-QoL index and correlations where calculated between self- and proxy-ratings. From Mini-Mental State Examination scores (median 22, range 13 to 29) tertile-groups with different severity grades of cognitive impairment were derived (1st: <=21, 2nd: >21 and <25 and 3rd: >=25) for comparing self- and proxy-rating of EUROHIS-QoL index. RESULTS: Internal consistence of EUROHIS-QoL index was good in the total study population (patients: Ca=0,81, proxy: Ca=0,83) and for all tertile-groups (1st: Ca=0,82 , 2nd: Ca=0,81, 3rd: Ca=0,75). There was a positive correlation of EUROHIS-QoL index scores between self- and proxyrating in the whole study population (Spearman R= 0,51, p<0,001); when tertile-groups were analyzed separately, correlations were significant for the 1st and 2nd only. There was no significant difference of EUROHIS-QoL index scores between the patientgroups while patient-rating was significantly higher than proxy-rating for all tertiles. CONCLUSIONS: The EUROHIS-QoL index was applicable to all individuals of the study population, internal consistency was good even in the most cognitively impaired group. Considerable differences between self-assessment and proxy-rating point at the importance of using a self-rating instrument for assessing QoL in patients with MCI/ mild to moderate dementia.

52/Abstract 1168 The Influence of Neuroleptic Therapy on Quality of Life in Schizophrenic Patients Suzana M. Tosic Golubovic, Srbobran A. Miljkovic, Dragoslava M. Gugleta, Nada M. Djordjevic, psychiatry, University of Nis, Gornja Toponica, Serbia, Serbia & Montenegro

54/Abstract 1760

AIMS: Reduction of the florid psychotic symptoms and reaching of stable remission are certainly the results of neroleptic therapy of schizophrenia. Never the less important treatment goal is assessment of quality of life as a subjective perception of life satisfaction. METHODS: The symptoms of sch were evaluated using PANSS, Calgary depression scale, while quality of life was evaluated using quality of life scale in schizophrenia (QLS) and WHOQOL Brief version RESULTS: The aim of our study was evaluation and comparing quality of life among schizophrenic patients without prominent psychotic symptoms who were on several months neuroleptic treatment, one group with typical (N=20) and other with atypical neuroleptics risperidon and clozapine (N=20). CONCLUSIONS: Better quality of life was assess among younger patients with longer and stable remission periods, with better compliance, treated with atypical neuroleptics and also with better interpersonal, social and occupational functioning. Adequate treatment of chronic and progressing illness like sch except prevention of relapse and resocialization of the patients, have to involve a battle for better quality of life percepted by the schizophrenic patients.

Psychometric Evaluation of a Most Troubling Sympton Scale for Generalized Anxiety Disoder Clinical Trails Georges Gharabawi, Gahan Pandina, Medical Affairs, Janssen Pharmaceutica, Inc., Titusville, NJ, Dennis Revicki, Leah Kleinman, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Ibrahim Turkoz, Quantitative Methodology, Luella Engelhart, Outcomes Research, Ortho-McNeil Janssen Scientific Affairs, L.L.C., Titusville, NJ AIMS: Generalized anxiety disorder (GAD) is often unrecognized and fewer than a third of patients are correctly diagnosed. The Patient-Rated Troubling Symptoms Scale for Anxiety (PaRTS-A) provides patient assessment of GAD symptoms. We examined the psychometric properties and clinical responsiveness (MCID) of PaRTS-A in GAD patients. METHODS: GAD patients (n=385/390 ITT) in a 6-week randomized, placebo-controlled trial of adjunctive risperidone contributed data at baseline, weeks 4 and 6 via interactive voice response on PaRTS-A; Quality of Life Enjoyment and Satisfaction Questionnaire Short Form (Q-LES-Q); Sheehan Disability Scale (SDS), and chief complaint item; and Patient Global Impression of Severity (PGIS). Clinicians completed HAM-A (baseline, week 2, 4, 6) and Clinical Global Impressions-Severity (CGI-S) scale (baseline, week 4, 6). Descriptive exploratory factor item response (IRT) and ANCOVA analyses were conducted using SAS version 8.2 (SAS Institute, Cary NC) and MULTILOG (IRT).

A-86

RESULTS: Patients were predominantly employed white women with a mean 14.9±12.6-year history of anxiety and a first anxiety episode at age 29.4 ±12.9. Fewer than 15% were ever hospitalized for anxiety. Strong evidence for a single factor (eigenvalue=3.54; 51% variance explained), strong correlation of items (r=0.48-0.85), ability of items to assess anxiety severity (IRT), and high reliability (>0.7) were found. Validity correlations were moderate and supportive (QLES-Q: -0.59; SDS: 0.62; chief complaint: 0.65) at baseline. PaRTSA discriminated between high and low HAM-A (p<.0001) and CGI-S (p=0.027) ratings. MCID was 4 points (stable and slight improvement). CONCLUSIONS: Using PaRTS-A is feasible and results in reliable and internally consistent scores, good construct validity, discriminant validity, and responsiveness over time in a GAD sample. Importantly, PaRTS-A was responsive to changes in clinical status. Further research would confirm whether such changes would be meaningful in clinical practice. Supported by Ortho-McNeil Janssen Scientific Affairs, L.L.C.

56/Abstract 1121 Longitudinal Study on QOL Assessment in Orthodontic Patients Required Jaw Surgery Shiori Azuma, Oral Dysfunction Science, Tohoku University Graduate School of Dentistry, Sendai, Japan, Masahiro Kohzuki, Internal Medicine and Rehabilitation Science, Tohoku University Graduate School of Medicine, Sendai, Miyagi, Japan, Mayumi Tajima, Orhodontics and Dentofacial Orthopedics, Shuichi Saeki, Oral Dysfunction Science, Tohoku University Graduate School of Dentistry, Sendai, Miyagi, Japan, Yoko Goto, Occupational Therapy School of Health Science, Sapporo Medical University, Sendai, Miyagi, Japan, Kaoru Igarasi, Oral Dysfunction Science, Jnnji Sugawara, Orthodontics and Dentofacial Orthopedics, Tohoku University Graduate School of Dentistry, Sendai, Miyagi, Japan AIMS: The improvement of QOL in orthodontic patients has been a principal objective in modern clinical orthodontics. We found that the patients with malocclusion, especially in need of surgical correction, had lower disease specific QOL and higher anxiety. The aim of the present study was to clarify the changes of QOL and psychological status following jaw surgery in the patients with facial deformities. METHODS: Thirty-one adult orthodontic patients (Female:24, Male: 7) diagnosed as malocclusion for indication of jaw surgery were selected. They were divided into three groups according to severity of malocclusion determined by severity score (SS); Low-SS, Moderate-SS and High-SS. QOL before surgery (T1) and after surgery (after debonding of multibracket appliances) (T2) was compared among the three groups. Generic QOL was assessed using the SF-36. Disease-specific QOL was evaluated with the Subjective Oral Health Status Indicators (SOHSI), the Orthognathic QOL Questionnaire (OQLQ), and recognition and satisfaction scores of surgical correction. In addition, anxiety and depression were assessed using STAI and SRQ-D, respectively. RESULTS: 1) There was no significant difference in any scale scores of SF-36 between T1 and T2 in each groups. 2)Both disease-special QOL and psychological status improved significantly after the treatment in any of the groups. Moreover, their anxiety improved after the treatment irrespective of individual anxiety level before surgery. CONCLUSIONS: The results of this study suggested that QOL of orthodontic patients improved after the treatment irrespective of SS before surgery.

55/Abstract 1764 Quality of Life and Depression in Primary Care Unit Ana Flávia B. da Silva Lima, Psiquiatria e Medicina Legal - UFRGS, Hospital de Clínicas de Porto Alegre, Porto Alegre, Rio Grande do Sul, Brazil, Marcelo P. de Almeida Fleck, Psiquiatria e Medicina Legal -UFRGS, Hospital de Clínicas de Porto Alegre, Porto Alegre, Rio Grande do Sul, Brazil AIMS: Objective: To evaluate changes in quality of life of people with major depression who seek treatment in a primary care services in a Brazilian sample, over one year. METHODS: Material and Methods: This is a one year naturalistic cohort study. Patients with a current major depressive episode, between 18 and 75 years, without currently under treatment for depression during the past three months were included. Participants completed a structuralized interview Composite International Diagnostic Interview Instrument (CIDI )and filled questionnaires at baseline visit, 6 weeks, 3 and 9 months after the beginning of the study . The measures of quality of life were WHOQOL _ BREF - World Health Organization Instrument to Assess Quality of Life, QLDS _ Quality of Life _ depression, MHI-5 - and SF-12 - both subsets of the 36 items of the Medical Outcomes study-short forms - the SF-36. The CES- D - Centers for Epidemiologic Studies _ Depression instrument was used to measure symptoms of depression. RESULTS: Results: The sample was composed by 179 primary care users with major depression. For the statistical analysis the variance (ANOVA) and intention to treatment (ITT) was used. At baseline the means scores of WHOQOL varied from 50.3(sd=0.9) to environmental domain until 55.6(sd=16.9) to physical domain. At 9 months the patients reported significant higher scores in all dimensions, but the difference of means was not bigger than 7 points. For the MHI-5 and SF-12 the scores not exceed the 9 points between baseline and 9 months. In the QLDS the mean was 11.2 (sd=7.6) at baseline and 8.6 (sd=8.2) at 9 months. In the CES- D the mean score at baseline was 31.7 (sd=10.1) and 24.5 (sd=13.8) one year later, showed that people continued with depressive symptoms at the end of study. These means differences were significant statistical CONCLUSIONS: Conclusions: The results suggest that patients with major depression remained with depressive symptoms with little improvements in quality of life along one year in a naturalistic follow-up.

57/Abstract 1445 Effectiveness of Traditional Chinese Medicine in Primary Care Wendy Wong, Cindy, LK Lam, Medicine, The University of Hong Kong, Hong Kong (SAR), Hong Kong, Hong Kong AIMS: To evaluate whether Traditional Chinese Medicine (TCM) consultation can significantly improve the quality of life of patients in primary care and to find out whether the Chinese Quality of Life instrument (ChQOL) is more responsive than Short form-36 (SF-36) in detecting the treatment effect of TCM. METHODS: This is a prospective, longitudinal study on 361 patients who consulted the TCM outpatient clinic for the first time. Methods: New patients at the clinic were recruited. Each patient was asked to complete a structured questionnaire on ChQOL and the SF-36 before the consultation with the TCM practitioner. Each subject was followed up after two weeks when he/she were re-assessed by the TCM practitioner and completed the ChQOL, SF-36 again and a global rating on change of the condition. The primary outcome measure was the change in scores in ChQOL and SF-36 scores after the TCM consultation. The ChQOL and SF-36 scores were correlated with the TCM practitioner assessment and the subjects' global rating of change. RESULTS: All domains scores of ChQOL and SF-36 of patients were improved for

A-87

all patients investigated in internal medicine, bone-setting and acupuncture departments of outpatient clinics (p<0.05). In subgroup analysis, the improvement of SF-36 and ChQOL is statistically significant in groups with TCM differential diagnosis of exterior, heart, brain, kidney, meridian and limbs syndromes. The patients global rating is statistically correlated to the Physical component score (PCS) and Body Pain domains of SF-36 and physical form of ChQOL (p<0.01). While only role emotional domain of SF-36 statistically correlated (p<0.05) to the TCM practitioner global assessment after 2 week treatment. CONCLUSIONS: The results showed TCM consultation and treatement have significant improvement in health-related quality of life. Moreover, patients rating do consistent with ChQOL and SF-36 but not for TCM practitioner assessment. Further research is needed to examine the HRQOL in accurately reflect changes in TCM treatment with longer followup period.

59/Abstract 1580 Incidence and Prevalence of the Pressure Sores in Portugal Pedro L. Ferreira, Centre for Health Studies and Research, University of Coimbra, Coimbra, Portugal, Cristina Miguéns, Health Centre of Pampilhosa da Serra, João C. Gouveia, Centre for Health Studies and Research, Ministry of Health, Pampilhosa da Serra, Portugal, Katia Furtado, Portuguese Institue of Oncology, Lisboa, Portugal AIMS: The number of patients with pressure ulcers represents a serious problem. In each care setting, it is extremely important to identify those patients who are at risk and to define a care plan according to the specific factors that increase the vulnerability to develop pressure ulcers. The aim of this study was to obtain, for the first time in Portugal, figures for the incidence and the prevalence of pressures sores. METHODS: This study began in July 2003 and the data collection started in January 2004 and finished in September 2005. Eight hospitals were chosen to participate, according to their geographic distribution. Previously to the data collection, we culturally adapt the Braden Scale for predicting pressure sore risk. Data was split by six waves by hospital. For each department, all inpatients were included in the study. RESULTS: Data from a total of 78,306 patients were collected from the eight hospitals. 5.66% of these patients had ulcers, with an average of 1.98 ulcers per patients. The incidence of ulcers was higher in departments of medicine and lower in the intensive care units. For the prevalence study, we collected data from a total of 10,202 patients. Almost half of these patients (49%) showed no risk, 25.8% a small risk, 10.1% a moderate risk, 11.5% a high risk and 3.6% a severe risk. The more severe ulcer was located in the sacrum. The general prevalence was 11.5% and medicines showed the highest prevalence (17.2%), followed by the intensive care units (13.0%) and the emergence care units (12.1%). Among the departments with lower prevalence values we include the medical specialities (7.8%), the surgical specialities (7.9%) and the surgeries (6.9%). CONCLUSIONS: Pressure sores yield pain and suffering, increase the mortality and the length of stay, are costly and, above all, can be avoided. This project produced the first indicators of incidence and prevalence of pressure sores for the whole Portuguese population. We were able to obtain meaningful relationships between sensory perception, moisture, activity, mobility, nutrition, friction and shear, and incontinence.

58/Abstract 1382 Household Income and Health-Related Quality of Life in Japan: Nationwide Survey in 1995 and 2002 Shin Yamazaki, Environmental Epidemiology, National Institute for Environmental Studies, Tsukuba, Japan, Yoshimi Suzukamo, Physical Medicine and Rehabilitation, Tohoku University, Sendai, Japan, Joseph Green, International Academic Affairs, University of Tokyo, Tokyo, Japan, Shunichi Fukuhara, Epidemiology and HealthCare Research, Kyoto University, Kyoto, Japan AIMS: Evaluating the association between socioeconomic status (SES) and health could provide information that is important for planning integrated economic and public health policies, because SES strongly reflects individual economic status. We examined the association between annual household income as a measure of SES and the eight scale scores of the SF-36 Health Survey (SF-36) as a quantifier of health-related quality of life (HRQOL) in Japan in 1995 and 2002. METHODS: Cross-sectional survey. Both data in 1995 and 2002 were from the SF-36 national survey in Japan. In 1995, a total of 4500 people 16-years or older were selected from the entire population of Japan using stratified-random sampling and 3395 responded. We restricted the subjects to 20-years or older. In 2002, a total of 4500 people 20-years or older were selected from the entire population of Japan using stratified-random sampling and 2966 responded. For each datasets, we used analysis of covariance to estimate the mean differences among five 'annual household income' groups for each SF-36 domain. Men and women were analyzed separately. The adjustment factors in this model included age, district of residence, population of city of residence, educational background, number of illnesses, family-living status, marital status, and occupational status. RESULTS: In 1995, men with lower levels of annual household income had lower scores in all SF-36 domains. However, only "general health perceptions" and "social functioning" showed statistically significant trends among the women surveyed. In the subgroup of women working full time, there were no domains that showed significant trends. In 2002, during those 7 years of economic decline in Japan, a new association developed between women's household income and their mental health. CONCLUSIONS: An association exists between annual household income and SF-36 scores in Japanese employment. The employment and economic policies that affect annual household income potentially influence HRQOL.

60/Abstract 1550 Impact of Sleep Problems on Health-Related Quality of Life in Transplant Patients Cynthia R. Gross, Maryanne Reilly-Spong, Valerie G. Russas, Carol L. Skay, Mary Jo Kreitzer, Hassan N. Ibrahim, Academic Health Center, University of Minnesota, Minneapolis, Minnesota, Michel Cramer-Bornemann, Minnesota Regional Sleep Disorders Center, Hennepin County Medical Center, Minneapolis, Minnesota AIMS: The prevalence of sleep problems is rising, with current costs of insomnia in the US, in terms of health and lost productivity, estimated at over $11 billion (http://www.cdc.gov/nccdphp/overview.htm). Information is limited for the impact of sleep problems on healthrelated quality of life (HRQL). Our aims are to describe sleep quality in organ transplant recipients and determine the impact of selfreported sleep quality and quantity on HRQL. METHODS: Recipients (n=87; M=47, F=40; age 21-75, mean 53.9 yrs; kidney/pancreas=47, liver=19, heart/lung=21; post-transplant 6 mos22 yrs) completed the Pittsburgh Sleep Quality Index (PSQI), SF-12 (MCS-12 & PCS-12) and SF-36 subscales for pain and vitality at baseline for a clinical trial. PSQI subscales range from 0-3 (lower

A-88

(p<0.0001, HR 1.194) as predictive. The PI was also highly significant (p<0.0001; HR = 2.367). The final model retained the PI and pain as independently predictive of survival. Hence, the only patient reported factor that is significant prognostic in the final model is self-reported pain. Its HR translates into a 15% increase in patient_s hazard of death for every 10-point increase (i.e. worsening) on the scale measuring pain. CONCLUSIONS: Our results suggest that pain may be an important prognostic factor in addition to the PI. Further research is needed to confirm these results.

scores for better sleep), and total scores range from 0-21 with >5 considered poor sleep. The MCS, PCS and SF-36 scores range from 0-100, with higher scores for better health. RESULTS: While female gender is a risk factor for insomnia, sleep problems were common in men and women. Overall, 61% were poor sleepers by PSQI >5, yet only 12.6% had been previously diagnosed with a sleep disorder (9 M, 2 F) and only 12.6% were taking sleep medication (6 M, 5 F). PSQI was strongly and negatively correlated with MCS, PCS, pain and vitality (r2s -.35 to -.58, p<.01, all). In multiple regression, MCS was explained by predictors: PSQI score, age and heart/lung tx with adjustments for time from tx and menopausal (r2=34%). PCS was explained by predictors: PSQI and age, adjusted for heart/lung tx, time from transplant and menopausal status (r2=21%). PSQI was the most influential predictor of HRQL, followed by age which is positively associated with mental health, but negatively associated with physical health. CONCLUSIONS: Confirmation in longitudinal studies is needed to distinguish risk factors for poor sleep from its consequences, however, our findings clearly indicate that poor sleep is a frequent and under-treated problem after organ transplant and is associated with a pervasive, negative impact on HRQL.

62/Abstract 1359 Quality of Life (QoL) for Head and Neck Cancer Patients Enrolled in RTOG Trail 99-01 Darlene J. Johnson, Senior Adult Oncology, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, R. Suzanne Swann, RTOG Statistics, American College of Radiology, Philadelphia, PA, Francis LeVeque, Harper Hospital/Wayne State University, Detroit, MI, Charles W. Scarantino, Radiation Oncology, Rex Healthcare Cancer Center, Raleigh, NC, Janice K. Ryu, Radiation Oncology, University of California Davis Cancer Center, Sacramento, CA, Andre Fortin, Radiation Oncology, Hotel-Dieu de Quebec du CHUQ, Quebec City, Quebec, Canada, JonDavid Pollock, Radiation Oncology, Schiffler Cancer Center Wheeling Hospital, Wheeling, WV, Harold E. Kim, Radiation Oncology, Wayne State University, Detroit, MI, K. Kian Ang, Radiation Oncology, MD Anderson Cancer Center, Houston, TX

61/Abstract 1423 Predicting Survival Using Patient Reported Health Status Information From Malignant Pleural Mesothelioma (MPM) Patients. Andrew Bottomley, Fabio Efficace, Quality of Life Unit, EORTC Data Center, Brussels,, Belgium, Rabab Gaafar, National Cancer Center, Department of Oncology, Cairo, Egypt, Corneel Coens, Quality of Life Unit, EORTC Data Center, Brussels, Belgium, Christian Manegold, Department of Oncology, University Medical Center, Mannheim, Germany, Sjaak Burgers, Department of Oncology, Netherlands Cancer Institute, Amsterdam, he Netherlands, Mark Vincent, Department of Oncology, London Regional Cancer Center, Ontario, Giuseppe Giaccone, Department of Oncology, Free University Medical Center, Amsterdam, The Netherlands, Catherine Legrand, Quality of Life Unit, Alexandre Passioukov, Lung Cancer Unit, EORTC Data Center, Brussels, ., Belgium, Rob van Klaveren, Department of Pulmonology, Erasmus Medical Center, Rotterdam, The Netherlands, Jan van Meerbeeck, Department of Oncology, University Hospital, Ghent, Ghent, Belgium

AIMS: Radiation therapy (RT) is a primary modality for treating H&N cancer. Mucositis, a complication of treatment, can be painful. Preventing or reducing severity of mucositis may lead to better local control if treatment breaks are avoided. It is hypothesized that relief of symptoms benefits function, wellbeing, and QOL. Secondary objectives of RTOG 9901 included determining impact of mucosal injury and pain on QOL, and impact of symptom relief with GM-CSF measured by QOL. METHODS: 130 patients were randomized to receive RT and GM-CSF, 250mg/m2 sc (n=65) or RT and placebo (n=65). Patients were stratified whether they were or were not receiving concurrent cisplatin. All patients were scheduled to receive 60-70 Gy at 1.8-2.0 Gy/day 5 days per week for 6-7 weeks. GM-CSF or placebo was administered by SubQ injection one week prior to RT on Mon-Wed-Fri during treatment, stopping two weeks after treatment. QOL and function were evaluated using the University of Washington QOL/H&N Symptom Questionnaire at baseline, 4, 13, 26 and 48 weeks from the start of RT. Scores were categorized as Normal, Abnormal, Severe, and Lost (no longer able to perform that function). RESULTS: Comparing QOL at baseline to 4 and 12 weeks, a decline existed in all areas of QOL for patients receiving placebo. Significant areas included Eating, Saliva, Mucous or Phlegm and Taste. At 4 weeks there was Pain (throat), and decline of Recreation and Eating. At 12 weeks decreased activity persisted. In patients who received GM-CSF there was a decline in all areas except disfigurement. Declines in Activity, Eating, Saliva, Mucous or Phlegm and Taste at both 4 and 12 weeks were noted. Recreation showed a decline at 4 weeks. Mucous and Phlegm showed a similar pattern of decline at 4 weeks, not resolved by 12 weeks. There was no significant difference between the placebo and the GM-CSF groups in any areas. CONCLUSIONS: The results provide no evidence of a positive effect of GM-CSF compared to placebo. There were problems with the study including a shortage of drug supply before all patients completed therapy.

AIMS: Little is known if patient reported symptoms or QOL have any prognostic value in MPM patients. We report an evaluation of the prognostic value of these factors from data of a recent trial in MPM patients in addition to clinical factors. METHODS: Patients with histologically proven unresectable MPM, not pre-treated with chemotherapy, WHO PS d 2, and adequate hematological, renal and hepatic function were randomly assigned to receive cisplatin 80 mg/m2 iv on day 1, without or with preceding infusion of raltitrexed 3 mg/m2. HRQOL was assessed with the EORTC QLQ-C30 and the Lung Cancer module (QLQ-LC13). Only baseline scores were used for this analysis. The Cox proportional hazards regression model was used for the univariate and multivariate analyses of survival. Included were 10 key symptom and QOL scales and the prognostic index (PI) developed by Legrand et al (2005) which is composed of stage, histology type, time since diagnosis, haemoglobin difference and platelet count. RESULTS: 250 patients were randomised, with 80% male, median age 58 years, with WHO performance status 0, 1, 2 in 25, 62, and 13% of cases, respectively. 229 (91.6%) patients had a valid HRQOL assessment. Univariate analysis indicated symptoms/QOL scores revealing physical (p<0.0001, HR 0.829), cognitive (p=0.0010, HR 0.988), and social functioning (p=0.0009, HR 0.916), fatigue (p<0.0001, HR 1.162), nausea/ vomiting (p=0.0002, HR 1.158), appetite loss (p=0.0006, HR 1.083) and pain

A-89

only 6% were using strong opioids, although 38.8% reported severe pain. Patients who had adequate pain treatment compared with patients with inadequate pain control had significantly better global QoL(72.4 versus 50.8, p=0.01), role functioning (67.1 versus 30.11,p=0.01), emotional functioning (57.4 versus 42.5, p=0.04), physical functioning (65.7 versus 46.02, p=0.01), financial impact (24.07 versus 43.3,p=0.04), intensity of pain (NRS=1.38 versus 8.64,p=0.001), status performance (Karnofsky scale= 83.9 versus 71.3,p=0.02) and depression scores (BDI=11.9 versus 17.5,p=0.02). With regard to symptom scales, adequate pain control group experienced less pain (31.9 versus 79.6, p=0.01), fatigue (32.1 versus 49.5,p=0.01) and constipation (16.7 versus 43.01,p=0.02) than those from inadequate pain treatment group. The predictors of global QoL were depression (beta=-0.6, p=0.01) and age (beta=0.2,p=0.01) adjusted by the worst intensity of pain in last 24 hours and PMI (r2a=0.44). CONCLUSIONS: Results suggest that inadequate pain control can affect the QoL of cancer pain patients. Thus, pain control and assessment should be a primary goal in clinical practice.

63/Abstract 1363 Quality of Life in Patients Following Surgical Treatment for Esophageal Cancer Jan-Hendrik Egberts, Beate Bestmann, Bodo Schniewind, Fred Fändrich, Jürgen Tepel, Thomas Küchler, General and Thoracic Surgery, University of Schleswig-Holstein, Campus Kiel, Kiel, Germany AIMS: AIMS: For patients undergoing oncologic surgery quality of life (QoL) is generally accepted as an additional outcome parameter aside from mortality and complication rates. This is particularly important for patients with comparatively poor prognosis like oesophagus cancer. Therefore, the objectives of this study were to assess outcome in terms of quality of life in patients with esophageal cancer and to investigate differences between patients with adenoand squamous cell carcinoma. METHODS: METHODS: In a prospective single center study, the EORTC-QLQ-C-30 and a tumorspecific module were administered to patients before surgery, at discharge, three, six, 12 and 24 months after surgery. One-hundredand two patients who had undergone surgery for oesophageal cancer in our department from 1998 to 2005 were included; eighty-eight of them provided QoL data for at least one QoL assessment. Statistical analysis was performed using chi-square-tests, t-tests und ANOVA. A global alpha of p<.05 was considered significant (adjusted for multiple tests). RESULTS: RESULTS: Pre-operative (baseline) QoL scores were comparable to healthy controls. Most QoL scores dropped significantly below baseline in the early postoperative period. Even though they recovered slowly during the follow-up period, they never reached pre-operative levels again. Patients with squamous cell carcinoma report poorer QoL than patients with adeno carcinoma. Statistically and clinically significant differences were found on the scales "fatigue", "dyspnoea" and "reduced efficiency". CONCLUSIONS: CONCLUSION: Oesophageal resections are associated with significant deteriorations of QoL, which persist during the follow-up period. Patients with squamous cell carcinoma report stronger impairment which can be attributed the extent and localisation of resection, adjuvant therapy (Radiochemo) and rapid progression of the primary disease. These findings should be considered for postoperative patient-management, which may include as well adjuvant psychooncological support/ supportive treatment.

65/Abstract 1101 Tailored vs. Targeted Interventions to Increase Mammography Adherence in Low-Income African American Women Victoria L. Champion, Research, Indiana University, Indianapolis, Indiana, Jeffrey K. Springston, Advertising & Public Relations, University of Georgia, Athens, Georgia, Terrell W. Zollinger, Robert M. Saywell, Jr, Family Medicine, Kathleen Russell, Environments for Health, Indiana University, Indianapolis, Indiana, Susan K. Steele, Nursing, Louisiana State University, New Orleans, Louisiana AIMS: Breast cancer is the second leading cause of cancer deaths for African American women. The purpose of this randomized prospective study was to compare a tailored interactive computer intervention with a targeted video for increasing mammography screening in African American women. METHODS: Intervention groups were compared to a usual care group to determine effect on mammography adherence. The Transtheoretical Model (TTM) and Health Belief Model (HBM) guided development of intervention content. A total of 344 African American women who were 41 to 75 and had not had a mammogram in the last 15 months were recruited from multiple community sites, consented and randomly assigned to one of three groups: 1) Usual Care, 2)Targeted Video, and 3)Tailored Interactive Computer. Mean age was 50. 63 and the average educational level attained was 12.34 years of formal school. Data were collected in person at baseline and at 3 months post intervention by telephone. Mammography status was measured by self-report and women were classified as to Pre Contemplation (not thinking about having a mammogram), Contemplation (thinking about having a mammogram) or Action (had a mammogram after the intervention). RESULTS: Results indicated that the interactive computer intervention produced the greatest level of adherence (40.0%) compared to the video group (24.6%) and the usual care pamphlet group (32.1%). When subjects in the pamphlet and video groups were combined to form a non-interactive group, this group had a significantly lower adherence than the group who received the interactive computer intervention (27.0% vs. 40.0%). When combining the non interactive technology (pamphlet and video) there was also more forward movement in mammography stage of readiness for subjects in the interactive intervention group (52.0% moved 1 or 2 stages) compared to those in the combined noninteractive group (36.2%). CONCLUSIONS: Tailored and interactive computer based programs can be effective in increasing African American women's mammography adherence in the community setting.

64/Abstract 1280 Impact of Adequacy of Pain Treatment on Quality of Life of Cancer Patients Karine A. São Leão Ferreira, Miako Kimura, School of Nursing, Manoel J. Teixeira, School of Medicine, University of São Paulo, São Paulo, São Paulo, Brazil AIMS: pain is often inadequately treated in patients with cancer and this could damage their quality of life. Thus, we assessed the adequacy of prescribed analgesic drugs using guidelines developed by the World Health Organization, and estimated the impact of pain control on the cancer patient_s quality of life(QoL). METHODS: it is a cross-sectional study.We studied 67 outpatients attending in a General Hospital in Brazil using EORTC QLQ-C30 to assess QoL, Numeric Rating Scale(NRS) to assess pain, Pain Index Management(PMI) to assess adequacy of pain treatment and Beck Depression Inventory(BDI). RESULTS: Mean age was 62.8 years(sd=13.1), 52.2% were female, 62.2% were married and 40.3% had as many as 8 years of education.46.6%(31) had inadequate pain control(PMI=-3 to 0), 77.8% had metastasis, 55.5% had moderate to severe pain (NRS³5) and 62.7% have taken analgesic drugs daily. Among these 31.3% were using NSAIDS, 19.4% weak opioids and

A-90

longevity. RESULTS: 1. The Radio Program _New Dimensions : communications by Radio AM _O Povo , (1010) (1997/2005) and Radio FM University (107,9 MHz) (2006) is broadcasting every Saturdays - from 12 to 1 p.m.- interviewing special guests (953 up to this date) and making important communications. On this liveprogram listeners may call our guests for information, questions and solicitations. Since 1997 the program performed 382 presentations. The most important issues were: to communicate the therapies, practices and techniques for a better QOL and a healthy aging; the mobilization of politicians, leaders and chairpersons of Aging Organizations; the improvement of healthy life styles and the political promotion for the creation of State and County Councils for the Rights of Aging Persons. 2. The Internet Program: since January 2000 our program did have 574 participants making QOL evaluations and 83 of those became permanent members. We can testify that the main risk indicators of QOL are: 1. Financial problems; 2. Stress; 3. Drinking; 4. Lack of physical activities; 5. Smocking; 6. Insecurity; 7. Low self-esteem; 8. Unhealthy nutrition; 9. Health problems and 10. Environmental problems. 3. On a local level (medical offices) personal attention is offered and consulting programs are performed in different cities all over the Country. CONCLUSIONS: This Program, according several communications organs is very important for the health and active longevity for the State of Ceará population.

66/Abstract 1693 The European Organisation for Research and Treatment of Cancer (EORTC) Translation Procedure for Quality of Life Questionnaires: Accomplishments and Implications for Clinical Research Michael Koller, Center for Clinical Studies, University Hospital Regensburg, Regensburg, Germany, Neil K. Aaronson, Psychosocial Research and Epidemiology, Netherlands Cancer Institute, Amsterdam, North Holland, The Netherlands, Jane Blazeby, Surgery, Bristol Roayl Infirmary, Bristol, UK, Andrew Bottomley, Quality of Life Unit, EORTC, Brussels, Belgium, Colin Johnson, University Surgical Unit, Southampton General Hospital, Southampton, United Kingdom, Peter Fayers, Public Health, University of Aberdeen, Aberdeen, United Kingdom, Ramage John, Gastroenterology, North Hampshire, Basingstoke, United Kingdom, Neil Scott, Public Health, University of Aberdeen, Aberdeen, United Kingdom, Karen West, Quality of Life Unit, EORTC, Brussels, Belgium AIMS: The EORTC Quality of Life (QoL) questionnaires are used in international trials and standardized translation procedures are therefore required. This report summarizes the EORTC translation procedure, its accomplishments and problems. METHODS: Translations follow a forward-backward procedure, independently carried out by two native-speakers of the target language. Discrepancies are arbitrated by a third consultant, and solutions are reached by consensus. Translated questionnaires undergo pilottesting. Suggestions are incorporated into the final questionnaire version. Requests for translations come from the module development groups or pharmaceutical industry, and most translations are performed by professional translators. The translation procedure is managed and supervised by the Translation Committee within the EORTC QoL Group. RESULTS: To date, the EORTC QLQ-C30, has been translated and validated into 72 languages, with a further 6 translations in progress. Condition-specific modules have been translated in up to 70 languages. Translations include all major Western, and many African and Asian languages. The following major translation problems were encountered: lack of expressions for specific symptoms in various languages, the use of old-fashioned language, recent spelling reforms in several European countries, and different weights of social issues between Western and Eastern cultures. The EORTC measurement system is now registered for use in over 9000 clinical studies in 80 countries worldwide CONCLUSIONS: The EORTC provides a strong infrastructure and methodology to produce high quality translations of their QoL questionnaires. Translation problems have been identified. The EORTC Quality of Life Group is using a unique cross-cultural database of over 38,000 questionnaire responses to help shed light on whether observed problems are due to procedural/methodological shortcomings or subtle cross-cultural differences in concepts of health, illness and QoL.

68/Abstract 1201 Effects of Socioeconomic Factors on Cancer Survivors's Quality of Life (QoL) Yuka Uchida, Keiko Saito, Respiratory Department, Saitama Medical University, Irumagun, Saitama, Japan, Tatsuyuki Miyashita, Respiratory Department, Saitama Medical University, Irumagun, Saitama, Saitama, Mari Simonagayoshi, Day Care Center, Saitama Cancer Center, Adachigun, Saitama, Japan, Satoshi Morita, Epidemiological & Clin Res Info Management, Kyoto University Graduate School of Medicine, Sakyo-ku, Kyoto, Japan, Kunihiko Kobayashi, Respiratory Department, Saitama Medical University, Irumagun, Saitama, Japan AIMS: In 2002, a national survey was performed in Japan to assess distress among cancer patients (n=7885) using a semi-structured questionnaire (SSQ). At one institution, relationships between patients_ distress and their QOL were also investigated. METHODS: The EORTC QLQ-C30 and the FACIT-Sp were distributed with the SSQ in Saitama Cancer Center and were analyzed two dimensionally. The relationships of each item of the SSQ with the scales of QLQ-C30 and FACIT-Sp were analyzed using multivariate regression analysis with adjustment for gender, age, and cancer type. RESULTS: Of a total of 210 cancer outpatients, 130 patients (male:42%; average age:59 y; performance status rating (PSR) 0-2:89%; breast/lung/gastrointestinal cancer:38/22/21%) completed all the questionnaires. Regression analysis demonstrated that a better PSR significantly correlated with a better QOL status in all of the scales, and that having an occupation correlated with a better QOL in Physical functioning (p=0.014), but that losing a job negatively correlated with all of the scales on functioning and several symptom measures of the EORTC QLQ-C30 (p<0.05) and spiritual well-being (p<0.05). Higher family income significantly correlated with better status in physical and cognitive functioning, global QOL, and spiritual well-being. Unexpectedly, regression analysis indicated that having a partner negatively correlated with emotional and cognitive functioning (p<0.05) and spiritual well-being (p<0.05). Narrative answers on the SSQ indicated that this might be due to anxiety over the family_s future and poor mutual understanding with the partner about the illness. CONCLUSIONS: Losing an occupation through experiencing cancer had a highly negative impact on patients_ QOL.

67/Abstract 1287 Quality of Life Promoting and Healthy Aging Antero C. Neto, Medicine-Gerontology, Federal University of Ceará, Fortaleza, Ceará, Brazil AIMS: The IQVida-Institute for the Quality of Life (www.secrel.com.br/iqvida) associated to the Brazilian Citizen Foundation (www.brasilcidadao.org.br) and the Cooperation Forum (www.pacto.com.br), develops the program _Quality of Life Promotion and Healthy Aging ONLINE METHODS: On a Radio Broadcast, the Internet, Local and Consulting work as part of its modern projects on quality of life and active, creative and healthy

A-91

measured in the form of the function, Y= f (X). Here, the Y is discharge cycle time and availability of bed, and X's are all the factors that affect the Y.Then, the data were analyzed using SAS statistical one-way ANOVA procedure to find if there is any significant variations in outcomes. RESULTS: As a result of these methodologies, patient discharge cycle time was reduced from 375 min to 275 min, a 27% decrease.There was no significant difference in all variance. CONCLUSIONS: Six sigma and Lean integrated methodoligies extents total quality management efforts by using detailed metrics to identify and eliminate process variation.This analysis showed that the use of these integrated methodologies improved internal quality by decreasing admitting delays and improved bed management.In essence, six sigma and lean can take total quality management to the next level, a level focussed on process improvement in quality and service.

69/Abstract 1530 HRQoL & Malignant Melanoma: A Qualitative Assessment Christina Donatti, Oxford Outcomes Ltd, Oxford, United Kingdom, Michael Christensen, Mette Hammer, Novo Nordisk, Bagsværd, Copenhagen, Denmark AIMS: Interviews were conducted with patients with malignant melanoma (MM) about how the condition has affected their healthrelated quality of life (HRQoL). There is limited research examining the changes associated with treatment-related HRQoL for patients with MM. Evidence suggests that psychological, emotional and social factors have an impact on illness perception, HRQoL and survival for patients with melanoma. METHODS: Five patients with MM and eight oncologists participated in individual semi-structured interviews. Each interview was analysed using Atlas ti. RESULTS: From the content of the interviews, a conceptual model was developed. This model presents the key issues that were raised by both the patients and the experts. From this model it was observed that HRQoL issues are affected slightly differently depending on whether focus is given to the actual diagnosis of the malignancy, or, in the treatment of it. Overall, both patients and experts reported HRQoL being affected in a number of different but interrelated areas. CONCLUSIONS: Melanoma survival is a process of three phases: acute survival (or the coping employed in the peridiagnositic period); extended survival (or the coping employed as the patient completes initial treatment and begins a 'watch and wait' strategy); and permanent survival (or long term coping in a disease free state). The experience of melanoma impacts the patient's HRQoL in different ways depending on the stage of the disease with which they have been diagnosed. The HRQoL issues for patients with MM are less severe for those at Stage I and II of the disease. Patients at Stage III and IV do suffer significant impairment in their HRQoL. Stage III and IV patients can be distinguished insofar as the treatment options available depend on the location of the cancer. At Stage IV, there is a general consensus that interferon treatment does not translate into actual survival benefit to patients and thus raises the issue of whether extending a patient's life by only a very small period does in fact improve HRQoL for these patients.

71/Abstract 1662 Question-Induced Response Variation (QRV): The Effects of Comprehension and Judgement Problems on Longitudinal Assessments of Health-Realted Quality of Life (HRQoL) Sally Stapley, Elaine McColl, John Bond, Centre for Health Services Research, University of Newcastle, Newcastle upon Tyne, Tyne and Wear, U.K. AIMS: Response shift is an important issue in interpreting longitudinal health-related quality of life (HrQOL) outcomes. At the same time, CASM (Cognitive Aspects of Survey Methodology) also aims to understand how patients interpret HrQOL, self-report measures. Anomalous post-test findings, often attributed to response shift, may instead be due to problematic question items: here termed Question-induced Response Variation (QRV). Therefore the aim of the current study was to investigate the presence and impact of QRV on longitudinal HrQOL assessments. METHODS: Type 2 diabetes patients were randomly selected from a hospital database. At Time 1, 21 patients enacted both present perspective and when-test thinkalouds. Participants completed items from the ADDQOL, the SF-36 (version 2) and the WHOQOL-Bref. At Time 2 (6-12 months), 16 patients from the original sample enacted both present perspective and then-test think-alouds with the same question items from Time 1. Analyses of verbatim transcripts were aided by pattern matching and reference to the QAS-99, Questionnaire Appraisal System. RESULTS: The impact of age and co-morbidity on self-report judgements was observed at both study time-points. Psychological items were perceived as ambiguous and seemed to evoke general self-perceptions, which were unrelated to health status. A key finding was that response categories on the SF-36 (version 2) Physical Functioning domain brought about comprehension problems and therefore an erroneous pattern of responding. However, at Time 2, where participants did not repeat this pattern, their scores shifted to reflect worsened HrQOL; although their explications for these scores were the same as at Time 1. Therefore QRV was evident within this study population. CONCLUSIONS: QRV is an alternative explanation to response shift, with particular implications for intervention assessments. The presence of QRV also suggests that face and content validity should be crucial criteria when evaluating HrQOL measures.

70/Abstract 1102 Six Sigma and Lean Integrated Methodologies in Health Care to Improve Patient Quality of Life Vijayveer Bonthapally, Pharmacy Administrtion, Joseph Feldhaus, Harvey Rappaport, Pharmacy Administration, Univ Of Louisiana At Monroe, Monroe, LA AIMS: Six sigma and Lean methodologies are used to measure the undesirable variations in health care processes. The objective of this research is to improve operations, reduce costs, better serve the patients and improve the patient quality of life by reducing the variability and admitting delays, and improving bed management. METHODS: Six sigma stands for six standard deviations, which translates into an error rate of 3.4 parts per million. It uses DMAIC (define, measure, analyze, improve, control) and statistical tools to uncover root causes and to improve internal quality and increased patient satisfaction. Lean ensured on-time delivery, equipment productivity, eliminate waste, and improved returns. Lean uses PHYSICAL MAPS. When these are integrated, they reduce patient discharge time and patient waiting time variability, and they improve patient care and satisfaction which leads to more productive employees.The patient discharge cycle time data were taken from a hospital consultant firm. The data were measured in terms of patient discharge cycle time and patient waiting time. The outcomes were

72/Abstract 1219 Treatment Strategies and Improvement in Quality of Life (QoL) in Patients with Chronic Obstructive Pulmonary Disease (COPD) Randi Andenaes, Nursing, Oslo University College, Oslo, Norway, Astrid K. Wahl, Medicine, Oslo University, Drammen, Norway

A-92

reliable and valid for wound care, as they were confirmed two psychometric properties, i.e., the internal consistency and convergent validity.

AIMS: The purpose of this study was to search for evidence of treatment strategies for improvement in QoL in COPD patients. According to European Respiratory Society (ERS), the objectives of the management of stabile COPD are to improve the symptoms and the patients_ quality of life, reduce the impairment of lung function, prevent and treat complications, and increase survival with maintenance of quality of life. METHODS: The present study was a literature review. The following treatment strategies were included; Pulmonary rehabilitation and education, Pharmacology, Domiciliary oxygen, Physical training, Nurse led follow-up in patients_ homes, Nutritional support, and Antidepressants. Surgical procedures were left out. A broad literature search for RCT studies was generated within these areas, and both original studies and review articles from the Cochrane Library, as well as other reviews were analysed. We categorized the findings in three levels of QoL; Global, Health related, and Disease specific QoL. RESULTS: A variety of quality of life measures were used. Generic measures like the SF-36, as well as disease specific such as the Transitional Dyspnea Index, Chronic Respiratory Disease Questionnaire (CRQ)(dyspnea, fatigue, emotional function and mastery), or St. Georges Respiratory Questionnaire (SGRQ) were most used. None of the reviewed studies discriminated between the three levels of QoL, and none included measurement of global QoL. Regarding health related QoL, effect was found with Pharmacologic treatment, where Long-acting beta2agonist demonstrated significant improvement in the SF-36 domain Physical Role. Regarding disease specific QoL, Pulmonary rehabilitation and education showed some improvement in dyspnea, fatigue and activity. Likewise, Long-acting beta2-agonist showed significant improvements in two domains of CRQ; dyspnea and fatigue, and Antidepressants for both emotional function and mastery. For Nutritional support, QoL improved in several areas. CONCLUSIONS: The results suggest that despite the common referrals to the concept of QoL, only a few studies can document any QoL effects.

74/Abstract 1057 Quality of Life of Gastric Cancer Patients in Taiwan: Validation of the Taiwan Chinese Version of the European Organization for Research and Treatment of Cancer (EORTC) QLQ-C30 and EORTC QLQ-STO22 Wei-Chu Chie, Preventive Medicine, National Taiwan Iniversity, Taipei, Taiwan, Chi-Cheng Huang, General Surgery, Cathay General Hospital Sijhih, Sijhih, Taipei, Taiwan, Heng-Hui Lien, General Surgery, Yung-Chuan Sung, Hematology and Oncology, Ching-Shui Huang, General Surgery, Han-Ting Liu, Hematology and Oncology, Cathay Medical Center, Taipei, Taiwan AIMS: The purpose of this study is to report the cross-cultural application and psychosocial property of the Taiwan Chinese version of the EORTC QLQ-C30 and QLQ-STO22. METHODS: Guidelines of translation and pilot test of EORTC were followed, and a total of 100 patients (36 for active treatment group and 64 for follow-up group) were enrolled and interviewed. Multi-item correlation and Cronbach's alpha coefficients were used to test the structure of the tool, Clinical validity was tested by known groups comparison. RESULTS: Both the QLQ-C30 and QLQ-STO22 showed good itemsubscale correlation and internal consistency (Cronbach's alpha coefficient: 0.70-0.94) except cognitive functioning of the QLQ-C30 and eating restriction of the QLQ-STO22. Patients in the active treatment group experienced compromised functional status and worse treatment associated symptoms than those in the follow-up group. Similar results were found when comparisons based on Eastern Cooperative Oncology Group (ECOG) Performance Status and dysphagia grade. CONCLUSIONS: The Taiwan Chinese version of the EORTC QLQ-C30 and QLQ-STO22 showed high internal consistency and expected ability to distinguish different groups of gastric cancer patients in Taiwan.

73/Abstract 1576 WHOQoL-BREF in Wound Care: Reliability and Validity Beatriz F. Yamada, Nursing, Enfmedic Saúde, São Paulo, São Paulo, Brazil, Vera L. Santos, Medical Surgical Nursing, Nursing College/ University of São Paulo, São Paulo, São Paulo, Brazi

75/Abstract 1816 Asthma Quality of Life Questionnare: Translation and Validation Study of the Iranian Version Sara Miri, Immunology, Asthma and Allergy Research Institute, Tehran University of Medical Sciences, Tehran, Iran, Ali Montazeri, Mental Health, Iranian Institute for Health Sciences Research, Tehran, Iran, Zahra Pourpak, Hassan Heidarnazhad, Immunology, Asthma and Allergy Research Institute, Tehran University of Medical Sciences, Tehran, Iran

AIMS: to analyze the reliability and validity of the WHOQOL-bref in patients with different kinds of wounds. METHODS: this study was developed as a major project related to the development of a Wound Version of the Ferrans & Powers QOL Index (QLI_WV). The study was conducted after accomplishment of ethical issues. The overall sample was composed of 362 outpatients from 16 health facilities. They were selected according to some pre established criteria. Reliability was obtained through internal consistency _ IC (Cronbach´s alpha > or = 0.70). The convergent validity though the analysis of correlation between the WHOQOL-bref domains and domains and overall QLI_WV mean scores (Pearson or Spearman correlation coeficient > or = 0.30), based on the following hypothesis: strong to moderate correlation between similar domains and weak between the different ones. RESULTS: Cronbach´s alpha coeficients were 0.76, 0.73, 0.52, 0.72 and 0.68 respectively for Physical Health, Psychological Health, Social Relationships and Environment domains and overall QOL. Correlation from 0.17 to 0.60 (p<0.001) were obtained among the domains and overall QOL mean scores of both QOL instruments except between WHOQOL Physical Health domain and QLI_WV Family domain (r=-0.58; p=0.444). The strongest correlation was obtained between overall QOL from both instruments (r=0.60; p<0.001). CONCLUSIONS: the study points out that the WHOQOL-bref can be considered

AIMS: To translate and validate the Iranian version of the Asthma Quality of Life Questionnaire (AQLQ). METHODS: After permission by Professor Elizabeth Juniper, forward-backward translation method was applied to translate the AQLQ from English into Persian (the Iranian language). It is a 32-item specific quality of life measures for patients suffering from asthma. Each item is rated on a 7-point Likert scale and the questionnaire provides one Overall Quality of Life score and four subscales namely: Activity Limitation, Symptoms, Emotional Function and Environmental Stimuli. Scores range from 1 to 7 and the higher scores for overall index and each subscales indicates a better condition. The questionnaire was pilot tested and then a sample of patients attending to an asthma clinic in Tehran, Iran were interviewed. Statistical analysis was carried out to test reliability and validity of the scale. Internal consistency was used to assess reliability (Cronbach's alpha coefficient) and known groups

A-93

analysis was performed to examine validity. RESULTS: In all 92 patients were interviewed. The mean age of the patients was 46.7 (SD = 14.9), and mostly were female (64%). The asthma grade for patients was as follows: intermittent (25%), mild (27%), moderate (21%), and severe (27%). The analysis showed satisfactory results for the scale reliability. The Cronbach's alpha coefficient for Overall Quality of Life score was 0.94 and for Activity Limitation, Symptoms, Emotional Function and Environmental Stimuli were 0.87, 0.89, 0.64, 0.87, respectively. The Known groups comparison also showed excellent scale validity. The scale discriminated between patients by the disease severity as anticipated (P < 0.0001). CONCLUSIONS: The findings suggest that the Iranian version of the AQLQ is a reliable and valid measure of quality of life in patients with asthma and now can be used in outcome studies in Iran.

similar outcomes; the final results will be updated. We welcome international collaboration in the next phases of development.

77/Abstract 1324 Timing of Transition from Curative to Palliative Care - Based on the Data Obtained from a Series of QoL Surveys in Japan Kikuko Miyazaki, Takeo Nakayama, Health Informatics, Kyoto University School of Public Health, Kyoto, Japan, Mari Saito, General Medicine, Yokohama City University, Yokohama, Japan, Fumi Hayashi, Human Sciences, Toyo-Eiwa University, Yokohama, Japan AIMS: Palliative care is an approach that improves the quality of life (QOL) patients facing the problem associated with life-threatening illness. In the current study, patients QOL, which is the outcome of palliative care, was measured repeatedly over a specific period of time in order to address the issues related to the patients timing of transition from curative to palliative care. METHODS: The current survey was conducted during a period extending from June 2002 to January 2005. Subjects were selected from 91 terminal cancer patients in general wards of a university hospital. Thirty-two of them, who were capable of completing the questionnaire and who consented in writing to participate in the study, were deemed eligible for participation. As a rule, a Japanese ver. of EORTC QLQ-C30v3 was administered to the subjects every once a week. Besides the QOL assessment, the patients medical data as well as the record of the interview of the subjects by psychological counselors were analyzed. RESULTS: Subjects consisted of 18 male and 14 female patients. The average age was 63.5 years old. Through the survey period, a total of 245 responses were obtained from the subjects. The frequency and timing of the present survey differed across subjects depending on their physical/mental condition and stage of cancer. Thus, the survey period was divided into five time periods. The average QOL scores were calculated for each of these time periods. Generally, the physical and psychological QOL of cancer patients decreases dramatically two to three months prior to death. The results obtained from our survey indicate that, in order to provide patientoriented end-of-life care, it is important to ensure sufficient communication among patients, their families, and healthcare providers prior to the arrival of the turning point where patients begin to suffer from a great deal of physical and psychological pain. CONCLUSIONS: In order to respect and protect the dignity of terminally ill patients, it is necessary to ensure early introduction of palliative care.

76/Abstract 1366 Early Phase in the Development of Bone Metastases Quality of Life Kristin Harris, Sukirtha Tharmalingam, Radiation Oncology, Toronto Sunnybrook Regional Cancer Centre, Toronto, Ontario, Canada, Andrea Bezjak, Radiation Oncology, Princess Margaret Hospital, University of Toronto, Toronto, Ontario, Canada, Jackson Wu, Radiation Oncology, Tom Baker Cancer Centre, Calgary, Alberta, Canada, Michael Barton, Faculty of Medicine, CCORE, University of NSW, Liverpool Hospital, Liverpool BC, NSW, Australia, Richard Eek, Medical Oncology, Liverpool Hospital, Liverpool, Sydney, NSW, Australia, Jesmin Shafiq, CCORE, Liverpool Hospital, Liverpool BC, NSW, Australia, Galina Velikova, Cancer Research UK Clinical Centre, St James's University Hospital, University of Leeds, Leeds, UK, Andrew Bottomley, Quality of Life Unit, EORTC, Brussels, NONR, Belgium, Edward Chow, Radiation Oncology, Toronto Sunnybrook Regional Cancer Centre, Toronto, Ontario, Canada AIMS: To report the results from phase 1 in the development of a bone metastases module to accompany the EORTC QLQ-C30 core questionnaire. METHODS: Phase 1 generated a list of relevant QOL issues from literature search, interviews with bone metastases patients and their health care professionals (HCPs). This list was administered to other patients and HCPs as a questionnaire. For each issue, patients were asked to indicate their extent of experience and HCPs the extent of relevance on a scale of 1 (not at all) to 4 (very much) and whether the issue should be included in a final questionnaire. The final stage of phase 1 involves the adaptation of the list following decision rules set out by EORTC Quality of Life Group (QLG). The parameters of the decision rules are: mean score, prevalence ratio, range of responses and priority ratings by patients and HCPs. RESULTS: The 61 issues related to: symptoms (13), function (15), treatment side effects (3), psychosocial domains (27), treatment expectations (2) and other (1). A total of 106 HCPs and 328 patients at 3 cancer centers in Canada were interviewed. The top 15 QOL issues using the EORTC decision rules were: difficulty carrying out meaningful activity, difficulty carrying out usual daily tasks, worry about loss of mobility compromising independence, worry about becoming dependent on others, long-term (or chronic) pain, pain aggravation with movement or weight bearing, lack of energy, able to perform self-care, able to perform role functioning, have a clear, alert mind, feel in control, positive, and confident, emotional stress of diagnosis of advanced, incurable cancer, worry about disease progression, deterioration in condition, and future complications, hope for sustained pain relief, hope treatment will reduce pain as much as possible. CONCLUSIONS: These findings provide provisional data, which will help build a robust module on bone metastases. Phase 1 interviews on-going in Australia have

78/Abstract 1303 Cross Cultural Equivalence in the Quality of Life Scoring Procedure: Difference of QoL Scores of Cancer Patients Between Japan and Netherlands K Yamaoka, Technology Assessment and Biostatistics, National Institute of Public Health, Wako, Saitama, Japan, AA Kaptein, W.Chr Kleijn, Medical Psychology, CL Creutzberg, JWR Nortier, Clinical Oncology, Leiden University Medical Center, Leiden, The Netherlands, K Kobayashi, Respiratory Medicine, Saitama Medical University, Iruma-gun, Saitama, Japan, K Ogoshi, Surgery, Tokai University, Sagamihara, Japan, F Hayashi, Human Sciences, Toyo Eiwa University, Yokohama, Kanagawa, Japan, K Inoue, T Tabei, Endocrinology, Saitama Cancer Center, Kita Adachi-gun, Saitama, Japan AIMS: It is essential to conduct international bridging studies to clarify cross-cultural validity of QOL scales, evaluating each scale

A-94

the lower between Impact on HRQoL and General Satisfaction (r = .001). CONCLUSIONS: The Glausat questionnaire shows good reliability and structural validity properties. The 7 dimensions proposed are stable and well defined in a 22-item form. Results support that the questionnaire can be used to compute an overall meaningful score. Further validity evidences should be collected.

item equally in divergent contexts of cultures, demonstrating appropriate item-response correlations and construct validity of the scales. The Japan and the Netherlands QOL Collaborative Study Group aimed at establishing a standardized, cross-culturally equivalent, evaluation procedure for QOL assessment as an outcome measure in cancer clinical trials. For this major goal, the presentation aimed at to examine QOL scores and scoring methods of the questionnaires using the data from the Netherlands and Japan. METHODS: Comparison of the scores and scoring methods of the EORTC QLQ-C30 and HRQOL-20 was conducted. Subjects were cancer patients (lung, breast, digestive, and prostate cancer) in the Netherlands and Japan. The data were summarized using summary statistics. The structures of the questionnaires were examined using factor analysis and correspondence analysis. RESULTS: The subjects were 265 patients for Japan and 172 patients for the Netherlands. The perspective of the structure for the HRQOL-20 using correspondence analysis, one-dimensional structures were observed in both countries except for the item of _being needed by one's family. As for EORTC QLQ-C30, principal factors analyzed by factor analysis were somewhat different from each other. Especially economic condition was classified into the same factor with social function in Japan. These results suggest that total score (QL2), subscales of role function (RF2), emotional function (EF), social function (SF), fatigue (FA), nausea and vomiting (NV), and pain (PA) could be useful for cross-cultural comparison of the two countries. CONCLUSIONS: In order to establish a standardized evaluation procedure for QOL assessment as an outcome measure in cancer clinical trials, cross-cultural validity between two different cultures in the QOL scoring procedure is examined for the patients in the Netherlands and Japan.

80/Abstract 1284 The Oswestry Disability Index, The Roland-Morris Disability Questionnaire and the Quebec Back Pain Disability Scale: Translation and Validation Studies of the Iranian Version Sayed Javad Mousavi, Faculty of Rehabilitation Sciences, Tehran University of Medical Sciences, Tehran, Iran, Mohamad Parnianpour, Faculty of Mechanical Engineering, Sharif University of Technology, Tehran, Iran, Hossein Mehdian, Centre for Spinal Studies and Surgery, University of Nottingham, Nottingham, UK, Ali Montazeri, Mental Health, Iranian Institute for Health Sciences Research, Tehran, Iran, Bahram Mobini, Orthopedic Surgery, Iran University of Medical Sciences, Tehran, Iran AIMS: To translate the Oswestry Disability Index (ODI), the Roland-Morris Disability Questionnaire (RDQ), and the Quebec Back Pain Disability Scale (QDS) into Persian (the Iranian language), and then to investigate the psychometric properties of the Iranian versions provided. METHODS: The translation and adaptation of the original questionnaires were carried out in accordance with published guidelines. One hundred patients with chronic low back pain were asked to complete a questionnaire booklet (the Iranian versions of the ODI, the RDQ, the QDS, the SF-36, and a visual analogue scale (VAS) for measuring pain. Of the original sample, thirty one randomly selected patients with chronic LBP completed the questionnaire booklet at two points in time: baseline and 24 hours later. Statistical tests were performed to carry out reliability and validity analyses. RESULTS: Cronbach's alpha for the ODI, the RDQ, and the QDS was 0.75, 0.83, and 0.92, respectively. The ODI, the RDQ, and the QDS showed excellent test-retest reliability (Intraclass correlation coefficient = 0.91, 0.86, and 0.86; and all P values < 0.01). The correlation between the ODI, the RDQ, the QDS and physical functioning scales of the SF-36 was 0.66, 0.62, and 0.69 respectively (P < 0.001)indicating very good convergent validity for all questionnaires. The correlation between the ODI, the RDQ, and the QDS and the VAS also was 0.54, 0.36, and 0.46, respectively (P < 0.001). CONCLUSIONS: The findings suggest that the Iranian versions of the Oswestry Disability Index, the Roland-Morris Disability Questionnaire and the Quebec Back Pain Disability Scale are reliable and valid instruments to measure functional status in Iranian patients with low back pain. These are simple and short instruments and their use could be recommended in clinical settings and the future outcome studies in Iran.

79/Abstract 1626 Structure of a New Specific Questionnaire Measuring Glaucoma Treatment Satisfaction: The Glausat Miguel Ruiz, Antonio Pardo, Department of Methodology, School of Psychology, Universidad Autónoma de Madrid, Madrid, Spain, Jose M. Martinez de la Casa, Ophthalmology Service, Fundación Hospital Alcorcon, Alcorcon, Madrid, Spain, Vicente Polo, Ophthalmology Service, Hospital Universitario Miguel Servet, Zaragoza, Spain, Javier Soto, Isabel Fernandez-Arias, Medical Unit, Pfizer Spain, Alcobendas, Madrid, Spain, Ines Fernandez, Scientific Affairs, Artac Bioestudios, Madrid, Spain AIMS: To develop a new questionnaire able to measure Treatment Satisfaction of Glaucoma, and to clarify the underlying dimensions of this construct. Item reduction and factorial validity are discussed. METHODS: After conducting 3 focus groups with chronic patients, literature review, and specialized practitioners debriefing, an expert panel proposed a candidate version composed by 38 items in 7 dimensions: Believes and Expectations towards Treatment (5 items), Treatment Convenience (5), Treatment Efficacy (4), Undesirable Treatment Effects (7), Impact on HRQoL (7), Medical care (5), and General Satisfaction (5 items). A pilot sample of 19 patients was used to asses the candidate questionnaire comprehensibility and wording. A sample of 194 patients, recruited in Madrid and Barcelona was used for item reduction by means of Factor Analysis and Reliability assessment. RESULTS: The questionnaire was reduced to a version composed by 22 items organized in the original 7 dimensions, accounting for 79.8% of variance. Item score distributions tend to be markedly skewed and negative wordings are worse understood. The reduced version presents an overall Cronbach alpha of 0.722, and dimensional alphas ranging from .773 to .926. Dimensions are well formed and correlate in different degrees. The larger correlation is observed between Eficacy and General Satisfaction (r = .489), and

81/Abstract 1729 Evaluating the Placebo Effect when Treating Overactive Bladder Karin S. Coyne, Chris Thompson, Donald Stull, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Zhanna Jumadilova, Outcomes Research, Pfizer, Inc, New York, NY AIMS: Placebo effects are common in many disease conditions, including overactive bladder (OAB). The contribution of placebo effects to changes in treatment outcome measures is poorly understood. The purpose of this analysis was to evaluate differences between drug and placebo groups in changes in bladder diary variables on symptom bother and health-related quality of life

A-95

(HRQL). METHODS: Secondary analyses were performed on data from a 12-week trial of tolterodine. Patients completed the Overactive Bladder Questionnaire (OAB-q) and bladder diaries assessing urinary frequency, urgency, and incontinence episodes at baseline, 4, and 12 weeks. The OAB-q includes a Symptom Bother scale and 4 HRQL subscales (Coping, Concern, Sleep, Social Interaction). Latent growth modeling (LGM) multi-sample analyses between drug and placebo groups were used to evaluate the simultaneous effects of changes in diary variables on Symptom Bother and HRQL. LGM estimates intercepts and slopes of change as latent variables in a structural equation modeling context to assess the relationship of one growth curve with another while modeling measurement error. RESULTS: Among the 505 patients (257 drug, 248 placebo), mean age was 59 years, 76% were women, and 90% were white (no differences between groups). The explanatory model linking urinary variables, HRQL, and Symptom Bother fit well (CFI=0.98; SRMR=0.05). However, when the model was constrained to be equal for drug and placebo groups, there was a significant difference in the models (p<0.001), indicating that the relationships between changes in diary variables and in HRQL and Symptom Bother were not the same for both groups. Improvements in diary variables were generally associated with greater positive impact on HRQL and Symptom Bother in the drug group, particularly with urgency. CONCLUSIONS: A model of the impact of treatment on HRQL and symptom bother was validated using LGM. Results confirm that there are differential effects for drug vs placebo groups. LGM methods have important utility for evaluating outcome differences between treatment groups.

studies that are being developed in various medical settings in our country. Further research is being conducted aimed at determining the population norms to calculate physical and mental components, to pursue item banking and item calibration.

83/Abstract 1542 Validation of a Total Gastrointestional Sympton Rating Scale (GSRS) Score for Use in a Renal Transplant Population Anne Kilburg, PH4325, HE & OR Transplant, Novartis Pharma AG, Basel, Switzerland, Leah Kleinman, Center for Health Outcomes Research, United BioSource Corporation, Seattle, WA, Wen-Hung Chen, Dennis Revicki, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Ari Ganasakthy, Health Economics and Outcomes Research, Novartis Pharmaceuticals, East Hanover, NJ AIMS: Renal transplant (RT) patients often experience gastrointestinal (GI) complaints (e.g. diarrhea, reflux, dyspepsia) as a result of their immunosuppressive regimen. The GSRS subscales have been validated for assessment of symptom burden in this population. The objective of this study was to develop and validate a total GSRS score for use in the RT population. METHODS: Data from three non-randomized clinical studies with similar inclusion/exclusion criteria were used in developing the total score. Two studies were longitudinal with medication intervention allowing for analysis over time; 2 studies included patients with and without GI complaints. Validity was assessed in different studies using the Gastrointestinal Quality of Life Index (GIQLI), PGWBI and EQ-5D. Item analysis including ability to discriminate between presence and absence of GI complaints and between GI complaints was conducted to determine if all items should be retained in the GSRS total score. Internal consistency and test-retest reliability, construct and discriminant validity (using clinician severity assessment) were assessed. RESULTS: All GSRS items demonstrated significant change over time (p < 0.05) in medication groups where change was expected and were able to significantly differentiate between the presence and absence of GI complaints (p < 0.05). The GSRS total score is the mean of the 15 items of the GSRS. Cronbach_s alpha was 0.90. ICC for test-retest reliability over 4-6 weeks was 0.76. Correlation coefficients between the GSRS total score and the GIQLI ranged from 0.29 to 0.89, on the PGWBI from 0.33 _ 0.60 and -0.39 on the EQ-5D. The GSRS total score differentiated between patients without GI complaints and those with mild (p <0.001), moderate (p <0.05) and severe effects (p <0.001). CONCLUSIONS: The GSRS total score provides a valid and reliable method to measure the impact of the variety of GI symptoms suffered by RT patients. It has demonstrated excellent psychometric properties in the RT population.

82/Abstract 1543 Evaluation of the SF-36 in Various Clinical Settings in Uruguay Ana Inés Galain, Juan J. Dapueto, Laura Schwartzmann, Medical Psychology, Facultad de Medicina, Universidad de la República, Montevideo, Uruguay AIMS: The SF-36 has been widely used in Uruguay in different medical situations, and yet the psychometric properties of the instrument have not been reported. In this study, we collected data from three studies to evaluate the performance of the instrument in a large sample of clinical patients using a classic psychometric approach as a step before item banking and item calibration using IRT methods. METHODS: The sample consisted of 823 subjects in three clinical situations: people attending an outpatient program for controlling cardiovascular risk factors (RF group, n=407, 49%); patients with end stage renal disease (ESRD group, n= 354, 43%; hemodialysis 86%, transplanted 14%), and patients with Parkinson's disease (PD group, n=62, 8%). Reliability was evaluated using Cronbach's alpha. Factor analysis was used to depict the structure of the instrument. Differences among groups were studied with ANOVA and Scheffée comparisons. RESULTS: Cronbach's alpha coefficients were high in all the subscales when studied in the overall sample (PF=.90; RP=.88; BP=.84; GH=.71; VT=.79; SF=.76; RE=.77; MH=.81) as well as in the different clinical subgroups. Factor analysis showed that seven factors explained 63% of the variance: PF, PR, VT, GH, BP, RE with one factor loading the items of MH and SF. The PF, RP, SF and RE subscales showed significant differences among known groups, with patients of the RF group getting higher scores than those with chronic diseases, and among the latter PD group having the lower ratings. The effect size of these differences was >.50. CONCLUSIONS: Here we have reported on the psychometric properties of the SF-36 when used in clinical samples in Uruguay showing similar standards for reliability, factor analysis and criteria validity as the original English version and those studied in Spain. These findings validate the data coming out of

84/Abstract 1684 Impact of Treatment with Agalsidase Beta Treatment on Health Related Quality of Life in Patients with Fabry's Disease Xavier Badia, Pilar Perez, Miquel Layola, Health Outcomes Research Europe, Barcelona, Spain, Dolores Checa, Nefrology, Hospital Universitario Virgen de la Macarena, Sevilla, Spain, Andrés Tutor, Internal Medicine, Hospital Virgen de la Salud, Toledo, Spain, Roser Torra, Nefrology, Fundació Puigvert, Barcelona, Spain, Juan Torras, Nefrology, Hospital de Bellvitge, Hospitalet de Llobregat, Spain, Gabriel de Arriba, Nefrology, Hospital General Guadalajara, Guadalajara, Spain, Carlos García-Moncó, Neurology, Hospital de Galdácano, Galdácano, Spain, Josep Mª Puig, Nefrology, Hospital del Mar, Barcelona, Spain, Vicente Torregrosa, Unit of Renal Transplant, Hospital Clínic i Provincial, Barcelona, Spain

A-96

principle component analysis indicated that the variance explained by measures was up to 92.4% and the first and second factor loadings of the residuals were 2.2% and 0.8%. The results of the 35 items model consist to the 36 item model. CONCLUSIONS: Results of the present study indicated that the SF-36 might be statistical unidimensional regarding Rash model. Further discussion and researches were asked to clarify the difference between theoretical construct and empirical data structure of the SF-36. The methodological issues about quality of life measurement were also discussed.

AIMS: Fabry Disease (FD) is caused by the lack of or faulty enzyme needed to metabolise lipids. Agalsidase alfa and beta are enzyme therapies which have been increasingly used in FD treatment. Nevertheless, their effects on health related quality of life (HRQoL) have not been previously analysed. The study objective is to assess the effects of agalsidase beta treatment on HRQoL. A maintenance of HRQoL of patients under treatment is expected, due to the therapeutic objective of treatment. METHODS: A prospective, observational and multicenter study was performed. Patients were followed-up for a maximum period of 24 months, with visits every 3 months. The study was performed by 8 specialists (nefrology, neurology and internal medicine). 17 patients with FD beginning or following treatment with agalsidase beta were included in the study, but only 14 (included in final analysis) completed the follow-up period. Treatment effectiveness was assessed in terms of clinical measures and HRQoL, using the EQ-5D questionnaire. Due to the low number of patients, only descriptive statistical measures were used. RESULTS: 87.5% of patients were male with a mean(SD) age of 36.6(12.1) years. 64.3% of patients presented family history and the mean(SD) of evolution years was 10.3(10.9). 36% of included patients were still receiving treatment with agalsidasa beta for a mean(SD) time of 168.2(145.6) days. During the follow-up period it was shown a reduction in the percentage of patients with pain/discomfort problems (from 57.1% at baseline to 14.3% at 24 months). Other EQ-5D dimensions remained without important changes. Slight improvement of HRQoL was also shown in EQ-5D VAS scores, changing from 74.4(17.0) at baseline to 83.1(12.3) points at month 24. In terms of clinical variables patients were stable during the study follow-up period. CONCLUSIONS: Despite the reduced number of patients included in the study, it provides information about the effect of agalsidasa beta on HRQoL of patients with FD, showing a slight improvement during the 24 follow-up months.

86/Abstract 1466 Cross-Cultural Issues in Measuring Health: Issues in Population Norms for the SF-36 Version 2 Graeme Hawthorne, Psychiatry, Richard Osborne, AFV Centre for Rheumatic Diseases, University of Melbourne, Parkville, Victoria, Australia, Anne Taylor, Population Research & Outcome Studies Unit, South Australia Department of Health, Adelaide, South Australia, Australia, Jan Sansoni, Australian Health Outcomes Collaboration, University of Wollongong, Woden, Australian Capital Territory, Australia AIMS: The SF36 Version 2 (SF36V2) will become a commonly used health status measure across the world. It is important that guidelines for interpreting scores are available. A key issue relates to cultural effects on how health is interpreted. The IQOLA project _ which validated the SF36 Version 1 in different countries _ showed that there were important cultural differences. These differences contributed to the development of the SF36V2. The health status of any individual may be broadly thought of as comprising two parts: a part that reflects the local cultural environment relating to how health is conceptualised and a part that reflects the individual_s health assessment. In constructing local population norms for the SF36V2, these issues become important because they affect the implicit values upon which health is measured METHODS: Based on a random sample of Australians (N=3014), this study reports the derivation of SF36V2 population norms using the identical methods of the instrument developers RESULTS: Small but significant differences between the US norms published by the SF36V2 developers and Australian responses to it were observed on 7/8 scales, and on the mental health summary scale. Although the cause of these differences is unknown, cross-cultural emic effects cannot be ruled out. Australian weights were therefore derived and reported. Population norms by age cohort, gender and health status are reported by T-score as recommended by SF36V2 developers. Additionally, the proportions of cases within SF36V2 T-score deciles are presented. The findings suggest there are statistical artefacts associated with the use of T-scores that have implications for how the data from the SF36V2 are interpreted and analysed. CONCLUSIONS: It cannot be assumed that SF36V2 health weights are universal. The procedures reported in this paper may be used by other researchers where emic effects are suspected and who wish to develop local weights for the SF36V2. The population norms presented may also be of considerable interest.

85/Abstract 1617 Is the SF-36 Unidimensional or Multidimensional? Chia-Huei Wu, Grace Yao, Ping-Chi Lee, Psychology, National Taiwan University, Taipei, Taiwan AIMS: The SF-36 is a widely-used instrument for measuring health status. It has been proposed that there are eight first-order factors and two second-order factors underlying the SF-36. Therefore, two summary scores for physical and mental health usually have been used to indicate an individual_s health status. However, a recent psychometric study on Taiwan population using confirmatory factor analysis revealed that the factor correlation between physical and mental health were highly correlated (r = 0.93), implying that there is only one factor underlying the SF-36. Thus, the main purpose of this study was to examine the dimensionality of the SF-36 from item response theory with Rasch model, to see if the SF-36 is unidimensional. METHODS: The 2001 Taiwan National Health Interview Survey Information System (NHISIS) database were used. There were 12966 subjects included in the analysis . Rasch Partial Credit Model was performed to analysis all 36 items using WINSTEPS. The Chi-square fit indices named infit and outfit MNSQ were as the criteria. MNSQ larger than 1.4 was considered poor fit. A principle component analysis on the standardize residuals of Rashtransformed score was also conformed to detect the possible multidimensionality. Low factor loadings of the residuals indicate unidimensionality. The analysis above was also conducted using 35 items, without the item about health related to one year age. RESULTS: The infit and outfit MNSQ of 36 items were all in the acceptable array, ranging from 0.64 to 1.33, respectively. The

87/Abstract 1601 The Comparison Between Likert Type Scale and Visual Analogue Scale (VAS) in the Measurement of Quality of Life -Take the WHOQOL-BREF as an Example YU-YU HSIAO, Chia-huei Wu, Grace Yao, PSYCHOLOGY, National Taiwan University, Taipei, Taiwan

A-97

Moreover, adjustment of the WTP and the expected rates mentioned above would occur due to the newly introduced information in different scenarios. CONCLUSIONS: According to PT, the general public would under-evaluate their true need for medical resources due to risk-aversion in the domain of gains while patients would over-evaluate the need because of risk-seeking in the domain of losses. The _irrational_ judgments are caused by QOL, expected rate of suffering from specific health problem and expected cure rate. Hence, they may be rationalized by introducing objective index, such as prevalence of disease and effectiveness of treatment to replace expected variables.

AIMS: The WHOQOL-BREF is a commonly used questionnaire in quality of life (QOL) research. It contains 26 items for measuring four domains of life quality with a 5-point Likert-type scale. Though previous studies showed that the WHOQOL-BREF has adequate psychometric properties when the Likert-type scale was used, however, some studies on scaling methods suggested that the Visual Analogue Scale (VAS) may be a more desirable scaling method for tapping a continuous construct, such as the construct of quality of life. Thus, the aim of this study was to evaluate the psychometric properties of the WHOQOL-BREF measured by the Visual Analogue Scale (VAS) and the Likert-type scale, to see if the VAS is more desirable than the Likert scale in measuring quality of life. METHODS: A total of 186 healthy respondents participated in this study. Among the 184 subjects (two subjects_ demographical data are missing), there are 62 males and 122 females. Their ages ranged from 16 to 71, with a mean of 31.88 and a SD of 11.84. Each subject was asked to fill in the WHOQOL-BREF in two different versions, one was measured by Likert-type scale and the other was by VAS. Confirmatory factor analysis was applied to examine the construct validity of the WHOQOL-BREF for each version with a hierarchical model with one second-order factor and four first-order factors. RESULTS: According to the CFA results, fit indices of the Likerttype scale version indicated a bad fit (RMSEA = 0.065; SRMR= 0.094; NNFI = 0.79; CFI = 0.81). On the other hand, fit indices for the VAS version indicated a good fit (RMSEA = 0.081; SRMR= 0.069; NNFI = 0.94; CFI = 0.95). The result showed that the VAS version performs better than the Likert-type scale version. CONCLUSIONS: According to the results, for the WHOQOLBREF questionnaire, scores measured by VAS fit the theoretical model of the WHOQOL-BREF much better than scores measured by the Likert-type scale. This means that the VAS is a more desirable scaling method for the WHOQOL-BREF than the Likert-type scale.

89/Abstract 1686 The Prolabels Database: A New ON-Line Tool to Explore the Wording and Types of Pro Included in Approved Medicinal Products Labels M Caron, Caron, MP Emery, C Acquadro, Mapi Research Trust, Lyon, France, P Marquis, Mapi Values, Bonston, E Piault, Mapi Values, Boston AIMS: Patient-reported Outcomes (PRO) are used in clinical studies to assess patients' treatment benefit. There is an increased interest in examining PRO claims included in approved product labels. To date, there is no single database summarizing PRO claims approved by the FDA or the EMEA by therapeutic area, type of PROs or sponsors. In order to provide easy access to this information, Mapi Research Trust and Mapi Values have developed the PROLabels database. METHODS: Summary Product Characteristics (SPC) of drugs approved through the centralized procedure were gathered from the EMEA website since January 1995. Approved labels of New Molecular Entities posted on the FDA CDER website since January 1998 were also examined. Only data pertaining to efficacy endpoints were considered. Once a PRO claim was identified in an approved label, the corresponding product was added to the database. Further information displayed include: description of clinical studies supporting the claim, the product_s pharmacological action, and data source. RESULTS: At the date of 12/31/2005, the database contains 131 records (61 from the FDA and 70 from the EMEA) for 96 different International Nonproprietary Names (INN). The database can be searched by INN, commercial name, marketing authorization holder, indication, PRO, approval date, and agency. The five most represented therapeutic areas include nervous (32.8%), immune (26.0%), skin and connective tissue (24.4%), musculoskeletal (22.9%), and respiratory (15.3%) systems. Signs and symptoms are the most frequently measured PROs while HRQL represents 22.1%. CONCLUSIONS: The PROLabels database is a useful on-line tool to find which products have obtained a PRO claim when reviewed by the EMEA or the FDA. The database will be weekly updated. Other data sources will be added to include PRO approved through the mutual recognition/decentralized procedure in Europe and claims approved in Canada.

88/Abstract 1735 Applying Quality of Life to Medical Resources Allocation Kai-Chun Chang, Psychology, National Taiwan University, Shulin, Taipei County, Taiwan, Chiung-Ting Chang, Environmental Resources, UNESCO-IHE Institute for Water Education, Delft, The Netherlands, Grace Yao, Psychology, National Taiwan University, Taipei, Taiwan AIMS: This study aims at applying quality of life (QOL) using Prospect Theory (PT, Kahneman and Tversky, 1979; Lenert et al., 1999), incorporating preferences of the general public and that of patients, to medical resources allocation. Adjusting and merging the over-evaluated demand for treatments from patients and underevaluated demand from the general public is predominant for resources allocation. Hence, PT is introduced to serve this purpose while an empirical study based on the patients with kidney problems is conducted to support the theoretical results. METHODS: Study 1: The subjects are 60 patients, 60 cured patients and 60 healthy people. The questionnaire is composed of items from SF-36, the WHOQOLBREF of Taiwan version, and questions related to utilities affected by kidney disease and non-kidney diseases in terms of willingness-topay (WTP), expected cure rate and expected rate of having kidney problems. Study 2: 100 healthy subjects provide answers to WTP, expected cure rate and expected rate of having kidney problems in five fictional scenarios and in five clinical scenarios respectively. RESULTS: Study 1 finds those who have higher QOL (the normal public) would under-evaluate their needs for medical resources while who have lower QOL (patients) would over-evaluate the needs. Study 2 finds expected contraction rate of a disease and expected cure rate of a disease, cause different preferences of resources allocation.

90/Abstract 1644 Treatment Satisfaction and QoL in Hypertensive Patients Carmen Suárez, Internal Medicine, Hospital de la Princesa, Madrid, Madrid, SPAIN, Miguel A. Ruiz, Psychology, Universidad Autónoma de Madrid, Madrid, Madrid, SPAIN, Verónica Sanz de Burgoa, Cardiology, Pfizer SPain, Alcobendas, Madrid, SPAIN, Antonio Pardo, Psychology, Universidad Autónoma de Madrid, Madird, Madrid, SPAIN AIMS: Background: The development of new pharmacological

A-98

treatments demands the assessment of the patient's quality of live and treatment undesired effects, variables which have been traditionally considered as secondary. It has been posted that quality of live variations might influence treatment satisfaction and hence treatment adherence. Aim: To determine the proportion of hypertensive patients satisfied with their pharmacological treatment. Special attention is paid to combinations of good/bad clinical efficacy and good/bad tolerability. METHODS: A sample of 1226 patients was gathered by 258 clinicians from specialized hypertension units representative of the national distribution in Spain. The following data were collected: Sociodemographic and anthropometric data, hypertension risk factors, concurrent diseases, pharmacological treatment, and the SATMED-Q (Treatment Satisfaction), MOS SF-12 (HRQoL) and Morisky-Green (Adherence) questionnaires. RESULTS: Blood pressure was under control for 27.9% of patients and only 12.6% of patients had low added CV risk. Differences are found in the 6 dimensions of treatment satisfaction: Treatment Efficacy and Impact on QoL showed lower mean levels (p<.001) than Convenience, Treatment Believes, Satisfaction with Medical Care and, Undesired Effects. Treatment Satisfaction correlations with HRQoL are lower (r from -.005 to .240) than with adherence (r from -.154 to -.314). Patients with controlled BP levels show better treatment satisfaction, HRQoL, and adherence (p<.001) than those not controlled. Interactions between Treatment Satisfaction dimensions and EficacyTolerability groups are found. Adherence shows a small correlation with cardiovascular risk severity (r = .116) but not with the number of medicines used (r = .005). CONCLUSIONS: Patient reported outcomes show to be sensitive to patient's hypertension severity and disease control, and depict a fair picture of patient_s situation. Treatment Satisfaction, Health Related Quality of Life, and disease control appear to be related with treatment adherence.

measure in the items symptoms (edema, shortness of breath, angina and arrhythmia) and systolic ventricular dysfunction. CONCLUSIONS: Data proved that MacNew is a reliable and valid HRQL instrument for the population studied. It is advisable to conduct further studies to assess the psychometric performance of MacNew, especially with respect to responsiveness.

92/Abstract 1296 Quality of Life in Patients With Peacemaker Treatment Shunsuke Yoshii, Rie Udagawa, Takahide Maenaka, Haruyasu Fujita, Kozaburo Adachi, Kanehisa Morimoto, Takashi Mandai, Japanese Society of Quality of Life Research, Kobe, Hyogo, Japan, Ken Matsuoka, Masami Nishino, Jun Tanouchi, Yoshio Yamada, Osaka Rosai Hospital, Sakai, Osaka, Japan AIMS: The purpose of this study was to evaluate the Quality of Life (QOL) in patients with pacemaker treatment. METHODS: Thirty patients with pacemaker treatment participated in this study. Selfadministered questionnaire including 37 questions divided into 14 categories and 15 pacemaker treatment specific questions with the Life Satisfaction Index (LSI) were used. RESULTS: Pearson's correlation coefficients between our questionnaire and LSI was r=0.87 (P<0.01). Cronbach's alpha coefficients of our questionnaire were high enough to accept for clinical use; 0.92 in well-being, 0.91 in psychological problems, and 0.89 in medical service etc., respectively. Our questionnaire contained 9 main factors and cumulative contribution was 0.85. Compared with before pacemaker treatment, beyond our expectation, two fifth patients indicated the improved total QOLs, two fifth patients showed the deteriorated total QOLs and one fifth patients demonstrated the no changes on total QOLs in after pacemaker treatment. There was no significant total QOL difference between before and after pacemaker treatment. The remarkable QOL impairments were demonstrated in the strata of well-being and mental function etc. after pacemaker treatment. All patients whose total QOLs were deteriorated before pacemaker treatment showed the improved total QOLs after pacemaker treatment, but many patients whose total QOLs were excellent before pacemaker treatment demonstrated the impairments of total QOLs after pacemaker treatment. There was significant positive correlation between passion for life and total QOLs (r=0.76, P<0.01). CONCLUSIONS: These findings indicate that our questionnaire has high enough reliability and potency of validity to use for patients with pacemaker treatment. Although we are easy to understand the merits of pacemaker treatment from the standpoint of medical specialists, we should also pay more attentions for the lack of postoperative cares and counselings on well-being and mental function etc. in the complicated cases from the viewpoint of patient, too.

91/Abstract 1768 Health-Related Quality of Life in Patients With Coronary Heart Disease: Psychometric Properties of a Specific Instrument Karin M. Nakajima, Roberta Cunha R. Colombo, Maria Cecilia Bueno J. Gallani, Nursing, University of Campinas, Campinas, São Paulo, Brazil AIMS: The aim of this study was to verify the psychometric performance of a heart disease-specific instrument of health-related quality of life (HRQL), the MacNew Heart Disease Health-related Quality of Life Questionnaire (MacNew), adapted to the Portuguese language from Brazil, and with the specific objectives of assessing practicality, acceptability, floor and ceiling effects, reliability and validity. METHODS: One hundred and fifty-nine patients with myocardial infarction participated in the study, all of them users of the public health system and under follow-up at the ambulatory of the university hospital. The convergent construct validity was demonstrated by correlation between MacNew and a generic HRQL, The Medical Outcomes Study 36-Item Short-Form Health Survey SF-36. The discriminant validity was assessed by association between scores of each domain with sociodemographic and clinical variables expected to affect HRQL. RESULTS: The MacNew demonstrated practicality and acceptability in the studied group. A ceiling effect was detected in the subscales and global score, with no floor effect. The internal consistency of MacNew domains was satisfactory with alpha values ranging from 0.87 to 0.92. MacNew and SF-36 scores were significantly correlated (p<0.0001), with Pearson correlation coefficients ranging from 0.45 to 0.78. Findings confirmed the majority of convergent correlation hypotheses, especially in the physical, social and global domains and partially supported the convergent validity in the emotional domain. The MacNew questionnaire discriminated difference in the HRQL

93/Abstract 1627 Biological Markers of Importance for Self Rated Health Pia Svedberg, Carola Bardage, Nancy L. Pedersen, Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden AIMS: The predictive value of self-rated health (SRH) for future morbidity and premature mortality is well-known. Less is known about the biological mechanisms behind a good or a poor SRH. The purpose of this study was to examine the associations of frequently used biomarkers and SRH. METHODS: The study population consists of 810 randomly sampled twins, aged 17-82, from the Swedish Twin Registry. Data on SRH were collected by a computer assisted telephone interview. The participants were asked to attend

A-99

the local care centre for blood sampling. Blood was analysed concerning the biological markers used in this study. The associations were evaluated using Pearson correlation coefficient for normally distributed variables and Spearman correlation coefficient for skewed variables. RESULTS: In general, significant relationships were found in ages above 30 years and most commonly in ages 40-59 years. Higher levels of total Cholesterol, Triglycerides, low-density lipoprotein (LDL) and Apolipoproteins B (ApoB) were associated with poorer SRH. High-density lipoprotein cholesterol (HDL) showed significant positive correlations in ages 50-59 years. The positive association is consistent with the knowledge of the protectiveness of these substances against cardiovascular disease. Markers associated with damage to the liver showed negative associations with SRH with significant correlation coefficient for ASAT and for GGT. CONCLUSIONS: We conclude that common biological markers are of importance for SRH among middle aged persons.

95/Abstract 1790 Quality of Life for Caregivers in the SAO Paulo Home Care Program Fernanda Amendola, Nursing, University of São Paulo, São Paulo, São Paulo, Brazil, Maria Amélia C. Oliveira, Miako Kimura, Elizabete F. Sportello, Nursing, University of São Paulo, São Paulo, São Paulo, Brazil AIMS: Home health care comprises a set of integrated actions carried out by a caregiver who receives guidance from the health team. The tasks given to the caregiver, the alternation of routines and the time spent on care can affect the caregiver's quality of life (QOL). The purpose this study is characterize the profile and evaluate the QOL of the caregivers served by the University of Sao Paulo Hospital Home Care Program. METHODS: The WHOQOL-bref, a generic questionnaire composed of 24 questions divided into four domains (Physical, Psychological, Social Relationships and Environment), in addition to two questions about QOL in general, was administered to 38 of the 45 caregivers enrolled in the programme. Mostly were married, catholic, women, aged between 41-70 years of age with incomplete primary education, living at home with low income. Approval of the study was obtained from the Hospital Ethics Committee. Each participant signed an informed consent in which their right to withdraw without penalty was assured. RESULTS: The results of the WHOQOL-bref showed satisfactory reliability (Cronbach_s Alpha= 0.88). On a scale of zero to 100, the Physical domain obtained the highest score (65.04), and the Environment, the lowest (54.11). The aspects that contributed most to high quality of life in each domain were mobility (Physical domain), spirituality (Psychological domain), social support (Social Relationships domain) and the physical environment of the home (Environment). The worst aspects of quality of life were related to sleep and rest (Physical domain), positive feelings (Psychological domain), sexual activity (Social Relationships domain) and opportunities for recreation/leisure (Environment). CONCLUSIONS: Caregivers_ self-recognition and satisfaction derived from caring must be recognized, but national health care systems should take responsibility for caregivers physical and mental health, as well as their human and social rights.

94/Abstract 1742 Impact of Caregiver Depression and Pain on Proxy Assessments of Patient Health-Related Quality of Life After Stroke Shih-Ying Hung, Social and Administrative Pharmacy, University of Wisconsin at Madison, Madison, WI, A. Simon Pickard, Pharmacy Practice, University of Illinois at Chicago, Chicago, IL, Whitney P. Witt, Medicine, Northwestern University, Chicago, IL, Bruce L. Lambert, Pharmacy Administration, University of Illinois at Chicago, Chicago, IL AIMS: Informal caregivers often serve as proxy raters of healthrelated quality of life (HRQL) when patients cannot report on their own behalf. Caregiver depression has been associated with bias in proxy ratings, but few studies have examined the role of caregiver pain. This study aimed to determine if caregiver depressive symptoms and/or pain systematically affect patient-proxy agreement on patient HRQL after stroke. METHODS: Secondary data analysis of 95 stroke patients and their caregivers (dyads) at 6 months poststroke. Mental health and depression were measured by CES-D scale. Effect of caregiver depressive symptoms and pain on patient-proxy differences scores on Health Utilities Index 3 (HUI3) and vitality and social functioning domains of the SF-36 was examined using multivariate regression. RESULTS: Caregivers who reported only pain (mean age 59.7 years, SD12.4) were significantly older than caregivers with neither factors (mean age 53, SD12) (p<0.01), while younger caregivers tended to report having depressive symptoms (mean age 47, SD15). Patients had poorer mental health when caregiver had depressive symptoms and pain (CES-D score 17.4, SD10.2) compared to 9.3 (SD8.0) for patients whose caregivers had neither depressive symptoms nor pain(p=0.01). Controlling for other factors, caregiver depressive symptoms and pain significantly explained dyadic difference scores on the pain attribute (p<0.05). Depressed caregivers tended to underestimate pain experienced by patients, and caregivers with pain overestimated patient pain. An interaction between caregiver depressive symptoms and pain was identified on the pain attribute and vitality domain. CONCLUSIONS: Caregiver pain was found to contribute their perception of the patients' HRQL, particularly patient pain, and the co-occurrence of depressive symptoms in caregivers modified the assessment. Results suggest that caregivers who do not have pain and depressive symptoms would provide less biased proxy assessments of HRQL.

96/Abstract 1794 Analysis of the WHOQOL-Bref and SF-36: Reliability, Validity and Agreement Between Patients of Intensive Care Units and Their Families Miako Kimura, Medical-Surgical Nursing, Nursing College of the University of São Paulo, São Paulo, São Paulo, Brazil, Suely S V. Zanei, Medical Surgical Nursing, Nursing College of University of São Paulo, São Paulo, São Paulo, Brazil AIMS: This study aims to compare the psychometric properties of the WHOQOL-bref and Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) applied to adult ICU patients after discharge from Intensive Care Units and their families and to evaluate the agreement between patients_ and proxies_ responses. METHODS: The sample consisted of 71 patient-proxy pairs. Cronbach_s alfa coefficient was used to evaluate the internal consistency reliability of the instruments. The convergent validity was analyzed by the inter-domain correlations of the instruments. Intraclass Correlation Coefficient (ICC) and weighted kappa were used to analyze the agreement between the pairs. RESULTS: Cronbach_s alpha coefficient was acceptable (e 0.70) for the majority of the SF-36 domains for patients and their families. For the WHOQOL-bref, the coefficients were below 0,70 for all domains in

A-100

both groups. The majority of inter-domains correlation of SF-36 for patients and proxies were positive and significant. WHOQOL-bref inter-domain correlations were positive and significant for the patients, but not for the families. The strength of the agreement between the pairs for the SF-36 domains (ICC from 0,31 to 0,64) was greater than the WHOQOL-bref (ICC=0,25 to 0,41). In both instruments, the agreement was better when the relatives were the parents. Domains and items that evaluated objective conditions presented much higher values of agreement between the pairs. Domains of SF-36 demonstrated greater agreement when patients had 2 or more comorbidities. Regarding the WHOQOL-bref, it was better for patients with 1 or 2 comorbidities. Correlations were weak (<0,3) among domains of the two instruments hypothesized to measure similar constructs both for patients and proxy scores. CONCLUSIONS: This study suggests that the SF-36 is a more suitable instrument for evaluation of quality of life of Brazilian ICU patients than the WHOQOL-bref and the patient-proxy agreement was generally better.

Barcelona, Spain, Jordi Alonso, Health Services Research Unit, Institut Municipal d'Investigació Mèdica (IMIM-IMAS), Barcelona, Spain AIMS: There are no available scales to assess life events and their impact on HRQL addressed to Spanish children. The Coddington Life Events Scales (CLES) is an easily administered self-report instrument developed in the USA designed to assess a range of life events that children/adolescents may have experienced in the previous year. The CLES allows to identify events that the child/youth is having difficulty adjusting to. Positive and negative life events have been linked to increased risk of physicial and emotional problems and requiere cross-cultural adaptation for its use in different culture. The higher the Life Change Unit (LCU) score, the greater the number of significant life events and the risk for illness.The objective of this work was to adapt the CLES into Spanish. METHODS: Two forward translations, expert assessment of the translation and a backtranslation was performed. Pilot testing was based on in-depth interviews and focus groups with children/adolescents, ages from 11 to 20 years old (n=17). The level of understaing of the scales, suggested modifications of text, understandability of the time frames and time of administration were assessed. A test re-test reliability study is currently underway. RESULTS: From a total of 53 events, 14 were not easily understood.'Divorce and marital separation' were difficult to differentiate, 'to become an adult member of a church' was reported as uncommon in Spain and 'being send away from home' was interpretated in a negative sense.The introduction of the scales had also some minor understanding problems for younger children. Time frames were acceptable and understood. Administration time ranged from 10 to 20 minutes (median=17). Reliability estimates will be presented. CONCLUSIONS: The content of the CLES was mostly understandable in Spain, but some linguistic modifications were necessary. Administration time is similar to the original. Before its use,it is necessary to fully assess its psycometric properties. Funding: Instituto de Salud Carlos III grants: FIS Expte PI042315 and Red IRYSS G03/202. DURSI-GENCAT (2005-SGR-00491).

97/Abstract 1608 Response Shift in the Global Well-Being of Forensic Psychiatric Outpatients Yvonne H. Bouman, Research, Pompestichting, Nijmegen, the Netherlands AIMS: In a study of the subjective well-being of patients treated in forensic psychiatric community based settings, patients rated their global well-being more positively after six months, although social indicators of quality of life and domain-specific subjective indicators hardly changed. Using the concept of response shift (Schwartz and Sprangers, 2000) these changes in the uniscales of subjective wellbeing will be explored. METHODS: In a longitudinal multi-site study 135 patients treated in a community based setting were interviewed on their quality of life. 100 of these patients were again interviewed after six months. The population studied, consists of male patients diagnosed with an Axis 2 disorder or Axis 2 traits (DSM-IV, 1994), and all exhibiting delinquent behavior. Using the Lancashire Quality of Life Profile (Oliver et al, 1996) these forensic psychiatric outpatients' life satisfaction and social indicators of QoL on seven life-domains and four global well-being indicators are measured with an interval of six months. Using the non parametric McNemar test and a paired sample t-test the differences between t=1 and t=2 are measured. Using a linear regression the different conceptualization of the global measures on t=1 and t=2 are explored. Patients are asked to rate the importance of each domain and to fill out the Life Regard Index for existential state with two factors: fulfillment and framework. Differences between t=1 and t=2 will be presented. RESULTS: During the presentation three different causes of response shift: recalibration, reprioritization, reconceptualization will be explored. CONCLUSIONS: Implications for the assessment of quality of life in forensic psychiatry and its usefulness in treatment evaluation will be discussed. Due to the dual character of forensic psychiatry risk assessment plays an equally important role as treatment results in terms of health status. Attention for this specific characteristic is necessary.

99/Abstract 1595 Does Eletroconvulsivetherapy Influence Quality of Life? Paula B. Antunes, Marcelo P. Fleck, Psychiatry, Hospital de Clinicas de Porto Alegre - UFRGS, Porto Alegre, Rio Grande do Sul, Brazil AIMS: To evaluate the effect of electroconvulsive therapy on clinical improvement and quality of life of inpatients at a university hospital in Porto Alegre, Brazil. METHODS: Using an observational and longitudinal design, 30 patients were assessed for clinical symptoms and quality of life (QOL) right before and 24 hours after the whole treatment. Symptoms were assessed with Behavior and symptom identification scale (BASIS-32), Brief Psychiatry Rating Scale _ Anchored (BPRS-A) and Clinical Global Impression (CGI). QOL was assessed with the World Health Organization Quality of Life Instrument-bref (WHOQOL-bref). RESULTS: From the 30 assessed patients, 15 had Major Depression (50%), 11 had Bipolar Disorder (36,6%) and 4 had Schizophrenia (13,3%). ECT was indicated in 25 (83,3%) patients for depressive episode, which 5 (16,7%) of them with psychotic symptoms. There were significant difference in the pre and post ECT scores at the BASIS-32 (68,9±19,4 vs. 42,3±22,2; p<0,001), BPRS-A (23±7,1 vs. 13,5±8,5; p<0,001) and CGI (5,9±0,6 vs. 3,2±1,5; p<0,001). There were also significant difference in all WHOQOL domains, with the most difference in physical (33±15,6 vs. 58,7±18,4; p<0,001) and psychological (31,6±18 vs. 56,6±19; p<0,001) domains, then in social relationships (44,7±23,5 vs. 58,3±20; p0,003) and environment (51,5±13,8 vs. 59,5±13,5; p0,003)

98/Abstract 1700 Cross-Cultural Adaptation of the Coddington Life Events Scales into Spanish Ester Villalonga Olives, José María Valderas, Health Services Research Unit, Institut Municipal d'Investigació Mèdica (IMIMIMAS), Barcelona, Spain, Jorge Andrés Palacio Vieira, Luís Rajmil,

A-101

domains. CONCLUSIONS: ECT was associated with a clinical and quality of life improvement.

at T0. Regression analysis was used to investigate the magnitude of the mean difference of utility between Hosp and No Hosp after adjusting for demographics, comorbidities, and inpatient costs. RESULTS: For all estimates except VAS, the Hosp group had a greater decline in utility over time than the No Hosp group (all p<0.01). Utility scores at T0 were lowest for both groups using the SF6DVAS estimate (Hosp=0.59, No Hosp=0.66) and highest using the Nichol OLS (Hosp=0.77, No Hosp=0.82). Change in utility score from T0 to T1 was largest for both groups using the Lundberg estimate (Hosp change=-0.03, ES=-0.21; No Hosp change=-0.01, ES=-0.10, p<.01). Change scores and ES were minimal for all other estimates. Controlling for underlying comorbidities did not change the magnitude of the effect. CONCLUSIONS: An increase of hospital days does not appear to have a significant impact on utility, as only the Lundberg estimate revealed a _small meaningful change. Because this is a sicker population afflicted with multiple comorbid disorders, one may conclude that hospitalizations are more common and may not have as profound an impact as in a healthier population.

100/Abstract 1134 Promoting Quality of Life for Patients with Schizophrenia in Hong Kong Ka Fai Wong, Alan Yat Kwan Tsang, Mei Kuen Li, Nursing, The Open University of Hong Kong, Hong Kong, China AIMS: Case management is a multidisciplinary patient care model. It is adopted in healthcare practice for promoting quality patient care. Despite existing evidence supports the effectiveness of case management in enhancing patient care, there is little work examining the issue particularly in the mental health setting. As quality of life (QOL) is an important patient care outcome and schizophrenia is a prevalent mental health problem in Hong Kong, this study aimed to explore the effect of a case management model in promoting QOL for schizophrenic patients. METHODS: A single case study design was used. Five registered nurses and five schizophrenic patients from a mental hospital were purposely recruited to participate in a case management model of care for six months. The nurse, as case manager, was responsible for planning and coordinating health services to meet the designated patient_s needs. Indicators of QOL including patients_ clinical and social functioning states were evaluated by validated questionnaires before and after the intervention. Patients_ experience was further assessed through semistructured interview. RESULTS: There was significant improvement in patients_ clinical and social functioning states. Patients_ positive symptoms had been reduced by 35%. Their self-care ability had been improved by 30%. Their social contact had been improved by 25%. Their perceived role responsibility had been improved by 22%. Content analysis of the interview data illustrated four themes: (1) enhancing own problem solving ability, (2) receiving more recognition from health care providers, (3) increasing participation in own patient care, and (4) family_s satisfaction. CONCLUSIONS: Findings supported the effectiveness of case management in promoting QOL for schizophrenic patients and justified its use in local mental care setting.

102/Abstract 1159 The Clinically Important difference (CID) From Primary Total Hip and Knee Arthroplasty for the Western Ontario McMaster University (WOMAC) OA Disability Index Bert M. Chesworth, Physical Therapy, University of Western Ontario, London, ON, Canada, Nizar M. Mahomed, Surgery, University Health Network, Toronto, ON, Canada, Robert B. Bourne, Surgery, University Hospital-LHSC, London, ON, Canada, Aileen M. Davis, Outcomes and Population Health, Toronto Western Research Institute, Toronto, ON, Canada AIMS: The objective was to determine the CID from primary total hip (THR) and knee (TKR) replacement surgery using the WOMAC OA disability index. METHODS: A pre-post within-subject design was used to collect WOMAC scores at decision for surgery in clinic and at 1-year follow-up by mail on primary THR and TKR patients (n=4,083). At 1 year, transition ratings (15-point scale) for pain and for function were obtained as well as a global assessment of willingness to go through surgery again (responses: yes, no, uncertain). Seventy-four percent of patients responded to the 1-year survey (n=3,003). A transition rating of "a good deal better" defined the CID. WOMAC scores were reverse scored and standardized to a score out of 100 (0 = extreme pain/disability; 100 = no pain/disability). Mean WOMAC change scores and 99% CIs (pain, function and total score) for grouped transition ratings and willingness to go through surgery again were compared. Patient characteristics across transition rating categories were examined separately for THR and TKR patients with the Chi square, ANOVA and Kruskal-Wallis tests. RESULTS: The CID for TKR surgery was 25 for both pain and function. For THR surgery, the CID was 30 for pain and 27 for function. WOMAC change scores for those who would have and were uncertain about having the surgery again validated the CID values. Patients with post-operative complications requiring hospitalization were less likely than those without a complication to rate their change in function as "a good deal better" or more, RR (95% CI) = 0.73 (0.63, 0.84), p<.0001. CONCLUSIONS: For THR and TKR surgery, the CID is a range of values that varies by joint and the construct being rated. Improvement that is at least "a good deal better" is an appropriate threshold for defining the CID. This means that successful THR/TKR surgery should be defined independently for hips and knees by large effects that consider pain and function separately. Future work should validate these findings.

101/Abstract 1368 Comparison of SF-36 Utility Estimates: The Impact of Hospitalization on Change in Utility Michael B. Nichol, Pharmaceutical Economic and Policy, Tom Dow, Tara K. Knight, Joanne Wu, Pharmaceutical Economics and Policy, University of Southern California, Los Angeles, CA, Femida GwadrySridhar, Medicine, University of Western Ontario, London, Ontario, Canada AIMS: 1) To assess the impact of an increase in hospitalization days on change in SF-36 based utility estimates from Baseline (T0) to Year 2 (T1); 2) To compare the impact across six utility estimates. METHODS: The study population was selected from a health maintenance organization dataset of 6,923 patients. Six utility estimates were derived for both T0 and T1 SF-36 scores, including Nichol OLS, SF6DSG, SF6DVAS, Lundberg, Shmueli, and Visual Analogue Scale (VAS). Subjects were categorized into one of two groups: those that had an increase in hospital days (Hosp) versus those that had no increase in hospital days (No Hosp) from T0 to T1. Change in utility scores and effect sizes were calculated. T-tests were used to compare in the change of utilities from T0 to T1 between Hosp and No Hosp. Effect size (ES) was calculated as the difference in utility between T1 and T0 divided by standard deviation of utility

A-102

of negatively-framed PCF (i.e., deficiency items). CONCLUSIONS: While capability related items cannot be scaled together with deficiency related items, as a group, all deficiency items are sufficiently unidimensional. This is one of the first studies to demonstrate that negative PCF can be described by a single score. Furthermore, because negative PCF is scaleable, it may be amenable to item response theory- based approaches such as item banking and computerized adaptive testing.

103/Abstract 1631 Minimally Important Difference and Standard Error of Measurement: Further Evidences Supporting the Connection of These Two Measures of Meaningful Changes in Patients Reported Outcomes Javier Rejas, Health Outcomes Research, Medical Unit, Pfizer España, Alcobendas (Madrid), Spain, Antonio Pardo, Miguel A. Ruíz, School of Psychology, Universidad Autonoma de Madrid, Madrid, Spain AIMS: Patient reported outcomes measurements have become a powerful source of evidence to both the appropriate management of patients from the clinic perspective in daily care, and from the health decision makers when collecting robust humanistic data to support their decisions. However, a unique estimator of clinical relevance is currently lacking. METHODS: Scores determined with four Patient Reported Outcome scales (MOS-sleep, Sheehan disability, Covi´s anxiety and Raskin´s depression scales), included in a cohort of patients with Neuropathic pain (n=603) assessing the analgesic effectiveness of gabapentin. SEM and MID values were determined for all scales and patients, then, were classified in three categories: improvement, no-change and impairment. Agreement between both criteria was assessed by Cohen´s kappa and Kendall´s tau-b coefficients. RESULTS: The 1-SEM criteria showed the higher values of agreement and correlation coefficients in all four scales. The values ranged between 0.83, and 1.00 for kappa coefficient and between 0.88 and 1.00 for tau-b coefficient; higher than those observed with 1.96- and 2.56-SEM criteria. Sensitivity analysis, performed with a sub sample of patients split according to analgesic response with gabapentin, confirmed the findings of base case analysis. CONCLUSIONS: One-SEM criterion may be an alternative measurement to the MID criteria when evaluating the meaningfulness of a change in Patient Reported Outcome scales.

105/Abstract 1428 A Randomized Controlled Study to Assess the Impact of a Spiritual Intervention in Patients With Cancer Aimee N. Swartz, David R. Buchanan, Public Health, University of Massachusetts, Amherst, Amherst, MA AIMS: To assess the effect of a spiritual intervention, an audio recording of prayer and Scripture, in improving QOL and relieving distress cancer patients. Four primary hypotheses are under study: the intervention will 1)increase spiritual well-being 2)decrease distress 3)decrease death anxiety 4)increase QOL. Two secondary hypotheses are under study: 1)patients with advanced cancer will show greater benefit (i.e.increased spiritual well-being, decreased psychological distress) than those with early stage cancer 2)patients with high religiosity will show greater benefit than those with low levels of religiosity. METHODS: Approximately 150 subjects are being recruited from Baystate Medical Center in Springfield, Mass. Inclusion criteria: cancer diagnosis, English-speaking, and capacity to understand informed consent and to complete questionnaires. Subjects with brain metastases are excluded. Subjects are stratified using block randomization by stage of cancer (advanced/early), gender, and religion (Christian/Jewish). Subjects assigned to the intervention arm receive a recording with Scripture and prayer overlaid on nature sounds. Subjects assigned to the control group receive a recording with nature sounds only. Subjects in both groups are asked to complete the following instruments at baseline and at the end of 30 days: 1)Functional Assessment of Chronic Illness Therapy - Expanded Spiritual Well-Being 2)Hospital Anxiety and Distress Scale 3)Death Attitude Profile Revised 4)Duke University Religion Index. All subjects are asked to complete a brief questionnaire to collect demographic data and brief medical histories. RESULTS: The presentation will feature interim results on the study's four primary hypotheses and two secondary hypotheses. CONCLUSIONS: Although spirituality has been identified as a critical domain in QOL and end of life care, few studies have tested spiritual interventions. It is our hope that this study will ultimately result in the development of an intervention that can be used broadly to improve the QOL in patients with cancer.

104/Abstract 1461 Evaluating the Dimensionality of Perceived Cognitive Function Jin-shei Lai, Zeeshan Butt, Lynne Wagner, David Cella, CORE, Jerry Sweet, Psychology, ENH & NWU, Evanston, IL AIMS: Cognitive dysfunction is common in various clinical groups and typically assessed by neuropsychological testing, which is expensive and time-intensive. Recently, there has been interest in measuring self-reported perceived cognitive function, PCF. Yet, it remains unclear what PCF scales measure. This study attempts to address this question by evaluating the dimensionality of PCF. METHODS: PCF was measured on 984 cancer patients using the Functional Assessment of Cancer Therapy-Cognition, FACT-Cog, which assesses self-reported mental acuity, concentration, memory, verbal fluency, and functional interference. Each subscale includes negatively-worded (i.e., deficiencies) and positively-worded (i.e., capabilities) items. Subscale reliability was assessed by Cronbach's alpha and item-total correlations, and dimensionality was assessed using confirmatory factor analysis (CFA) and a CFA-based bi-factor analysis (BFA). RESULTS: Cronbach's alpha ranged from 0.530.91. Capability items had low item-total correlations (range 0.050.27) and Spearman's rho (range 0-0.32) within each subscale. Alphas increased when they were considered separately as a capability subscale. CFA results indicated all revised subscales were unidimensional (CFI: 0.92-1; TLI: 0.98-1). BFA results showed that all deficiency subscales had higher loadings on general PCF than on their own subscales, while capability subscale had higher loadings on the subscale than on the general PCF. This suggests that capabilities and deficiencies are distinct from one another. A follow-up BFA using only negative items supported the sufficient unidimensionality

106/Abstract 1444 State of the QoL Data Bank for Germany Thomas Kuechler, Beate Bestmann, Maria Berend, Bjoern Malchow, Reference Center Quality of Life in Oncology, Bernd Kremer, General and Thoracic Surgery, University Hospital Kiel, Kiel, Germany AIMS: 1998 the German Cancer Help granted 1998 the Reference Center -Quality of Life in Oncology_ with the overall aim to actively support studies and activities on outcome research with focus on QoL and/or other patient-oriented measures. One of the long-term aims of this project is to establish a databank on QoL for most tumor entities and in different treatment modalities in Germany. METHODS: After now 8 years of data collection the state of the data bank can be

A-103

assess the stability of the outcomes. RESULTS: The final multivariate Cox regression model retained four parameters as independent prognostic factors for survival: gender with a hazard ratio (HR)= 1.32 (P=0.03); performance status with HR= 1.63 (P=0.032); patient's self-reported score of pain with HR= 1.11 (P<0.001) and dysphagia with HR= 1.12 (P=0.003). These variables were included in respectively 62%, 52%, 98% and 78% of the 5000 bootstrap generated multivariate analyses, thus supporting their independent prognostic value. In addition, the final model containing the above mentioned variables was the most selected out of all possible prognostic models during the bootstrap evaluation. CONCLUSIONS: These results suggest that patient's reported HRQOL provide independent prognostic information for survival in advanced NSCLC patients. Additional statistical approaches are useful in further exploring the robustness of the outcomes when undertaking such prognostic studies.

summarized as follows: a total of more than 8,700 datasets covers (from pre-op to two years post-op) patients with colorectal- (3,930), liver (839)-, gastric (999)-, esophageal (440)-, and pancreatic (884) cancer. Beside these data on gastrointestinal cancer patients, the data bank contains QoL data of 1617 patients with lung cancer, 2878 patients with prostate cancer, 746 patients with renal cell carcinoma and QoL data of 1129 healthy controls. RESULTS: Using this databank a total of 17 medline-listed publications mainly in the broad field of surgery has been published. It has to be stated, that receiving the benefits of _impact-points from significant publications completely changed _ not only at our department _ the attitude of surgeons toward QoL-assessments or the use of QoL related information: these and related topics (like psychosocial oncology) gained increasing awareness. This in turn facilitated the overall cooperation as well as the setup of new interdisciplinary trials. The greatest benefit may be seen in the fact that the quality of medical data assessment (reliability and completeness) substantially improved during the given timeperiod as well. CONCLUSIONS: Based on these data it is possible to sketch _typical courses of QoL in different diagnostic groups. Furthermore this large number of data sets allows for assessment of effects like different surgical procedures, adjuvant regimens or psychosocial interventions. Since the quality of medical data in our department of surgery improved substantially, we are positive concerning future impact of QoL-aspects in surgical decision-making.

108/Abstract 1452 Quality of Life in Patients with Oral Carcinoma: - Five Year Prospective Study Eva Hammerlid, Otorhinolaryngology Head&Neck Surgery, Sahlgrenska University Hospital, Göteborg, Sweden, Mats Nordgren, Otorhinolaryngology, Malmö University Hospital, Malmö, Sweden, Kristin Bjordal, Radiation Oncology, The Norwegian Radium Hospital, Oslo, Norway, Magnus Jannert, Otorhinolaryngology, Malmö University Hospital, Malmö, Sweden

107/Abstract 1300

AIMS: To evaluate the health-related quality of life (HRQL) of patients with oral carcinoma in a prospective longitudinal multicenter study at diagnosis, and after one and five years in relation to tumour location and treatment modality. METHODS: 122 patients (mean age 61; 62% males) with oral carcinoma were evaluated with standardized HRQL questionnaires; the European Organization for Research and Treatment of Cancer, Quality of Life Questionnaire Core30 (EORTC QLQ-C30) and the EORTC Head and Neck Cancer Module (EORTC QLQ-H&N35) six times during the first year and then after five years. RESULTS: After one year 75 % of the patients were alive after and after five years 52%. The compliance with regard to the HRQL questionnaires in surviving patients was 77% after one year and 94% after five years. Most functions and symptoms deteriorated significantly during treatment but the majority returned to pre-treatment scores after 6- 12 months. However, problems with teeth, dry mouth and sticky saliva remained a clinically significant problem after one year and also five years after diagnosis. Patients treated with surgery only had fewer problems over time compared with patients receiving other treatment regimes (radio and chemotherapy). Survivors reported better HRQL than the nonsurvivors at diagnosis and at the one-year follow-up. HRQL at diagnosis was associated with survival. CONCLUSIONS: HRQL at diagnosis for patients with oral carcinoma seems to be an important factor for the prognosis of both HRQL over time and survival. Treatment of oral carcinoma often results in long-term side effects such as dry mouth, problems with teeth and sticky saliva.

Investigating Model Selection Stability in Predicting Survival with Health-Related Quality of Life Data in Advanced Cancer Patients. A Multivariate Analysis of Prognostic Factors of EORTC 08975 Fabio Efficace, Andrew Bottomley, Quality of Life Unit, EORTC Data Center, Brussels, Belgium, Egbert Smit, Department of Oncology, Vrije Universiteit Medical Center, Amsterdam, The Netherlands, Pilar Lianes, Department of Oncology, Hospital de Mataro, Mataro,, Spain, Catherine Legrand, Channa Debruyne, EORTC Data Center, Brussels,, Belgium, Franz Schramel, Department of Pulmonology, St.Antonius Hospital, Nieuwegein, The Netherlands, Hans Smit, Rijnstate Hospital, Arnhem, The Netherlands, Rabab Gaafar, National Cancer Institute, Cairo,, Egypt, Bonne Biesma,, Jeroen Bosch Ziekenhuis's, Hertogenbosch, The Netherlands, Chris Manegold,, University Medical Center, Mannheim, Germany, Corneel Coens, Quality of Life Unit, EORTC Data Center, Brussels,, Belgium, Giuseppe Giaccone, Department of Oncology, Vrije Universiteit Medical Center, Amsterdam, The Netherlands, Jan van Meerbeeck, Department of Oncology, University Hospital, Ghent, Belgium AIMS: Several recent studies have suggested that health related quality of life (HRQOL) data independently predict survival in advanced cancer populations. However, one of the challenges of undertaking such studies is to control for the effect of possible harmful multicollinearity. This is particularly problematic for multidimensional measures of HRQOL. The aim of this study was to examine if baseline patient's self-reported HRQOL data provide prognostic information for survival and also to investigate the stability of the final model. METHODS: HRQOL was measured in 391 advanced non-small cell lung cancer (NSCLC) patients using the EORTC QLQ-C30 and the EORTC Lung Cancer module (QLQLC13). These data were collected in a recent international randomized controlled trial (EORTC 08975). The classical Cox proportional hazards regression model was used for both univariate and multivariate analyses of survival. In addition, a bootstrap validation technique, based on 5000 simulation datasets, was used to

109/Abstract 1015 Longitudinal Quality of Life Study in Patients with Metastatic Gastric Cancer Analysis Modalities and Clinical Applicability of QoL in Randomized Phase II Trial in a Digestive Oncology Franck Bonnetain, Biostatistics Epidemiology, FFCD & Centre regional de lutte contre le cancer, Dijon, France, Olivier Bouche, Gastroenterology, CHU REIMS, REIMS, France, Patrick Arveux, DIM, Centre Georges Francois Leclerc, Dijon, France, Thierry

A-104

0.17 +/- 0.04 (psychosis); 0.17 +/- 0.02 (GAD); 0.22 +/- 0.02 (phobia); and 0.24 +/- 0.02 (depressive episodes). These differences in preference-based scores exceed the values estimated for common physical disorders, and substantially exceed the levels previously suggested in the literature as economically important. CONCLUSIONS: Preference measurement over mental disorders has been largely neglected. This work has found large differences in mean preference scores between people who do/do not have a mental disorder. Further research is needed into the measurement of preference-based values in this population.

Conroy, Oncology, Centre Alexis Vautrin, Vandoeuvre les Nancy Cedex, France, Laurent Bedenne, Gastroenterology, CHRU la Bocage Dijon, Dijon, Dijon AIMS: The aim of this study was : to compare the longitudinal quality of life (QoL) between LV5FU2-irinotecan and LV5FU2 alone or LV5FU2-cisplatin in a randomized Phase II trial in patients with metastatic gastric adenocarcinoma; - to explore the interest of QoL assessment in a phase II study by investigating analysis modalities and their clinical applicability METHODS: The number of patients for was initially calculated according to an Ensign-MiniMax optimal three-stage design with 0.05 alpha and 0.1 beta levels. Among 134 eligible patients, QLQ-C30 scores were collected and described at each 2 monthly follow-up visit during 6 months. The frequencies of QLQ-C30 score improvement were calculated and mixed models for repeated measurements were applied with or without extreme poorest imputation for missing scores. The time until definitive global health score (GHS) deterioration was estimated. RESULTS: The QLQ-C30 completion rate was of about 90% at baseline. At the 3rd follow-up the completion rates were 55% (N = 12) in LV5FU2, 38%(N = 8) in cisplatin and 59% (N = 17) in irinotecan arms. Completion rates were systematically higher in patients who did not stop treatment.At the 3rd follow-up, patients with a stable or improved GHS ranged from 11% in the LV5FU2-cisplatin arm to 18% in the LV5FU2-irinotecan arm. With or without imputation, the irinotecan-based-therapy presented 14 to 15 scores with a better QoL. The median time until definitive GHS deterioration was: 5.5 months in the LV5FU2 and LV5FU2-cisplatin arms and 5.8 in the LV5FU2-irinotecan arm, respectively. CONCLUSIONS: This study highlights a better impact of LV5FU2-irinotecan. According to our study and binding phase II designs, we are convinced that QoL analyses would be relevant to complement the benefit-risk judgement from the patients point of view. Future research should be addressed in order to more accurately describe the standard of reporting and analysing QoL data in a phase II trial.

111/Abstract 1674 Estimating a Preference-Based Index for the Overactive Bladder Questionnaire John E. Brazier, Yaling Yang, Aki Tsuchiya, Health Economics and Decision Science, University of Sheffield, Sheffield, South Yorkshire, UK, Karin Coyne, Bethesda, Maryland, UK AIMS: The aim is to estimate a preference-based single index for the Overactive Bladder Questionnaire. METHODS: The Overactive Bladder Questionnaire (OAB-q) is a 33-item questionnaire that has been shown to be valid in both continent and incontinent overactive bladder (OAB) patients. A health state classification was derived using items from this questionnaire selected using a range of psychometric techniques. It has 5 dimensions (urge, urine loss, sleep, coping and concern); each with five levels and defines 3125 states. Ninety eight states were selected using a balanced design for valuation by a representative sample of the UK general population. Time trade-off was administered by interview and each respondent valued eight states. These data were modelled at the individual level using a random effects specification and mean values were modelled by OLS. RESULTS: A sample of 311 respondents was interviewed and in all they provided 2487 health state values. Mean values were above 0.6 for most states. Coefficients on the dimension levels broadly support the ordinality of the OAB-q health state classification. The RE model performs better than the mean model in terms of inconsistency of coefficients within levels in OAB dimensions, but the mean model does better than the RE model in prediction with smaller mean absolute error (0.044) and fewer of errors greater than 0.05. Given that the primary purpose of modelling is prediction, the mean model has been chosen. The results of this model were applied to a trial data set at baseline and 12 weeks after treatment and found to produce standardised effect sizes that were comparable to those of the original dimension scores. CONCLUSIONS: A preference-based scoring algorithm has been estimated for calculating QALYs using OAB-q data. This will permit the cost-effectiveness of new interventions to be assessed in patients with OAB using this questionnaire.

110/Abstract 1583 Modelling the Impact of Metal Disorders on Health State Values Isabel M. Towers, John E. Brazier, Kinley Dorje, School of Health and Related Research, University of Sheffield, Sheffield, UK AIMS: Studies have shown mental disorders to have adverse effects on health-related quality of life, yet there has been little research into quantifying these effects. This study models the impact of mental disorders on UK health state values. METHODS: The Office of National Statistics Psychiatric Morbidity Survey (2000) drew a random sample of 8580 adults from England, Scotland and Wales for interview to estimate prevalence of psychiatric morbidity, use of mental health services, extent of disability, and examine factors associated with mental disorder. Generalized anxiety disorder (GAD), mixed anxiety disorder (MAD), phobias, panic disorder, psychosis, depressive episode, avoidant personality disorder, and paranoid personality disorder were included in assessment procedures, which were conducted by clinical/lay interviews using pre-tested assessment instruments. Respondents completed the SF-12 and provided data on physical health and socio-demographics. Multiple linear regression models examined the impact of mental disorders on SF-6D values after controlling for background variables. RESULTS: Mean SF-6D values were higher when no mental disorder was present across all disorders, even after controlling for socio-demographic factors and physical disorders. Mean differences in scores between presence and absence of disorder were: 0.13 +/0.04 (panic disorder); 0.15 +/- 0.02 (MAD); 0.15 +/- 0.07 (avoidant personality disorder); 0.16 +/- 0.07 (paranoid personality disorder);

112/Abstract 1661 Impact of EQ-5D Dimensions on a Preference-Based Index Score: Results of a Pilot Study Using Best-Worst-Scaling. Oliver H. Günther, Hans-Helmut König, Health Economics Research Unit, University of Leipzig, Leipzig, Germany AIMS: For cost-utility analysis in economic evaluations, preferencebased social index scores usually help calculate quality adjusted life years. One prominent valuation method for deriving these preference weights is the time-trade-off (TTO) technique. Here respondents account for time and quality of life concomitantly in making a hypothetical trade off between living shorter and living healthier concerning a presented set of holistic health state specified by, for

A-105

example, the EQ-5D descriptive system. Derived TTO values are regressed on variables for problems indicated for the EQ-5D dimensions. Thus a set of TTO values results for all possible 243 EQ5D health states, with the state of full health as reference. Little is known empirically about the contribution of each EQ-5D dimension level to the overall utility. In a pilot study we measured the impact of each dimension on overall utility using best-worst-scaling. METHODS: A set of 18 different EQ-5D health states was created by an orthogonal main effect plan. A sample of 75 economics students was presented with the set of health states twice and was asked to choose the 'best' and the 'worst' dimension level of each state. In effect, for each health state the respondents had indicated the pair of dimension levels most distant on their latent utility scale. The data was aggregated over all pairs of best-worst-choices and was analyzed on sample level. A weighted least square regression model estimated the impact of the dimensions and their levels on the utility scale. Regression coefficients were evaluated in the face of short term test-retest reliability. RESULTS: In both best-worst scaling tasks the strongest impact relative to the dimension 'anxiety/depression' was 'usual activities'. The two rank orders of the EQ-5D dimensions were the same. In both tasks 'self care' has the largest range of level scale values compared to the other dimensions. CONCLUSIONS: Bestworst-scaling seems a feasible method to assess the impact of various dimensions of the EQ-5D on overall utility.

114/Abstract 1818 The Impact of Discussing Dealth on Health Preferences: A Qualitative Study in Multi-Etnic Asian Population Nan Luo, Centre for Health Services Research, National University of Singapore, Singapore, Singapore, Hwee-Lin Wee, Rheumatology, Singapore General Hospital, Singapore, Singapore, Shu-Chuen Li, Feng Xie, Xu-hao Zhang, Pharmacy, National University of Singapore, Singapore, Singapore, Yin-Bun Cheung, MRC Tropical Epidemiology Group, London School of Hygiene and Tropical Medicine, London, United Kingdoms, David Machin, Biostatistics Unit, National Cancer Centre Singapore, Singapore, Singapore, KokYong Fong, Julian Thumboo, Rheumatology, Singapore General Hospital, Singapore, Singapore AIMS: To characterize perceptions of and ease in discussing death (EID) and their potential influence on health valuation in a multiethnic Asian population. METHODS: Through in-depth interviews conducted in English or the mother-tongue among adult Chinese, Malay and Indian Singaporeans from both genders and a wide range of age/educational levels, subjects rated using a 0-10 visual analogue scale (VAS): (1) EID; (2) religiosity; (3) acceptability of 8 descriptors for death and (4) appropriateness of _pits and _all-worst in describing the worst possible health state. Subjects also valued 3 health states using VAS followed by time trade-off (TTO). The influence of sociocultural variables on EID was compared using Mann-Whitney or Kruskal-Wallis. Correlations between EID and health preferences were evaluated using Spearman_s correlation. RESULTS: Subjects (n=63, 35% Chinese, 32% Malay, median age 44 years) were generally comfortable with discussing death (median EID (IQR): 8.0 (6.0, 10)). Only education significantly influenced EID. All subjects felt _passed away was most acceptable while _sudden death and _immediate death were least acceptable (median acceptability (IQR): 8.0 (7.0, 10.0) vs. (5.0 (3.0, 7.0) vs. 5.0 (2.0, 8.0)). Subjects clearly preferred _all-worst to _pits (median (IQR): 7.0 (5.0, 9.0)) vs. (4.0 (0, 8.0)). Ethnicity influenced acceptability of _sudden death and _immediate death , appropriateness of _all-worst and 2 of 3 health preferences measured using VAS. EID did not correlate with health utilities. CONCLUSIONS: Singaporeans were generally comfortable with discussing death. Poor correlation between EID and health preferences suggests that the latter are unlikely to be influenced by EID.

113/Abstract 1205 Side-Effects of Localized Prostate Cancer Treatment: Are They Valued Differently by Patients and Healthy Controls? Ida J. Korfage, Harry J. de Koning, J. Dik F. Habbema, MarieLouise Essink-Bot, Public Health, Erasmus MC, University Medical Center Rotterdam, Rotterdam, the Netherlands AIMS: Prostate cancer is increasingly being detected and treated early, and more men are facing the long-term side effects of primary treatment, including urinary, bowel and sexual dysfunction. It has been suggested that patients take these side effects as all in the bargain because they feel they were saved from a life threatening disease. Therefore we asked men who had been treated for localized prostate cancer and experienced permanent side effects (n=54) and healthy controls (n= 53) to value 5 descriptions of health states associated with side effects of treatment for localized prostate cancer. The descriptions were presented without the label cancer to prevent the all in the bargain effect. METHODS: All men valued all descriptions using two valuation methods, visual analogue scaling (VAS) and time trade-off (TTO). We hypothesized that patients would value these health states as better than men without prostate cancer. Current functioning was assessed via EQ-5D completed with items on urinary, bowel and sexual function. RESULTS: Patients and healthy controls showed similar valuations for 9 out of 10 comparisons (5 health states * 2 methods). Valuations in both groups resulted in the same ranking order on the TTO and one switch in rank order on the VAS. Mean VAS valuations ranged from 0.46 for Serious bowel problems and severe erectile dysfunction (ED) to 0.74 for Severe ED. Mean TTO-valuations ranged from 0.61 for Serious bowel problems and severe ED to 0.89 for Some bowel problems and moderate ED. CONCLUSIONS: Patients with permanent side effects after treatment for localized prostate cancer and healthy controls did not differ in their valuations of urinary, bowel or erectile dysfunction which were presented without a cancer context.

115/Abstract 1824 Health-Related Quality of Life of Patients With Hip or Knee Arthrosis on the Waiting List for Prosthesis in Canary Islands, Spain Renata Linertová, Lidia García-Pérez, Lilisbeth Perestelo-Pérez, Service of Evaluation and Planning, Canary Health Care Service, Santa Cruz de Tenerife, Spain, León González-Massieu, Traumatology, USP Hospital La Colina, Santa Cruz de Tenerife, Spain, Armando Brito-Santiago, Traumatology, University Hospital Nuestra Señora de la Candelaria, Santa Cruz de Tenerife, Spain, José Luis Pais-Brito, Traumatology, University Hospital of Canarias, La Laguna, Spain, Marta González, Research Unit, Hospital of Basurto, Bilbao, Spain, Pedro Serrano-Aguilar, Service of Evaluation and Planning, Canary Health Care Service, Santa Cruz de Tenerife, Spain AIMS: To evaluate preliminary results of the health-related quality of life (HRQOL) of patients with hip or knee arthrosis before an arthroplasty, recruited within a research project of evaluation of health results in Canary Islands, Spain. METHODS: Preliminary analysis of baseline results of a prospective and longitudinal study.

A-106

Subjects are patients with a diagnosis of hip or knee arthrosis, on the waiting list for arthroplasty in three hospitals at Canary Islands (Spain). Clinical condition and HRQOL are assessed through the questionnaires WOMAC and EQ-5D. The principal descriptives were estimated, comparisons between men and women within each articulation sample were carried out and scales of questionnaires were correlated. The results were compared with referencepopulation values. RESULTS: Analysed sample (N=300, 65% female) consists of 70% patients with knee arthrosis and 30% patients with hip arthrosis. Average age of the hip sample is 63.81 years, while average age of knee sample is 70.40 years. WOMAC scores in patients with knee arthrosis show a greater pain, stiffness and functional inability than in patients with hip arthrosis, although these differences are not significant. These results correspond with values of reference population with arthrosis from other studies in Spain. WOMAC stiffness scores are significantly worse in women with knee arthrosis than in men. WOMAC stiffness and pain scores are significantly worse in women with hip arthrosis than in men. There is significantly and strongly correlation among WOMAC scores in both knee and hip groups. In the hip arthrosis sample, EQ-5D thermometer and VAS are not correlated, while in the knee arthrosis sample they correlate positively and significantly. CONCLUSIONS: The questionnaire WOMAC differentiates between patients with knee and hip arthrosis. Obtained results coincide with values of reference population with arthrosis from other studies in Spain.

117/Abstract 1671 Measuring Change During Recovery After Hip Fracture: What Does the Global Change Question Tell Us? C. Allyson Jones, Physical Therapy, University of Alberta, Edmonton, AB, Canada, David H. Feeny, Health Utilities Inc, Dundas, ON, Canada AIMS: To compare responses from a 9-point global change(GC) question that involves recall to estimates of change based on real time assessments during recovery after hip fracture. METHODS: In a longitudinal study of patients admitted for treatment of hip fracture, 99 patients who were 65 aged and older were enrolled within 3 to 5 days of surgery. Functional Independence Measure(FIM) and Health Utilities Index(HUI3) measures were done at baseline and at followup:1,3,&6 months. Patients were also asked at follow-up to assess the change in their health (GC). Differences for the FIM and HUI3 for 3 intervals: baseline to 1mon; 1-3mons; and 3-6mons were calculated and categorized into worse, same, or better on the basis of clinically important differences. Kappas were calculated to assess agreement in the classification of patients. RESULTS: The mean age was 78(sd 7.6) with 70% female. The mean Mini Mental State Examination score was 26.8(sd 2.8). Mean changes for the FIM were 28.2(sd 16.0), 6.8(sd 10.0) and .56(sd 7.8); and for HUI3 were .26(sd .24), .09(sd .24) and .02(sd .20) for the 3 intervals. Using the GC question, many patients viewed their health as the stable over the 3 intervals(30-32%), whereas improvements in the FIM and HUI3 were seen for 2 of the 3 intervals. Kappas were poor(0.16 to -0.08). CONCLUSIONS: In this patient cohort the GC question did not reflect the improvements seen over the first 3 months. Use of the GC question may be of limited value in hip fracture, a traumatic injury of elderly persons.

116/Abstract 1409 The Relationship Between Self-Report and Performance-Based Measures of Physical Functioning: Questioning the Content Validity of Self-Reports Caroline B. Terwee, EMGO Institute, VU University Medical Center, Amsterdam, the Netherlands, Rienk M. van der Slikke, Rob C. van Lummel, McRoberts BV, Den Haag, the Netherlands, Rob J. Benink, Orthopaedics, Gemini Hospital, Den Helder, the Netherlands, Wil G. Meijers, Orthopaedics, Atrium Medical Center, Brunssum, the Netherlands, Henrica C. de Vet, EMGO Institute, VU University Medical Center, Amsterdam, the Netherlands

Patients Improved (%) baseline to 1 mon 90 88

FIM HUI3 Global 12 Change

AIMS: To test the hypothesis that self-reported physical functioning is more influenced by pain than performance-based physical functioning. METHODS: 163 knee-osteoarthritis patients completed the performance-based DynaPort® KneeTest (DPKT), WOMAC and SF-36 (self-report measures of pain and physical functioning) before, 3, 6, and 12 months after knee replacement. RESULTS: Correlations between (two) self-reported measures of functioning and (two) pain measures were higher (0.57-0.74) than correlations between the performance-based measure of functioning and the two pain measures (0.20 and 0.26). In factor analysis, WOMAC and SF-36 pain and physical functioning subscores loaded on the first factor (eigenvalue 3.2), while DPKT KneeScore2 loaded on the second factor (eigenvalue 0.92). Before surgery, correlations between performance-based and self-reported physical functioning were higher in patients with less pain (0.43) compared to patients with more pain (0.17), for the WOMAC (as expected), but not for the SF36. After surgery, when the pain had diminished, the correlations between performance-based and self-reported physical functioning were higher, especially for the WOMAC. CONCLUSIONS: Our hypothesis was convincingly supported by the results of the WOMAC, and somewhat less by the results of the SF-36. We consider this as evidence for a lack of content validity of the WOMAC.

1 mon to 3 mon to 3 mon 6 mon 27 5 59 46 11

11

118/Abstract 1456 Cultural Adaptation of the Burn Specific Health ScaleRevised:Version for Brazilian Burn Victims Eneas Ferreira, Unit Burn at University Hospital, School of Medicine of Ribeirão Preto, USP, Ribeirão Preto, São Paulo, Brazil, Rosana Aparecida S. Dantas, Lídia A. Rossi, School of Nursing of Ribeirão Preto, University of São Paulo, Ribeirão Preto, São Paulo, Brazil AIMS: To report the translation and validation process of the Burn Specific Health Scale-Revised (BSHS-R) for Brazilian burn victims. METHODS: BSHS-R has 31 items and six domains (simple functional abilities, affect and body image, interpersonal relations, heat sensitivity, work, and treatment regimens). Each item ranged from 1 to 5 points and the total scale ranged from 31 to 155 points with higher scores indicating the better perceived health state. The adaptation process followed the methodological steps proposed in literature: translation to Portuguese; evaluation by an expert committee; back-translation to English; semantic analysis of scale

A-107

items; pretest of the adapted version and application of the final Portuguese version to burn victims. Internal consistency and construct validity (hypothesis testing) were performed on 82 outpatients aged 18 to 78 years old. These subjects had an average burned body surface area of 19.3%, mixed second and third-degree burns and all participants presented sequelae (64.6% of them with esthetical and functional sequelae). RESULTS: The average BSHSR item score was 3.8 (s.d=0.8) and the total mean score was 119.4 (s.d=24.6). To analyze construct validity and to confirm the hypothesis for convergence between BSHS-R and self-esteem and divergence between BSHS-R and depression was used Pearson´s correlation. Positive and moderate correlations were obtained between self-esteem and affect and body image (r=.64) and between interpersonal relations (r=.60). Moderate correlations were obtained between depression and heat sensitivity (r=-.55) and between interpersonal relations (r=-.67) and high correlation between depression and affect and body image (r=-.79). Internal consistency was good (Cronbach ± 0.74 to 0.94). CONCLUSIONS: These results showed that the adapted version of the BSHS-R into Portuguese (Brazil) fulfill the validity and reliability criteria required for a health state assessment instrument and maintains the properties of the original version.

120/Abstract 1281 Generic Quality of Life in Japanese Patients with Rectal Cancer Hitomi Kataoka, Internal Medicine and Rehabilitation Science, Tohoku University Graduated School of Medicine, Sendai, Miyagi, Japan, Yuko Sasaki, Setsuko Monma, Naoyoshi Minami, Masayuki Kanazawa, Masahiro Kohzuki, Internal medicine and Rehabilitation Science, Tohoku University Graduate School of Medicine, Sendai, Miyagi, Japan, Yoko Goto, Occupational Therapy School of Health Science, Sapporo Medical University, Sapporo, Hokkaido, Japan AIMS: A permanent stoma has shown to be a detrimental effect on perceived QOL in such patients. The aim of this study was to assess differences in perceived QOL over time among Japanese patients treated with stoma and without stoma METHODS: In this prospective study, 36 patients with rectal cancer who had undergone resection in a hospital from October 2003 to October 2005 were evaluated. The generic QOL questionnaire (SF-36) was administered to the patients before surgery, at discharge, 3 and 6 months after the operation. Comparisons were made between patients treated with a permanent stoma (stoma patients) and without stoma (non-stoma patients). RESULTS: Twenty-two patients out of 36 patients answered at 3 and 6 months. Eleven stoma patients (age 60.4yrs) and 11 non-stoma patients (age 60.8yrs) were evaluated. Stoma patients were significant worse QOL scores than non-stoma patients on the role-physical functioning, bodily pain, role-emotional functioning, general health, vitality and mental health. However, no significant differences were found between stoma patients and non-stoma patients at 6 months after the operation. In the different time points for each dimension, scores of stoma patients at discharge were worse than before surgery, however, there were some improvement at 3 and 6 months after surgery on the scores of role-physical functioning, bodily pain and role-emotional functioning. In contrast to stoma patients, non-stoma patients were no improvement of QOL scores at discharge, and scores of role-physical functioning, bodily pain and general health after discharg were significantly worse than before surgery. CONCLUSIONS: Concerning generic QOL of patients who had surgery for rectal cancer, stoma patients were worse than nonstoma patients. However, no significant differences were found between stoma patients and non-stoma patients at 6 months after the operation. Furthermore, stoma patients had some improvement at 3 and 6 months after surgery, however non-stoma patients had no improvement after discharge. The training for stoma patients may lead to a better perception of QOL.

119/Abstract 1007 Validation of the French Canadian Version of the Burn Specific Health Scale Nathalie Dube, Diane Morin, Camille Gagne, Nursing, Laval University, Quebec, Quebec, Canada AIMS: The aim of the study was to translate and adapt the abbreviated Burn Specific Health Scale (BSHS-A) into French Canadian, as well as assessing its reliability and validity. The BSHSA, an 80-items questionnaire, has proven useful in measuring quality of life (QoL) in burn survivors. It includes four domains: physical, mental, social and general health. METHODS: Standard procedures of forward and backward translation were first performed in order to ensure conceptual equivalence. Content validity was verified by 2 panels of experts. One was composed of health professionnels in burn care and the second was of burn survivors. The objective was to verify both clarity and relevance of the items. A pilot study (10 patients) confirmed that the French Canadian version (BSHS-FQ) was understandable and easy to self-administer. Subsequently, 87 adults admitted to the Burn Unit of a large university affiliated hospital in Quebec, Canada between 1997 and 2004, entered the validation study by completing the BSHS-FQ and the QoL measure, the SF-12. The first 40 responders completed a retest at 2 weeks. RESULTS: As concern validity, the overall internal consistency test (Cronbach's alpha=0.98) was very satisfactory, while the test-retest reliability (ICC ranging from 0.22-0.85) showed wide variability. Tests for criterion validity showed expected patterns of association between comparable domains in the BSHS-FQ and the SF-12 (r=0.63). Scores on the BSHS-FQ indicated that the variable "sex" was non significant while "age" although associated with lower QoL, didn't have any association with total burn area (TBA). Furthermore, t test showed that burn thickness (p=0.028) and severity index (thickness and TBA) (p=0.041) was associated with lower scores on the validated scale. CONCLUSIONS: Analyses showed that BSHSFQ is a valid instrument of QoL for French Canadian burn patients. However, more testing is needed to improve reliability. Also more studies need to be conducted in order to test the construct validity before the instrument can be considered as a reliable indicator of burn survivors QoL.

121/Abstract 1297 Norms for the Oral Health Related Quality of Life Questionnaire Gohai in the Japanese General Population Mariko Naito, Preventive Medicine, Nagoya University Graduate School of Medicine, Nagoya, Aichi, Japan, Yoshimi Suzukamo, Physical Medicine and Rehabilitation, Tohoku University Graduate School of Medicine, Sendai, Miyagi, Japan, Shunichi Fukuhara, Epidemiology and HealthCare Research, Kyoto University School of Public Health, Kyoto, Kyoto, Japan AIMS: We have developed and validated the Japanese version of the GOHAI as an oral-health related QOL measure. In this study, we attempted to establish the national norms of the GOHAI in gender/age categories in Japan. METHODS: To select the subject group, stratified multi-stage random sampling was performed at 200 points in the nation. Based on the population in the national census, stratification was performed according to areas and municipalities. As first-stage sampling units, municipalities were sampled for each stratum. Survey points as second-stage sampling units were randomly

A-108

sampled for each stratum. As third-stage sampling units, individuals eligible for the survey were systematically sampled at regular intervals from the Basic Resident Registry in the range of the survey points. A questionnaire survey was performed in 2,400 sampled males and females aged 15-79 years in 2006. The questionnaire included items such as self-rated health, oral health status, oral health behavior, use of dentures, and the present number of teeth in addition to the Japanese version of the GOHAI. RESULTS: Replies were obtained from 1,170 of the 2,400 subjects, of whom 1,169 (569 males and 600 females) were analyzed. Analysis according to characteristics showed no significant differences in the GOHAI score among residential areas, city scales or gender. There were significant differences between the scores by age. The GOHAI score was positively correlated with self-rated health as well as oral health status and negatively correlated with the use of dentures. CONCLUSIONS: The Japanese norms of the GOHAI were established by this survey. In addition, the correlation observed between health status and oral-health related QOL supported the results of previous studies. The results of this study may contribute to the development of QOL studies.

Pediatrics, University of British Columbia, Vancouver, British Columbia, United Kingdom, Stefan Cano, Neurology, University College London, London, England, United Kingdom, Constance M. Chen, Colleen M. McCarthy, Plastic and Reconstructive Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, E D. Collins, Plastic Surgery, Dartmouth-Hitchcock Medical Center, Lebanon, NH, Peter G. Cordeiro, Plastic and Reconstructive Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY AIMS: In cosmetic and reconstructive breast surgery, quality of life is a crucial research topic. Conventional surgical outcomes, focused on morbidity and mortality, remain important but are no longer sufficient on their own. Given this, reliable and valid patient-reported outcome measures (PROMs) are essential. This study examined existing breast surgery-specific questionnaires with the view to recommending future research strategies. METHODS: We performed a systematic literature review to identify PROMs developed and validated for use in cosmetic and reconstructive breast surgery patients. The content and psychometric properties of each measure were evaluated following guidelines proposed by the Scientific Advisory Committee of the Medical Outcomes Trust (2002). RESULTS: We identified 227 PROMs used in breast surgery studies. After excluding 135 generic instruments, 65 ad hoc instruments, 7 oncologic breast surgery instruments, 11 education and knowledge questionnaires, and 2 non-English-language questionnaires, seven measures remained. These seven instruments had been developed for use in the following categories: breast augmentation (n=4); breast reduction (n=1); and breast reconstruction (n=2). Of these, six measures had undergone limited development and validation. Only one measure, the Breast Reduction Symptom Questionnaire, demonstrated adequate development and validation in its target population. Nevertheless, this measure focuses on symptoms and does not include important content domains such as aesthetics and body image. CONCLUSIONS: Reliable, valid and responsive PROMs are lacking in cosmetic and reconstructive breast surgery. In order to show the benefits of aesthetic and reconstructive breast surgery, future research is required to develop breast surgeryspecific PROMs that are conceptually grounded in patient perceptions, span the continuum of impact, and are validated against appropriate psychometric benchmarks.

122/Abstract 1261 Use of the Patient Health Questionnaire in an International Study Nathalie Fernandez, Linguistic Validation, Mapi Research Institut, Lyon, France, Kurt Kroenke, School of Medicine, Indiana University, Indianapolis, Indiana, Christine Getter, Pfizer, New London, CT AIMS: Prior to use in an international study, the 11-item Patient Health Questionnaire (PHQ), underwent linguistic validation in 22 languages. The original scale was developed in US English to facilitate the recognition and diagnosis of the most common mental health disorders in primary care patients. A rigorous methodology was required to ensure conceptual equivalence and cultural relevance across different languages. METHODS: The translation process was conducted by a specialist in each target country using the following standardized methodology: (1) two forward translations by professional translators who were native speakers of the target language and fluent in English; (2) comparison and reconciliation of the translations by the specialist in the target country and the translators; (3) backward translation by a native English speaker; (4) comparison of source and backward version; (5) review by a clinician, (6) review by a psychiatrist and (7) comprehension test on 5 healthy subjects. RESULTS: The translation process revealed linguistic and conceptual challenges: For items containing more than one word to describe a single concept, translations in some languages had less words at their disposal to convey the concept. Also the concept of "more than half the days" when referring to the "last 2 weeks" required the addition of an explanation in some translations. CONCLUSIONS: The 22 language versions of the PHQ were established according to a rigorous standardized translation methodology. The process aims to ensure conceptual equivalence across language versions to facilitate international comparison and pooling of data. The linguistic validation process as a whole supports the integration of international feedback on concepts and wording during the development of questionnaires.

124/Abstract 1503 The Effects of Art on the Quality of Life of HIV Infected Adults: A Qualitative Study in Uganda Barbara N. Wakholi, Social Science, Medical Research Council, Entebbe,, Uganda, Antonieta L. Medina, HIV/AIDS knowledge Programme, Liverpool School of Tropical Medicine, Liverpool, ., United Kingdom, Paula Munderi, HIV/AIDS, Medical Research Council, Entebbe,, Uganda, Charles Gilks, HIV/AIDS, World Health Organisation, Geneva, Switzerland, Heiner Grosskurth, HIV/AIDS, Medical Research Council-Uganda, Entebbe, Uganda AIMS: A high level of adherence is vital for minimizing risk of treatment failure among antiretroviral therapy (ART) recipients. This qualitative research investigates and documents experiences of treatment adherence and perceived changes in the quality of life of Ugandan adults receiving in ART. METHODS: 8 predetermined Focus Group Discussions, 4 with men and 4 with women were conducted with participants of the Development of Antiretroviral Therapy(DART)Trial who had taken ART for a year. RESULTS: Participants reported that myths and misconceptions surrounding the purpose, effectiveness and side effects of ART initially convoluted their decision to seek ART. However, their deteriorating health compelled them to enrol for ART. Drug trust was established through

123/Abstract 1249 Measuring Quality of Life in Breast Surgery: Where are We Now and Where Do We Go From Here? Andrea L. Pusic, Plastic and Reconstructive Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, Anne Klassen,

A-109

the clear dosages, expiry date, and the perception that ART was manufactured professionally and hygienically unlike traditional herbal medicine. ART was reported to reduce disease symptoms and restore their physical strength thus enhancing their mobility, enabling them to resume their usual activities and care for themselves. Most importantly, ART had restored their self-esteem, hope and overall improved their quality of life, relieving them of depression and thoughts of death. Nevertheless, the positive effects of ART were limited by the high pill burden, treatment fatigue, and anxiety generated each time they needed to take the pills for those that had not disclosed their status. In addition, the visible side effects of ART, including change in colour of their nails, portrayed them as ART recipients thus exposing their HIV status and exacerbating stigma, which was described as an emotional encumbrance. CONCLUSIONS: ART improves the overall quality of life; however this positive gain is limited by the drug side-effects, pill burden and the likelihood of sero-status exposure. ART programmes should reinforce adherence through continuous and effective counseling and address stigma at household and community level. These qualitative findings are being tested in a quantitative study.

126/Abstract 1546 Does Presence of Meaningful QoL Data Influence Dermatology Outpatient Consultation? Sam Salek, Asrul Shafie, Welsh School of Pharmacy, Richard Evans, Andrew Freeman, Infectious Diseases Unit, Cardiff University, Cardiff, UK, Humphrey Birley, Mary Browning, Genito-urinary Medicine, Cardiff Royal Infirmary, Cardiff, UK, David Cohen, Health Economic Research Unit, University of Glamorgan, Pontrypridd, UK AIMS: New human immunodeficiency virus (HIV) treatment strategies have transformed the disease into a chronic state with an increased risk of treatment toxicity. It is recognized that improving a patient_s HRQoL should be part of the HIV infection treatment goal. This study aimed to assess the utility of assessing QoL in UK HIV population at routine outpatient monitoring. METHODS: The study was approved by local research ethics committee. All adult patients receiving HIV care in Cardiff were approached in the outpatient clinic and recruited into the study after giving written informed consent. Participants were required to complete the Health Utility Index 3 (HUI3) on recruitment (T0) and post another after three (T1) and six months (T2). Patients who completed at least two questionnaires were included in the analysis. Variables analyzed include HRQoL score at recruitment and each three months interval (HRQoL at T0 vs T1; T1 vs T2; T0 vs T2), and HIV stage (asymptomatic, symptomatic, AIDS) using Wilcoxon Signed Rank test. A separate analysis was also conducted including only patients who completed all questionnaires. RESULTS: In total, 150 completed the baseline questionnaire but only 115 (76.7%) met the inclusion criteria. The patient' mean age was 41.7 years (±10.3 SD) and 95 were male. The patients' HRQoL score demonstrated a decreasing trend from baseline (Median HUI3 T0=0.87935, T1=0.84158, T2=0.79882), p>0.05. Including only patients who completed all three questionnaire revealed that the HRQoL at T1 (Median=0.84158) to be lower than at T0 (Median=0.89032), p<0.05, r=0.2. Stratifying patients according to HIV stage at baseline found that for asymptomatic patients, the HRQoL score was significantly higher at T0 (Median=0.89032) than at T2 (Median=0.79435), p<0.05, r=-0.4. These findings were echoed in symptomatic and AIDS patients but not statistically significant. CONCLUSIONS: This study shows a significant deterioration in patient HRQoL measured at a routine HIV outpatient monitoring interval (3 months). This demonstrates its potential as a patient-reported outcome measure in routine clinical practice.

125/Abstract 1708 A Review of Patient Reported Outcomes Instruments in HIV/AIDS Diana Rofail, Katie Allan, Questionnaire Development and Validation, Mapi Values Ltd, Bollington, Cheshire, United Kingdom, Dominique Dubois, Katrien Peeters, Pharmaceuticals Services, Johnson & Johnson, Mechelen, Belgium AIMS: HIV/AIDS and its treatment have a great impact on HealthRelated Quality of Life (HRQoL). This study aims to identify the most appropriate instruments to measure HRQoL for future clinical trials and surveys. METHODS: A search of studies published between 2003 and 2006 was conducted using MEDLINE and EMBASE. Studies were selected if Patient Reported Outcomes (PRO) data were described in the title or abstract and original data was included. The PRO instruments identified were reviewed in terms of their domain coverage, psychometric properties, and practical assessment features (e.g. number of items, and language availability). RESULTS: The search strategy produced 412 articles, 130 met all of the inclusion criteria and were retrieved for review. A total of 26 HRQoL instruments were identified (15 generic and 11 HIV/AIDS-specific). No instruments covered all core domains and issues directly relevant to current treatment options. Amongst the generic measures the Short-Form 36 (SF-36) and EuroQoL (EQ-5D) provide utility and normative data for comparisons. Regarding HIVspecific instruments the Medical Outcome Study HIV (MOS-HIV), AIDS Clinical Trials Group Quality of Life (ACTG-QOL) and the Functional Assessment of HIV Infection (FAHI) comprehensively capture the domains relevant to patients living with HIV/AIDS and demonstrate robust psychometric properties. The assessment features for the recommended tools are good with most having a reasonable number of items and therefore are less burdensome to patients to complete, and available in several languages. CONCLUSIONS: No single measure assesses all aspects of HRQoL that are important to HIV/AIDS patients. Consequently, it is recommended that a battery of instruments be implemented in any HIV/AIDS study. The SF-36 or EQ-5D are recommended as generic tools for comparison to other populations, and either the MOS-HIV, ACTG-QoL or the FAHI as HIV/AIDS-specific tools. The most appropriate instruments will be dependent on a number of aspects including the research question, trial design, stage of disease, and target message as it pertains to PRO.

127/Abstract 1000 Multivariate Analysis of Influent Factors on Quality of Life of AIDS Patients Jing Xie, Xue Zhong Shi, Health Statistics, Public Health, Zhengzhou, Henan Province, China AIMS: To screen the important influent factors of QOL of AIDS patients under Chinese culture background and provide clues for improving QOL and policy-make. METHODS: Multi-stage sampling was taken to select 64 AIDS patients who were assigned to fill in the WHOQOL-120 HIV questionnaire. RESULTS: Male (12.36±1.82)have better QOL than that of female (11.26±2.21) in Psychological Domain (t=2.14, p = 0.036); married persons (13.47±1.85) better than those windowed ones (11.66±2.95 ) in Social Relationship Domain (t=2.37, p=0.021); higher income persons have better QOL in Social Relationship Domain (F=4.98, p=0.010), which are 12.10±2.27, 13.13±1.91, and 14.13±1.92 respectively for group A (annual incomed1000 ¥ ), group B (¥1000)

A-110

CONCLUSIONS: Conclusion Annual income, marital status, gender, and numbers of AIDS related symptoms do affect QOL of AIDS patients, in which AIDS related symptoms is the most important factor.

R), Coping (Brief-COPE). RESULTS: Results: Controlling time 1 (T1) Affect, structural equation modeling (Amos 5.0) indicates that Optimism (T1) predicts more Positive Affect at time 2 (T2). The use of Evasive Coping at T1 leads to higher levels of Negative Affect (T2). The use of Active Coping at T1 leads to higher positive affect at T2. Evasive and Active Coping partially mediates the relation between optimism and Affect. CONCLUSIONS: This study provides empirical information on how Coping mediate between Optimism and Affect. Attention should be given to initiatives which aim to improve Optimism and active coping in college students.

128/Abstract 1263 Subjective Wellbeing and Quality of Life Among Persons With Sickle Cell Disease Kingsley O. Akhigbe, Mental Health, Faustina K. Idu, Optometry, University of Benin, Benin City, Edo, NIGERIA

130/Abstract 1391 AIMS: Sickle cell disease (SCD) is a heterogeneous disorder, characterized by clinical manifestations including chronic haemolysis, an increased susceptibility to infections and vasoocclusive complications often requiring medical care.It is the most common inherited disease in Nigeria, and Sub-Saharan Africa. While the hematological and clinical profiles of these patients have been extensively studied, less is known about the patients satisfaction with life as well as their quality of life. Quality of life studies have become an essential complement to medical evaluation The overall aim of the study was to describe the quality of life of Sickle cell patients and compare it to that of non-sicklers. METHODS: The study design was a two-sample comparison at one point in time. 150 patients attending the Sickle Cell Clinic at a Teaching Hospital in Nigeria was the study population while 150 nursing students who are non-sicklers constituted the control group. The survey instruments used were the Medical Outcomes Study 36-item Short Form (SF-36) and the Satisfaction with Life scale (SWLS). Correlation and multiple regression analyses were calculated for the relation between subjective wellbeing and QOL scores, and compared scores using ttests. RESULTS: The study sample consisted of 92 males and 58 females aged between 16 and 39 years. Most of them were single and lived with their relatives, had low educational level and had low income. SF-36 scores showed wide variability with an acceptable internal consistency of Cronbach's alpha >0.70. Non-sicklers reported more satisfaction with all the domains of life than the sicklers. Expectedly, non-sicklers reported greater satisfaction with life than the sicklers. DICUSSION: The psychosocial burden of SCD on sufferers in a nation where there is no national social welfare provision is discussed. CONCLUSIONS: Our findings indicate there exist poor mental and psychosocial functioning among Sickle cell patients in Nigeria.QOL assessment should be an essential component of the medical evaluation of SCD patients.

An investigation into the value of symptom and health related quality of life data as prognostic factors in newly diagnosed anaplastic oligodendrogliomas patients M Mauer, QLU, EORTC, Brussels, Belgium, MJB Taphoorn, Dept of Neurology/Neuro-Oncology, Medisch Centrum Haaglanden Westeinde, Den Haag, The Netherlands, A Bottomley, F Efficace, QLU, EORTC Data Center, Brussel, Belgium, J-Y Delattre, Neurologie - Hopital B, CHU Pitie-Salpetrière, Paris, France, AA Brandes, Neurology, Ospedale Bellaria, Bologna, Italy, CCD Van der Rijt, Neurology, Erasmus University Medical Center, Rotterdam, The Netherlands, HJJA Bernsen, Neurology, Canisius Wilhemina Ziekenhuis, Nijmegen, The Netherlands, M Frenay, Serv. de Medecine, Centre Antoine Lacassagne, Nice, France, CC Tijssen, Neurology, St. Elisabethziekenhuis, Tilburg, The Netherlands, C Coens, QLU, D Lacombe, Medical Affairs, A Allgeier, Brain cancer unit, EORTC Data Center, Brussel, Belgium, MJ Van den Bent, Neurology, Erasmus University Medical Center, Rotterdam, The Netherlands AIMS: Few studies have explored the prognostic value for survival of health-related quality of life (HRQOL) parameters in anaplastic oligodendrogliomas patients. We investigated this using data from a recently conducted randomized controlled trial (RCT) of radiotherapy with adjuvant chemotherapy or radiotherapy alone in newly diagnosed oligodendrogliomas patients. METHODS: 368 newlydiagnosed oligodendrogliomas patients were randomized to receive radiotherapy alone or radiotherapy plus chemotherapy (Van den Bent ASCO 2005). HRQOL data was collected using the EORTC QLQC30 and Brain Cancer Module. Baseline scores were examined for the relationship with length of survival. Differences between survival curves were assessed using the log-rank test. The Cox proportional hazards regression model was used for the multivariate analyses. Treatment type was included as a fixed covariate in the starting model for the multivariate analyses, along with clinical factors of gender, previous resection for low grade oligodendroglioma, age, type of surgery, tumor location, performance status, 1p & 19q LOH, endothelial abnormalities, and necrosis. RESULTS: The prognostic factor analysis was conducted on baseline questionnaires available in 79% (n =288) patients. Univariate analyses and the final multivariate analysis revealed all the clinical variables predicted survival (except gender and treatment). Ten HRQOL factors were significant in the univariate analyses, but no major HRQOL scales independently contributed in predicting survival in the multivariate model. Further analysis is still ongoing to confirm these results. CONCLUSIONS: While initial analysis suggests that the major HRQOL scales do not add significant prognostic information to the clinical variables for predicting survival, additional ongoing analysis will be needed to confirm these findings

129/Abstract 1054 Optimism and Coping as a Predictors of Affect Over Time: Test of a Mediator Model Vítor Bertoquini, José Pais Ribeiro, Psychology, Faculdade de Psicologia e de Ciências da Educação UP, Porto, Portugal AIMS: Research suggests that Coping mediates the relationship between personality resources and immediate and long term effects outcomes. Coping is conceived to be an indirect path through which optimism may influence Affect. In contrast, Optimism has a positive influence on positive affect and a negative effect on negative affect. This study investigates the impact of Optimism on Affect and the mediational role of coping in the optimism affect relationship over a 19 week period. METHODS: The data (N=227) were obtained as part of an ongoing longitudinal study on the psychological determinants of SWB. The sample consists of college students aged between 17 and 42 (M=23 DP=3.77). Participants completed measures of negative and positive affect (PANAS), Optimism (LOT-

A-111

for a chronic condition and 50% of the mothers had to be under mental health treatment. A battery of instruments was included regarding HRQL, habits and risk behaviours, mental health, life events, health care services use, and academic achievement. Subjects were requested to fill the questionnaire at home. Mothers were interviewed by phone (CIDI-SF). Focus groups and in-depth interviews were carried out. Interviews were recorded and content analysis of interviews was performed RESULTS: Twenty out of 24 subjects were interviewed. Average time to complete the questionnaires was 40 minutes. Most of the questionnaire's sections were easily understood. The questionnaire seemed to capture changes in habits and major life events in those adolescents diagnosed with chronic conditions, as well as 3 years retrospective assessment. CIDI phone interviewing was feasible and administered in an average time of 20 minutes CONCLUSIONS: The questionnaire adequately collects information on factors with potential influence on HRQL. The CIDI-SF phone interview is a feasible way to assess mothers' mental health. Funding: Instituto de Salud Carlos III grants: PI042504, PI042315, and Network of Excellence IRYSS G03/202

131/Abstract 1658 Use of Pro Instruments: Licensing and Copyright Issues C Anfray, MP Emery, Lyon, France AIMS: The exponential development of Patient-Reported Outcome (PRO) measures in clinical research and their increasing use in international studies have led to question about their accessibility and licensing conditions. The recent release of the draft FDA guidance on PROs acknowledging the possibility of modifying instruments, and the emergence of item banking specific-projects make it even more necessary to establish copyright protection not only in the original works but in all their derivative versions e.g. translations, e-versions. In addition, to further the science of PROs and share their work, most authors are keen to have their instruments released in the public domain. However, this may be detrimental to the instrument itself as there is no control on its use and dissemination. The present work will review the methods, concepts and processes of the international distribution of PRO instruments, from a legal perspective and practical aspects. METHODS: To optimize the use of a questionnaire there is a need to (1) facilitate access to it; (2) get reliable and useful information about it (e.g. user manual, availability and validity of translations); (3) have this information updated. To respond these conditions, the establishment of a centralized distribution process is essential, especially at an international level. RESULTS: The distribution rules -which establish the conditions of accessing, using, and translating the questionnaires -should be determined with the authors according to their rights which come under the domain of intellectual property. The various aspects of authorship should be considered at all stages of the instrument's life: development, publication, translation, adaptation, taking into account the different parties involved: authors, translators, and sponsors. CONCLUSIONS: The ambiguity of international laws regarding copyright and the difficulty in interpreting these in the specific context of PROs have led to develop recommendations and procedures that help authors to protect their rights while allowing a wide and safe use of their instruments.

133/Abstract 1596 Factor Mixture Modeling and the Measurement Structure of the Multidimensional Students' Life Satisfaction Scale Richard G. Sawatzky, Pamela A. Ratner, Nursing, Bruno D. Zumbo, ECPS, Measurement, Evaluation, & Research Methodolog, University British Columbia, Vancouver, British Columbia, Canada AIMS: Huebner's (2001) Multidimensional Students' Life Satisfaction Scale (MSLSS) putatively measures five correlated latent factors that represent satisfaction with family, friends, school, living environment, and self. Although the measurement model has been supported by several published confirmatory factor analyses, we were unable to replicate it in a large sample of adolescents. We therefore used factor mixture modeling to examine whether the lack of model fit could be explained by heterogeneity in the sample. METHODS: The data were taken from a cross-sectional health survey that was completed by 8225 adolescents in British Columbia, Canada. Factor mixture models were specified with the thresholds of the observed ordinal variables and the factor loadings free to vary across latent classes. Global fit statistics (e.g., BIC and AIC), thresholds and factor loadings were compared for models with two to five latent classes. RESULTS: Acceptable model fit is obtained for each of the subscales if four latent classes are specified. There were substantial differences in the thresholds of the ordinal variables for each of the four latent classes. We used multinomial regression to examine whether predicted class membership could be explained by demographic variables (gender, ethnicity, grade), variables pertaining to social comparisons (perceived school performance compared with other students), and healthrelated variables (perceived mental and physical health, and depression). Class membership was not explained by these variables. CONCLUSIONS: Our study shows that adolescents do not respond consistently to the items of the MSLSS and that comparison of life satisfaction scores may therefore not be warranted. Although factor mixture modeling provides a powerful approach for examining samples that are heterogeneous with reference to a specified measurement model, we caution that the latent classes need to be carefully examined before assigning substantive significance to predicted class membership.

132/Abstract 1434 Pilot Test of the Spanish Kidscreen follow-Up Study J Palacio-Vieira, Catalan Agency Health Technology Assessment Research, Barcelona, Spain, Villalonga-Olives E, Health Services Research Unit, Institut Municipal d_Investigació Mèdica (IMIMIMAS), Barcelona, Spain, Berra S, Solans M, Catalan Agency Health Technology Assessment Research, Barcelona, Spain, Codina F, Primary care centre, Lloret de Mar, Spain, Alonso J, Health Services Research Unit, Institut Municipal d_Investigació Mèdica (IMIMIMAS), Barcelona, Spain, Rajmil L, Catalan Agency Health Technology Assessment Research, Barcelona, Spain AIMS: Few longitudinal studies have analyzed the influence of habits, life events and mothers' mental health on HRQL during adolescence. The KIDSCREEN project developed a questionnaire on HRQL collecting information from a representative sample of Spanish 8-18 year olds. The sample will be re-interviewed in 3 years (2006). The aims of the pilot test were to assess understandability and feasibility of the new sections of the questionnaires, methods of collecting information on changes that have happened in the last 3 years, and phone administration of the Composite International Diagnostic Interview Short Form (CIDI-SF) to measure mothers' mental health METHODS: A convenience sample of 12 pairs (adolescents and mothers) was recruited from the records of the Primary Care Centre in Lloret, Spain. Adolescents included were males and females from 11 to 21 years old; 50% had to be diagnosed

134/Abstract 1632 Quality of Life and Treatment of Obesity in Children and Adolescents. First Results From a Nationwide Prospective Study (EVAKuJ-Project) Monika Bullinger, Medical Psychology, University of Hamburg,

A-112

Hamburg, Germany, Ulrike Hoffmeister, Epidemiology, University of Ulm, Ulm, Germany, Ulrike Ravens-Sieberer, Child & Adolescent Health, Robert Koch Institute, Berlin, Germany

political actors can better define more effective social policies. In future studies the samples should be bigger and the scales should be applied in other cities.

AIMS: Obesity in childhood and adolescence is an urgent health problem, also in Germany. The existence of this _new epidemic is well known from epidemiological studies and has been met by a variety of treatment options tailored towards obese children (obesity defined as a BMI percentile score of 97 or above). METHODS: In a multicentre government funded prospective study (EVAKuJ-Project), 492 treatment providers were included from a representative survey on obesity treatment. Inclusion criteria were compliance of the intervention programmes with quality criteria such as multicomponent approach, number of patients treated, size and existence of formal documentation. Interventions range from comprehensive 3 week inpatient programmes in specialized pediatric rehabilitation clinics to outpatient courses offered by health professionals. 137 institutions consented to collecting medical, behavioural and psychosocial data from a consecutive number of attendees before and after the respective intervention as well as at one year follow-up. Among the outcome criteria, the KINDL was used to assess quality of life from patients and parents perspective. RESULTS: At present 1000 children and 1000 parents have been recruited, and respective data are available for the first measure point. Focus of the statistical analyses presented is the description of quality of life in this sample as compared to German population based reference data, the relationship between quality of life and medical, sociodemographic and psychosocial data as well as hypotheses related to the longitudinal study. CONCLUSIONS: This is the first nationwide German study addressing rehabilitation outcomes in obese youth and their families. Important is the representative approach to the identification of care-giving institutions as well as the longitudinal data collection from self-reports, proxy-report and medical documentation. The results are deemed to be politically relevant for the German health care system as well as for other countries.

136/Abstract 1473 Monitoring Health Related Quality of Life (HRQOL) in Adolescents with Diabetes. Baseline Data from a Rancomised Controlled Crossover Multicentre Study (RCT:ISRCTN65138334) Maartje de Wit, Medical Psychology, Henriette A. Delemarre-van de Waal, Reinoud J. Gemke, Pediatrics, Frank J. Snoek, Medical Psychology, VU University medical center, Amsterdam, Netherlands AIMS: Systematic monitoring of HRQoL in adolescents with type 1 diabetes may help to improve psychosocial adaptation and glycemic control. In an ongoing RCT (2005-2007), we investigate the effect of monitoring HRQoL on psychosocial adaptation, HbA1c and satisfaction with care in adolescents with type 1 diabetes. At baseline adolescents and parents underwent a comprehensive psychosocial assessment, presented here. METHODS: Four outpatient pediatric diabetes clinics in the Netherlands (n=90) were randomized over CAU and monitoring condition: 3-monthly computerised HRQoL assessments by the PedsQL generic and diabetes module, and discussion of outcomes with pediatrician or nurse. Demographic, medical and psychosocial data were gathered at baseline. Patients completed the CHQ-CF87, CES-D, Diabetes-related Family Conflict Scale (DFCS) and the WHO-5. Parents completed the CHQ-PF50, CES-D, DFCS and the WHO-5. RESULTS: Mean age is 14.9 years (±1.09), 46 boys, mean HbA1c 8.75% (±1,65). On the CES-D, 3,3% of the patients scored above 23, indicating moderate to severe depression. Compared to healthy controls, patients scored significantly lower on CHQ sub-scales Role/Social-Physical Limitations (p=.005), Behaviour (p<.001) and General Health (p<.001). Adolescents rated their Behaviour more positive then parents (p<.001). Compared to boys, girls reported more Bodily Pain (p=.04). Less Diabetes-related Family Conflict is associated with better scores on Psychosocial (p=.001) and Physical Health subscales (p=.05) and less Depressive symptoms(p=.02). Living in a one-parent family, not being Caucasian and reporting lower Psychosocial Health are associated with higher HbA1c values (p<.001). CONCLUSIONS: Preliminary results from this ongoing trial are in line with literature, pointing to the importance of psychosocial factors in diabetes management, in particular family conflict. The study will show if monitoring HRQoL periodically will contribute to improved psychosocial functioning and diabetes control.

135/Abstract 1903 Portuguese and Spanish Children Quality of Life Marta I. Calado, Psychology, Escola Superior de Educação João de Deus, Lisboa, Portugal AIMS: The research question is: _What is the perception that children (age 9 to 17) have about their Quality of Life (QoL)? The purpose of the study is to assess and adapt Huebner's scales (1994) SLSS (Student_s Life Satisfaction Scale) and MSLSS (Multidimensional Student_s Life Satisfaction Scale) - to the Portuguese population. METHODS: The research uses a sample of 277 Lisbon students. The applied instruments are SLSS and MSLSS. Data comparison with the Spanish study (Casas, F. and Alsinet, C.; 2001). The statistical techniques are descriptive statistics, correlations and Cronbach_s Alpha. RESULTS: The results demonstrated that QoL perception is better in boys than in girls. The study also pointed out the same results for younger kids. The results demonstrated good reliability, a consistent factor structure, a good functioning of the scale in the Portuguese sample, and a pattern of answers consistent with the Spanish (Casas, F. and Alsinet, C.; 2001) and the original results by Huebner (1991 and 1994). CONCLUSIONS: QoL perception in these scales is measured by the psychological wellbeing. The study demonstrated higher results of psychological wellbeing for both Portuguese and Spanish children. This is an excellent social indicator of the health of these children. This psychological well-being was assessed in five domains: school, family, friends, self and environment. Throughout the use of these results social and

137/Abstract 1529 Telehomecare Maintains Quality of Life for a CommunityDwelling Adults with Chronic Illness: Preliminary Research Michael Erhart, Child Adolescent Health, Robert Koch-Institut, Berlin, Germany, Stephane Robitail, Pascal Auquier, Public Health, University Hospital of Marseille, Marseille, France, Mick Power, Psychiatry, University of Edinburgh, Edinburgh, United Kingdom, Ulrike Ravens-Sieberer, Child and Adolescent Health, Robert KochInstitut, Berlin, Germany AIMS: Currently available IRT techniques encompasses various models, differing e.g. with regards to the number of parameters used to model the actual item response behaviour of the respondents. The choice of IRT models should take into account the nature of the testdata and test-theoretical arguments have to be considered. For the adolescents KIDSCREEN-27 Psychological Well-being dimension

A-113

(PWB) we compared the trait person parameter estimates (PP) issued from 5 different IRT models with regards to their construct validity. METHODS: The field study of the KIDSCREEN HRQoL project resulted in a data set of 22000 children and adolescents (8-18 y.) from 13 European countries who answered the items of the KIDSCREEN-27 PWB. Person parameters were estimated with the nonparametric TESTGRAF approach; the generalized Partial Credit Model (GPCM); the Partial Credit Model (PCM); the Rating Scale Model (RSM) and the Multidimensional Rasch model (MULTIRA) which uses outer-scale information for scoring. The construct validity of the measurement was assessed with Pearson- and multiplecorrelation between the PP-values and measures of physical- & psychological health-status (GH-item, HBSC-SCL, SDQ) and social HRQoL determinants (OSSS). RESULTS: All IRT models fitted the test data well. The PP-values estimated with the GPCM displayed the highest construct validity (r=.16-.54) whilst those gained from MULTIRA displayed the lowest (r=.15-.51). The proportion of score variance determined by the HRQoL determinants was highest for the GPCM and the RSM (R2=.41/.40), slightly lower for the PCM and TESTGRAF (R2=.39) and lowest for the MULTIRA (R2=.37). All differences between the coefficients were statistically significant. CONCLUSIONS: A consistent pattern of differences in the construct validity of different IRT-scorings was revealed. The complexity of IRT models seems to be only one aspect relevant for an adequate account of the test-data. As the actual differences are rather small, the choice of IRT models should be based on a scientific rationale rather.

domain had the most robust psychometric characteristics and was most able to discriminate the difference of quality of life in physical capacity for COPD patients.

139/Abstract 1411 Strategies for Developing an Item Pool for a Computerized Adaptive Testing (CAT) Version of EORTC QLQ-C-30 Morten A. Petersen, Palliative Medicine, Bispebjerg Hospital, Copenhagen, Denmark, Stein Kaasa, Palliative Medicine Unit, Jorunn L. Helbostad, Dept. of Neuroscience, Faculty of Medicine, St. Olavs University Hospital, Trondheim, Norway, Mogens Groenvold, Palliative Medicine, Bispebjerg Hospital, Copenhagen, Denmark AIMS: The EORTC Quality of Life Group has initiated a project to develop CAT for the fifteen subscales of the EORTC QLQ-C30. The aim is a CAT version that is in agreement with the QLQ-C30. Particularly, this means that the CAT version shall measure the same aspects of the HRQL dimensions as the QLQ-C30, but with higher precision, and will have the same response format. Here we describe the strategy for the item pool development with focus on the special considerations when developing a CAT version of an existing multidimensional HRQL instrument. METHODS: The development and selection of candidate items for our item pools consist of the following steps. First, a literature review to clarify what aspects of the HRQL dimensions the QLQ-C30 items measure and to identify items used/developed to measure the HRQL dimensions. Second, items are rephrased to fit the 'QLQ-C30 format'. Third, the list of rephrased items is presented to experts. Fourth, the possibly revised list of items is presented to patients. As an example, we present the development of the physical functioning (PF) item pool. RESULTS: The PF item pool was based on the literature review by the Norwegian PAT-C group supplemented with PF items from questionnaires listed in the PROQOLID database. A total of 970 PF items were identified. Items were classified according to the WHO International Classification of Functioning, Disability and Health. Approximately 100 items were classified as belonging to a PF subdimension measured by a QLQ-C30 PF item and could be reformulated to fit the QLQ-C30 format. CONCLUSIONS: Requiring that the CAT version shall be in agreement with the existing QLQ-C30 affects the development process but insures backward-compatibility, measurement within a meaningful and known conceptual framework, and a simple and user-friendly format. Researchers developing CAT based on existing instrument should take these advantages and disadvantages into consideration when selecting their strategy.

138/Abstract 1487 A Rasch Analysis of the World Health Organization Quality of Life (WHOQOL) Assessment Instrument in Patients with Chronic Obstructive Pulmonary Disease Yu-Jun Chang, Wen-Miin Liang, Yi-Chun Yeh, Institute of Environmental Medicine, China Medical University, Taichung, Taiwan, R.O.C. AIMS: This study examined the World Health Organization Quality of Life (WHOQOL)-Brief Taiwan version to determine psychometric properties for quality of life outcome measure in patients with chronic obstructive pulmonary disease (COPD). METHODS: A total of 146 COPD patients aged 50 to 88 years were recruited at a university hospital in Taiwan. Rating scale model based on item response theory was used to evaluate 3 psychometric characteristics of the WHOQOL-BREF(TW): unidimensionality, targeting, and item difficulty. In the calibration, the average difficulty level across items in each category was fixed at zero. RESULTS: Five items misfit the constructs. "Pain" and "Medical dependency" misfit the physical domain; "Negative feelings" misfit the psychological domain; "Satisfaction with sex life" misfit the social relationship domain; "Opportunity for leisure activities" misfit the social relationship domain. Personal reliabilities of the four domains were all greater than 0.7. As evidenced by average measure approaching zero (0.31), the physical domain was well targeted to the sample and had a wide range of items that capture difficulties in most COPD patients. However, the psychological, social relationship, and environmental domains with high average measures (0.79, 1.67 and 1.18, respectively), which indicated the items were relatively easy and were not sensitive to detect the traits for a group of COPD patients very impaired in these domains. Item difficulty of the four domains of physical, psychological, social relationship and environmental, were ranged from -0.47 to 1.68, -1.11 to 0.85, -0.58 to 0.65 and -1.43 to 0.62, respectively. CONCLUSIONS: Rasch analysis further established the validity of the WHOQOL-BREF(TW). The personal reliabilities of the four domains were acceptable. The physical

140/Abstract 1393 An IRT Analysis Of St. George's Respiratory Questionnaire In Patients With Chronic Obstructive Pulmonary Disease Tai-Chin Wu, Wen-Miin Liang, Ya-Wei Shiao, Institute of Environmental Medicine, China Medical University, Taichung, Taiwan, R.O.C., Jian-Jung Chen, Chinese Medicine, China Medical University Hospital, Taichung, Taiwan, R.O.C. AIMS: Using item response theory (IRT) analysis to assess multiple psychometric characteristics of the St. George's Respiratory Questionnaire (SGRQ) in patients with chronic obstructive pulmonary disease (COPD). METHODS: The study participants consisted of 147 patients with COPD. All patients completed the SGRQ questionnaire. The partial credit model of IRT analysis by using WINSTEPS program was used to assess unidimensionality, targeting property, item difficulty, and separation property.

A-114

Research Europe, Barcelona, Barcelona, Spain, Alfredo Avellaneda Fernández, Instituto de Investigación de Enfermedades Raras, Instituto de Salud Carlos III, Madrid, Madrid, Spain, Maravillas Izquierdo Martínez, Servicio de Análisis e Intervención en Salud Pública, D.Gral. Salud Pública y Alimentación (C.A.Madrid), Madrid, Madrid, Spain, José R. Ramón, Instituto de Investigación de Enfermedades Raras, Instituto de Salud Carlos III, Madrid, Madrid, Spain

RESULTS: Within each domain, most of the items measured a single construct. Only one item in SGRQ symptom domain and one itme in SGRQ impact domain misfit the constructs. Items in symptom and activity domains were well targeted to the sample; however the impact items were easier for the sample. The impact and activity domains had a wide range of items (-2.18 to 3.55 and -5.18 to 4.03, respectively) that capture difficulties that most individuals with COPD experience, while symptom domain included items that capture limitations in COPD patients (-0.55 to 0.55). The order of items from less to more difficult was clinically meaningful. The activity domain had a higher separation index (2.62), but lower lower separation index in symptom (1.36) and impact (1.40) domains. The nonlinear, S-shaped, relationship between the raw score and logit unit latent trait showed the IRT scoring method provides a more realistic functional status measure. The correlation coefficient between SGRQ domain (symptom, activity, and impact domain) scores and the IRT latent trait calibrations was 0.90, 0.99, and 0.95, respectively, which indicated they were highly correlated. CONCLUSIONS: IRT analysis showed the domains were unidimensional, the items had an excellent range of difficulty in impact and activity domains. The IRT scaling provides a more precise and reliable measure.

AIMS: Rare Diseases (RD) are a group of heterogeneous, chronic and disabling illnesses with scant curative treatment options affecting less than 5/10.000 inhabitants. RD are responsible for a spoil of patients_ health-related quality of life (HRQoL). There are no studies in Spain assessing HRQoL in RD. The study aim was to assess HRQoL in RD patients in Spain and the accessibility and patients_ satisfaction with medical and social services. METHODS: Crosssectional study using a postal survey carried out during 2005. The survey was send to Spanish patients older than 14 years with RD through 24 associations around Spain. Socialdemographic, clinical, HRQoL and patients_ satisfaction data were collected. HRQoL was measured with the EQ-5D questionnaire. Accessibility and patients_ satisfaction with medical and social services were assessed with a questionnaire according to a Likert scale. A descriptive statistical analysis was performed. Differences were analysed with chi2 and t tests using SPSS 9.0. RESULTS: 530 patients with RD were included. 52.7% were female. Mean(SD) age was 39.0(14.9) years. The length of time of RD was 13.3(10.4) years. 40% followed a drug treatment . The mean (95%CI) VAS score for health status for RD patients was 50.4(2). The highest percentage of reported problems were related to usual activities (85.6%) and mobility (82.7%). Statistically significant difference of 14.5 points (p<0.001) was observed among men and women in pain/discomfort. 46% of patients considered physician was accessible and 52.8% were satisfied with their medical service. Related to social worker_s service, 39.9% considered accessible and 41.7% were satisfied with it. CONCLUSIONS: HRQoL in RD patients is poor. Usual activities and mobility were the dimensions the most frequently declared as ones with problems. There is a low proportion of patients satisfied with medical or social worker services. These results may generate evidences for public administrations in terms of decision-making and obtaining resources to tackle RD.

141/Abstract 1471 Differential Item Functioning Using RASCH Analysis John Brodersen, Hanne Thorsen, General Practice, University of Copenhagen, Copenhagen, Denmark AIMS: The aim of the study was to test items for Differential Item Functioning (DIF) in a condition-specific measure for women having an abnormal screening mammography. METHODS: The draft version of a 33-item condition-specific measure was completed by 184 women who had received an abnormal screening mammography. Two weeks later the same women completed the questionnaire again having either a known falsepositive result, a diagnosed breast cancer or still undiagnosed. The measure was also completed by 240 women with a normal screening result. Data were analysed using the Rasch model. DIF relative to person covariates was checked by analyses of variance examining the degree to which individual residuals for specific items depended on the covariates. Absence of evidence of interaction between the covariates and the estimated trait parameters were taken as evidence of DIF being uniform. If uniform DIF was identified mean scores of the dimension were estimated from the means of the person locations on the latent trait according DIF found. RESULTS: One of six items in a dimension measuring "sense of dejection" had uniform DIF in two of four subgroups. After deleting this item the five remaining items still fitted the Rasch model with no DIF. In a 2-item dimension describing the impact on breast examination both items had uniform DIF. The item "examined my breasts" had uniform DIF in the subgroup diagnosed with breast cancer compared to the remaining subgroups. The item "examined my breasts in the mirror" had uniform DIF in the subgroup with normal screening mammography versus the remaining subgroups. There was difference in the mean of the subscales' raw score and the estimated scores calculated according to identified DIF. CONCLUSIONS: If an item functions differently in subpopulations and it is decided to keep the item in a subscale special precautions should be taken when calculating scores of this dimension. Otherwise, the difference between subpopulations could be under or overestimated.

143/Abstract 1325 A New Health Educational System for Preventing a LifestyleRelated Disease in Japan Toshiko Yoshida, Graduate School of Nursing, Miyagi University, Kurokawa-gun, Miyagi, Japan, Atsushi Togashi, Yuka Ohsuka, Hiroko Tadaura, Miyagi University, Kurokawagun, Miyagi, Japan, Goichi Itabashi, Sendai Foundation for Applled Information Sciences, Sendai, Miyagi, Japan, Ichiro Hourai, Syoichi Noguchi, Sendai Foundation for Applied Information Sciences, Sendai, Miyagi, Japan, Noriko Sato, Miyagi University, Kurokawagun, Miyagi, Japan AIMS: For the last few decades, patients with a lifestyle-related disease have increased in Japan. It is necessary to develop educational system not only for patients with a lifestyle-related disease but also for healthy people as primary prevention for the preliminary stage of process. However it is difficult for people to keep their motivation for modifying their lifestyle and also to have the learning opportunities after discharging hospital. In our study, we constracted and evaluated a new health educational system for preventing a lifestyle-related disease using Internet educational programs in Japan. METHODS: We provided the system consisted of exercise and conuseling course, and a health educational program

142/Abstract 1717 Assessing Quality of Life of Rare Disease Patients in Spain. ERES Study. (ERES: Acronym of Social-Sanitary Study of Rare Diseases) Xavier Badia, Miquel Layola, Pérez Pilar, Health Outcomes

A-115

Fear, QOL Overall) were responsive (p<.05). CONCLUSIONS: Early findings demonstrate acceptable reliability and validity of the ITP-PAQ. Further analyses on a larger sample are planned to confirm these findings.

on Intrenet for healthy people to prevent lifestyle-related disease.In order to build the system, we performed secure systems with personal information, based on the technical development of a presenting agent network. To modify helthy life-style, the educational program for a lifestyle-related disease were set up Web site. The conpoements of the program were; health educational kits, the exercise and diet diary, self-feeling and the medical cheak up records. The exercise and diet diary were set up about daily movement records, and nutrition analysis of a meal diary. The Medical check up items were based on the health survey in Japan. We evaluated the program with the QOL(SF-36), exercise frequency, and BMI(body mass index) before participation in the program, and 3,6 months after. RESULTS: Nineteen healthy people enrolled in the program. Exercise and counseling with nurses were conducted on the records of exercise, diet diary and medical checkup. Fifty-parcents of the people enrolled this progrem had reguler physical activity. QOL scores and BMI were meintaind 6 months after. CONCLUSIONS: The new educational system will be useful for not only preventing diseases in healthy people but also for patients with a lifestyle-related disease especialy as prevention for the preliminary stage of process.

145/Abstract 1341 Body Image Moderates the Realtionship Between Life Satisfaction and Body Mass Izabela Tabak, School Medicine, Joanna Mazur, Epidemiology, Anna Oblacinska, Maria J. Jodkowska, School Medicine, Institute of Mother and Child, Warsaw, Poland AIMS: The aim of the study was to examine the relationship between body mass index and life satisfaction in 13-15 yrs. old adolescents and to determine how body image modifies this relationship. METHODS: A study was conducted (Grant No RG1/05 Nutricia Poland Research Foundation) in 2005 with 953 gender- and agematched pairs of overweight and obese (BMI above 85th percentile) and non-obese (BMI range 15th-75th percentile) adolescents from 112 lower secondary schools in Poland. School nurses measured weight and height of 8300 students and selected 1906 participants, complying with requirements for groups. Body figure perception was measured by BFPQ (Stunkard & Sorenson), life satisfaction by Cantril ladder. A high body size discrepancy occurred when actual size differed from his/her ideal size equal or higher than 2 points; a low life satisfaction scores on Cantril ladder was defined as lower or equal 5. Stepwise logistic regression was estimated with SPSS v.12. RESULTS: Overweight or obese students more often reported low life satisfaction than non-obese (24% v 18%, p=0.001, OR=1.5), mean scores of life satisfaction: 6.97 v 7.29, p<0.001. Overweight or obese students also more often disclosed high discrepancy in body image (50% v 16%, OR=5.2). Mean scores of life satisfaction were lower for students with high discrepancy in body image among both overweight, obese or non-obese participants, p<0,001, but there were no differences between mean scores of life satisfaction between groups (overweight, obese or non-obese) when we adjusted our results for body figure perception. Stepwise logistic regression revealed that high BMI predicted low life satisfaction, which was 1,5 times higher in overweight and obese group than in non-obese, p=0,001. Adding body size discrepancy to the equation resulted in a significant change of Exp(B)(1.8; p<0.001) and excluding BMI from the final model. CONCLUSIONS: Overweight and obese adolescents are less satisfied with their lives than non-obese, but only when they disclose a high discrepancy in body image.

144/Abstract 1367 Initial Assessment of Reliability and Validity of the Immune Thrombocytopenic Purpura-Patient Assessment Questionnaire (ITP-PAQ) Gary J. Okano, Global Health Economics, Amgen Inc., Thousand Oaks, CA, James B. Bussel, Pediatric Hematology-Oncology, New York-Presbyterian Hosp, Weill/Cornell Med Ctr, New York, NY, James N. George, Hematology-Oncology, University of Oklahoma Health Sciences Center, Oklahoma City, OK, Robert McMillan, Del Mar, CA, Michelle Pritchard, Statistical Analysis, Ovation Research Group, Venice, CA, Janet L. Nichol, Oncology Supportive Care, Amgen Inc., Thousand Oaks, CA AIMS: To assess reliability and validity of the ITP-PAQ. METHODS: Patients with ITP, enrolled in an ongoing open-label study, completed the 44-item ITP-PAQ at baseline (BL), week 4 (W4), 12 (W12), and every 12 weeks to study end. 10 scales: 4 Physical Health (PH-Symptoms, PH-Fatigue, PH-Bother, PHActivity); 2 Emotional Health (EH-Psychological, EH-Fear); 3 Quality of Life (QOL-Overall, QOL-Social, QOL-Work); 1 Women's Reproductive Health (WRH) are scored 0 (worst)-100 (best). Reliability (internal-consistency; test-retest) was assessed by Cronbach's alpha and intraclass correlation coefficients (ICC). ICC was calculated using patients whose platelet (plt) counts were stable between BL and W4. Construct validity evaluated a priori hypothesized relationships among scales using Pearson correlation coefficients. Known-groups validity was examined by a t-test between: W24 plt responders (PR) vs non; W24 durable plt responders (DPR) vs non; BL splenectomy (SPL) vs non; BL concurrent ITP therapy vs non. Responsiveness was based on a paired t-test on the BL to W24 change score for PR or DPR. RESULTS: Data through W24 are available for 34 patients (mean age=50±13; 68% female; 79% white). Cronbach's alpha ranged from .70-.95 except QOL-Work (.69 at W24) and EH-Psychological (>=.95 at W4, W12). ICCs ranged from .73-.87 for all scales except PH-Symptoms (.68) and WRH (.59). Most inter-item/inter-scale associations were consistent with hypotheses. W24 DPR scores were better than non-DPR scores for: PH-Symptoms (79±16 v 64±17); PH-Bother (84±16 v 64±22); EH Psychological (84±18 v 65±26); QOL-Overall (72±22 v 52±26) (all p<.05). Differences were also observed for WRH (SPL 62±33 v non 96±7; p=.03). Trends were observed in the appropriate direction for all other scales. 6 of 10 scales (PH-Symptoms, PH-Fatigue, PH-Bother, PH Activity, EH-

146/Abstract 1704 Measuring Quality of Life Improvement from Migraine Prophylaxis: Psychometric Evaluation of the Migraine Prevention Assessment Clinical Tool (M-Pact) Jason C. Cole, Peggy Lin, QualityMetric, Lincoln, RI, Mike Ingham, Pharmaceuticals Global Strategic Marketing, Johnson & Johnson Pharmaceutical Services LLC, Raritan, NJ, Domenico D'Amico, Headaches and Cerebrovascular Disorders Unit, C. Besta National Neurological Institute, Milan, Italy AIMS: Given the increased availability of migraine preventive treatments, we undertook the development of the M-PACT, designed to measure the impact of migraine prevention on functioning and aspects of daily life. Following literature review, expert input, and patient focus groups (described elsewhere), the current study was designed to examine the M-PACT psychometrics with a clinical sample of migraine and headache patients. METHODS: The M-

A-116

factors, on HRQL in PD. There is the potential to improve HRQL in PD by addressing these non-motor symptoms.

PACT consisted of 19 items with 5 response categories (from not at all improved or worse to greatly improved), using a two-week recall period. The sample of 338 patients was 81% female, 66% migraineurs (vs. headache only), and ranged from 19-84 years in age. Confirmatory factor analysis (CFA) was used to test the dimensionality of the items. Item response theory (IRT) was used to examine measurement properties of the items and bias for modality (web vs. paper), gender, and patient group differences. Classical psychometrics were also examined, including discrimination between migraine and headache groups, item- and scale-level reliability, and discriminant validity with the Headache Impact Test (HIT-6). RESULTS: CFA confirmed a unidimensional construct with high factor loadings of .87-.96. IRT estimates showed a good range of scores with appropriate standard errors, and no items showed bias on modality, gender, or patient group differences. Moreover, the MPACT was found to have excellent classical psychometric properties: alpha=.98, average interitem correlation=.78, discrimination of migraineurs vs. headache patients was F=4.8 (p<.05). CONCLUSIONS: The M-PACT is a psychometrically strong measure of QoL benefit from migraine prophylaxis. Moreover, the M-PACT can be given in either web-based or paper-based modalities. Further examination of this useful instrument should be undertaken with longitudinal analyses (including 2- vs. 4-week recall) to examine the sensitivity to change. Details on M-PACT compliance with recent FDA PRO guidelines will also be examined.

148/Abstract 1311 The Reliability and Responsiveness of the Burden of Stroke Scale (BOSS) Patrick J. Doyle, Geriatric Research Education & Clinical Center, VA Pittsburgh Healthcare System, Pittsburgh, PA, Katherine B. Ross, Speech/Language Pathology, Phoenix VA Medical Center, Phoenix, AZ, James E. Bost, Medicine, University of Pittsburgh, Pittsburgh, PA, William D. Hula, Speech/Language Pathology, VA Pittsburgh Healthcare System, Pittsburgh, PA, Malcolm R. McNeil, Communication Science & Disorders, University of Pittsburgh, Pittsburgh, PA AIMS: To examine the reliability and responsiveness of the Burden of Stroke Scale (BOSS) (Doyle, 2005). METHODS: A prospective cohort of 178 stroke survivors were administered the BOSS twice (T1, T2) at 3 MPO of stroke +/- 14 days (mean T1-T2 interval = 3.8 days). The SF-36 was also administered at T1. At 12 MPO (T3), 145 participants were available for BOSS and SF-36 follow-up assessments. The internal consistency of BOSS scales at T1 and T2 was examined using item-total correlations and Cronbach's alphas. Test-retest reliability was examined by calculating 95% confidence intervals of mean T1-T2 difference scores and T1-T2 intraclass correlation coefficients (ICCs). Internal responsiveness was assessed by calculating T1-T3 standardized response means (SRMs) and the probability of change statistic described by Zou (2005) for subsamples of participants who met Ferguson's (2002) criteria for clinically meaningful change on SF-36 physical and mental components summary scales. RESULTS: Cronbach's alphas exceeded 0.83 for all scales, with item total correlations ranging from 0.31 to 0.87. T1-T2 confidence intervals of all but 2 scales (self-care and negative mood) contained 0 and were narrow. ICCs ranged from 0.78 to 0.94 indicating strong associations between T1 and T2 scale scores. In the subsample who demonstrated clinically meaningful change on the SF-36 PCS scale (n = 40), SRM and probability to detect change values exceeded 0.80 for six of the BOSS domain scales (mobility, self-care, domain mood, domain satisfaction, domain restriction and negative mood) and all composite scales (physical, cognitive, psychological distress). In the subsample demonstrating clinically meaniful change on the SF-36 MCS scale (n = 41), SRM and probability to detect change values exceeded 0.80 for 3 BOSS domain scales (domain mood, domain restriction, and negative mood) and the physical and psychological distress composite scales. CONCLUSIONS: The BOSS provides reliable score estimates of functioning and well-being that are responsive to change during the first year of recovery from stroke.

147/Abstract 1124 Health-Related Quality of Life in Parkinson Disease: A Population-Based Study Sheri L. Pohar, Canadian Agency for Drugs and Technologies in Health, Edmonton, AB, Canada, C. Allyson Jones, Physical Therapy, University of Alberta, Edmonton, AB, Canada, David H. Feeny, Health Utilities Inc., Dundas, ON, Canada AIMS: To examine the associations between 1) stress and healthrelated quality of life (HRQL) and 2) depression and HRQL in community dwelling persons with Parkinson disease (PD) in Canada. METHODS: 259 respondents from the Canadian Community Health Survey (CCHS 1.1) with self-reported PD were included in the analyses. HRQL was assessed with the Health Utilities Index Mark 3 (HUI3), a generic measure that incorporates eight attributes of health (vision, hearing, speech, ambulation, dexterity, emotion, cognition and pain). Respondents were asked to rate their level of life stress most days on a five-point likert scale. The Composite International Diagnostic Interview Short Form for Major Depression (CIDISFMD) was used to identify respondents with depression. Adjusted overall HUI3 and ambulation, emotion, cognition and pain attribute scores were compared according to depression and life stress using ANCOVA. Normalized sampling weights and bootstrap variance estimates were used in the analysis. RESULTS: The average age of respondents with PD was 68.9 years (95% CI: 66.6, 71.2), with an average disease duration of 7.3 years (95% CI: 5.6, 8.9). The mean overall HUI3 score for subjects with PD was 0.56 (95% CI: 0.48, 0.63). The difference in HUI3 scores between respondents who rated their lives as not at all or not very stressful and respondents who rated their lives as quite a bit or extremely stressful was 0.19 (adjusted mean = 0.42, 95% CI: 0.29, 0.55 versus 0.23, 95% CI: 0.10, 0.36, p<0.05). Respondents without depression had HUI3 scores that were 0.29 units higher those with depression (adjusted mean = 0.49, 95% CI: 0.39, 0.59 versus 0.20, 95% CI: 0.03, 0.37; p < 0.05). Clinically important differences in ambulation, emotion, cognition and pain were also observed between respondents with and without depression. CONCLUSIONS: These results highlight the substantial impact of depression and life stress, two potentially modifiable

149/Abstract 1370 Loneliness in Aphasia Katherine B. Ross, Audiology and Speech Pathology, Carl T. Hayden VA Medical Center, Phoenix, AZ, Leonard L. LaPointe, Communication Disorders, Florida State University, Tallahassee, FL, Richard C. Katz, Audiology and Speech Pathology, Carl T. Hayden VA Medical Center, Phoenix, AZ, Patrick J. Doyle, Audiology and Speech Pathology, VA Pittsburgh Healthcare System, Pittsburgh, PA, Robert T. Wertz, Hearing and Speech Sciences, Vanderbilt University, Knoxville, TN AIMS: Survey data reveal that 90% of aphasic people feel socially isolated. Loneliness is associated with negative outcomes that include health problems, divorce, substance abuse, and suicide. Speech pathologists suggest that treatment should be relevant and useful to aphasic

A-117

people and their significant others, thus clinicians might consider treating loneliness. Our aims were to determine whether loneliness in aphasic adults differs significantly from loneliness in non-aphasic adults, and, if so, to identify treatable variables associated with increased loneliness in aphasic adults. METHODS: An observational, between-groups design examined loneliness in 26 aphasic and 21 matched, non-aphasic adults. Data collected were: Loneliness: UCLA Loneliness Scale, Version 2 Demographics: age, gender, education, marital status, employment status Physical health: selfreported comorbidities; physical disability (Modified Rankin Scale) Mental health: depression (Self-Rating Depression Scale); social support (Inventory of Socially Supportive Behaviors); Stroke (aphasic participants): presence of motor involvement; months post-stroke; language impairment (Porch Index of Communicative Ability); functional communication (Communication Activities in Daily Living). To determine whether groups differ significantly in loneliness, an independent samples t-test was used. To determine significant associations between demographic, health-related, stroke-related variables and loneliness within groups, correlations were used, with an alpha level of .05 divided by the total number of analyses. RESULTS: Aphasic adults were significantly lonelier than non-aphasic adults. In both groups, depression was significantly related with loneliness. In non-aphasic adults, social support was also significantly related with loneliness. No other variables were associated with increased loneliness in either group. CONCLUSIONS: Aphasic adults appear at risk for loneliness. Because depression is significantly related with loneliness, determining its presence and the outcome of its treatment on loneliness in aphasia is recommended.

with cognitive functions, age and health. Spirituality could represent a missing element explaining differences between patients with similar neurological conditions. Such findings support a new conceptualization and multidimensional assessment of QOL in these patients.

151/Abstract 1439 Parent-Reported Health-Related Quality of Life in Individuals with Williams-Beuren Syndrome Angela Gosch, Social Sciences, University of Applied Sciences, Munich, Germany, Rainer Pankau, Finkelstein Children's Hospital, Kölpingsee, Germany AIMS: So far no studies are available in which HRQoL has been examined in individuals with Williams-Beuren syndrome (WBS). WBS is a rare genetic disorder caused by a microdeletion on chromosome 7q11.23. A behavioral phenotype with somatic (e.g. heart disease), cognitive (e.g. intellectual disability with better verbal than visuomotor abilities) and behavioral characteristics is described. A high overall level of behavior problems (e.g. ADHS) and a lower level of adaptive behaviors has frequently been found, but at the same time individuals with WBS are described as being overfriendly and "hypersocial". Objectives of this study were to report 1) parent-reported health-related quality of life (HRQoL) in individuals with WBS of different age-groups and 2) effects of gender, the familial socio-economic status, and behavior problems on parent-reported child HRQoL. METHODS: 60 parents of children, adolescents and young adults with WBS (8-23 years) answered the KIDSCREEN-10 Index (Ravens-Sieberer et al., 2001, 2005), a HRQoL instrument with 10 items, the Strength and Difficulty Questionnaire (SDQ, Goodman, 1997) to assess behavior problems and questions addressing the socio-economic status of the family. RESULTS: Parents report a lower HRQoL in individuals with WBS irrespective of the child's age. The impact of gender - a higher HRQoL is noticed in males -, behavior problems, and the familial socio-economic status on the children's HRQoL will be reported. CONCLUSIONS: This is a first study addressing the parent-reported HRQoL in individuals with Williams-Beuren-Syndrome. Like in other studies with individuals with an intellectual disability a lower HRQoL is found in this group. The impact of gender and behavioral problems is similar to the findings in healthy individuals. Consequences of the results and recommendations for prevention and intervention programmes will be discussed.

150/Abstract 1470 Quality of Life in Chronic Neurological Diseases: The Contribution of Spirituality Anna R. Giovagnoli, Neuropathology and Neurology, National Neurological Institute Carlo Besta, Milano, Italy, Rute F. Meneses, Psychology, FCHS-Fernando Pessoa University, Porto, Portugal, Antonio Martins da Silva, Neurological Sciences, Hospital Santo Antonio, Porto, Portugal AIMS: Recent studies extended the assessment of quality of life (QOL) in patients with chronic neurological diseases but most failed to explain the variability of QOL of people with the same diagnosis. This study explored the contribution of spirituality to QOL with respect to cognitive, affective, clinical, and demographic factors. METHODS: 72 patients with epilepsy, brain tumor, vascular or immune-mediate brain lesions compiled inventories for QOL (WHOQOL 100), Spirituality (Spiritual, Religious and Personal Beliefs, SRPB), depression (Beck Depression Inventory, BDI) anxiety (State-Trait Anxiety Inventory, STAI), and cognitive selfefficacy (Multiple Ability Self-Report Questionnaire, MASQ) and underwent neuropsychological testing. RESULTS: Factor analysis of the SRPB facets, STAI, and BDI scores yielded four factors (Personal Meaning, Inner Energy, Awe and Openess, and Mood); factor analysis of the MASQ and neuropsychological test scores produced three factors (Control Functions, Cognition, Memory). Correlation and regression analyses showed that the WHOQOL 100 total score was significantly predicted by the Mood (40% of variance) and Inner Energy factors (48% of residual variance). In addition, Mood predicted the Psychological and Level of Independence domains, and Cognition predicted the Environment domain. Age, work level, and health status perception predicted the WHOQOL 100 total score. No association was found between QOL scores and disease duration, diagnosis, gender, schooling, marital status, number of family members, financial resources, religiosity or belonging to a religious community. CONCLUSIONS: The results show that, in patients with chronic neurological diseases, spirituality is distinct from affect and that both significantly contribute to determine QOL, interacting

152/Abstract 1283 Quality of Life in Patients with Multiple Sclerosis: A Comperative Study Saharnaz Nedjat, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran, Ali Montazeri, Mental Health, Iranian Institute for Health Sciences Research, Tehran, Iran, Kourosh Holakouie, Kazem Mohammad, Reza Majdzadeh, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran AIMS: The objective of the current study was to investigate quality of life in patients with multiple sclerosis and compare it with that for the general population. METHODS: The Iranian version of the WHOQOL-BREF (including Physical Health, Psychological, Social Relationship, and Environment domains) was used to measure quality of life. The scores for each domain range from 4 to 20, with higher scores indicating a better condition. Randomly selected sample of patients with multiple sclerosis were completed the questionnaire. The scores then were compared with scores for the general population derived from a normative study in Tehran. RESULTS: In all 145 patients were interviewed. The mean (SD) scores for the Physical, Psychological, Social

A-118

Relationship and Environment domains were 11.5(3.2), 11.8(3.2), 12.7(3.1) and 11.5(2.7) respectively. The scores in all domains were significantly lower in multiple sclerosis cases as compared to that for the general population (Table). Educational level has a significantly positive effect in scores for all domains; also the severity of the disease has a significant negative effect in physical and psychological domains. CONCLUSIONS: The findings suggest that quality of life in patients with multiple sclerosis is very poor. The results also showed that the WHOQOL-BREF in addition to the general population is a valid instrument for measuring quality of life in patients such as multiple sclerosis.

Physical Health Psychol-ogical Social Relationship Environ-ment

General Population (n=906) Mean (SD) 14.8 (2.3) 13.7 (2.5)

Multiple Sclerosis (n=145) Mean (SD) P 11.5 (3.2) < 0.001 11.8 (3.2) < 0.001

14.0 (2.4)

12.7 (3.1)

< 0.001

12.7 (2.6)

11.5 (2.7)

< 0.001

154/Abstract 1779 A Practical Sympyom Severity Measure for Urinary Incontinence John T. Wei, Rodney Dunn, Gary Faerber, Richard Dorr, Edward Mcguire, Urology, University of Michigan, Ann Arbor, MI AIMS: To develop and validate a practical, patient self-completed measure for assessing urinary incontinence severity and impairment. METHODS: Survey items were piloted and refined following psychometric principles. Patient and expert endorsement of items, factor analyses, item-scale correlations and response distributions were employed for item selection. Factor analysis yielded 3 domains (stress symptoms, urge symptoms, pad use). Two additional items assess impairment. 99 men and women with urinary incontinence participated in the final validation phase. RESULTS: Urologists, urogynecologists, and patients with urinary incontinence confirmed face and content validity. Predictive validity was demonstrated with decreased (improved) scores for the stress, urge, total severity and impairment scales (all p < 0.05) following therapy. Cronbach's alpha for the total ISI score and Impairment score were high at 0.90 and 0.82, respectively. Cronbach's alpha for the subdomains were each > 0.85. Correlation with the SF-12 and established measures of pelvic floor function demonstrated convergent (Hunskaar, IIQ) and divergent validities (PISQ, MCS, PCS)(Table). Test-rest reliability exceeded 0.8 in all domains. CONCLUSIONS: The ISI is a concise, self-administered urinary incontinence measure appropriate for both the research and clinical settings.

153/Abstract 1797 Correlation between the SF-36 and WHOQoL-BREF Ali Montazeri, Mental Health, Iranian Institute for Health Sciences Research, Tehran, Iran, Saharnaz Nedjat, Kourosh Holakouie, Kazem Mohammad, Reza Majdzadeh, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran

Convergent and Divergent Validity

AIMS: The objective of the current study was to investigate whether the SF-36 and WHOQOL-BREF are measuring the same quality of life dimensions. METHODS: As part of a validation study, the Iranian versions of the SF-36 and WHOQOL-BREF were compared. A random sample of the general population completed both questionnaires. Pearson correlation coefficient was used to estimate to what extent these are related measures. In addition an exploratory factor analysis was carried out to determine factor structures of the scales. RESULTS: In all 89 individuals were interviewed. The correlations between the SF-36 and WHOQOL-BREF are shown in the Table. Although for many subscales correlations were significant at 0.01 levels (figures with *), the analysis did not show strong correlations between subscales except for the WHOQOL physical health and the SF-36 subscales. The factor analysis also indicated two distinct factors accounted for more than 50% of variance in all subscales. CONCLUSIONS: The study results suggest that the SF-36 and WHOQOL-BREF are measuring different constructs. Further investigations are needed to confirm the findings. Physical Psycholo- Social Environhealth gical relations mental Physical 0.47* 0.04 0.05* 0.04 functioning Role physical 0.39* 0.17 0.07 0.10 Bodily pain 0.45* 0.20 0.20 0.08 General health 0.41* 0.35* 0.18 0.38* Social 0.49* 0.40* 0.22* 0.24* functioning Role 0.43* 0.25* 0.06 0.23* emotional Vitality 0.55* 0.40* 0.21* 0.29* Mental health 0.43* 0.40* 0.20 0.36*

Huns kaar score MCS score PCS score PISQ score IIQ Score

ISI ISI Pad ISI ISI Urge Stress use Bother score score score score

Total ISI score

0.71

0.80

0.76

0.73

0.85

-0.07 -0.12

-0.08

-0.18

-0.10

-0.39 -0.36

-0.36

-0.39

-0.41

0.44

0.37

0.43

0.51

0.46

0.71

0.66

0.66

0.85

0.75

155/Abstract 1791 Scoring and Concurrent Validity for the Korean Version McMaster Quality of Life Scale (K-MQLS) and Spitzer's Quality of Life (QLI) in Hospice and Palliative Care Patients Kyeong Uoon Kim, Nursing, University of Wisconsin, Madison, WI AIMS: Terminal cancer patients and their families sometimes experience serious crisis of quality of life due to cancer pain, physical or emotional suffering, etc. It should be emphasized to provide hospice and palliative care in these cases. Objective of this study were to measure the quality of life of hospice and palliative care patients by using K-MQLS tool (developed by Kim et al (2004) in Korea with proof of validation and reliability) and QLI (develped by Spitzer et al (1981)) and to verify the concurrent validity between the two tools. METHODS: Objects were seventy terminal cancer patients who have received hospice and palliative care in six hospitals and one facility in Korea and ten hospice and palliative care doctors.

A-119

The characteristics were described by descriptive statistics. Score of quality of life were analysed by average and standard deviation. Concurrent validity were calculated by Pearson correlation coefficient. RESULTS: Total K-MQLS (Mean±SD = 3.79±0.74) and QLI (Mean±SD = 4.58±1.63) showed moderate degree of quality of life. There is moderate degree of correlation and significant difference between total score of K-MQLS and that of QLI (r=0.55, p=0.002). CONCLUSIONS: K-MQLS and QLI were verified as useful tools for measuring quality of life in hospice and palliative care patients by using concurrent validity. These tools could be used as basic data for the study of interventional effect of hospice and palliative care.

Lyon, Rhone, France, Gerard DURU, LASS, Universite Claude Bernard, Lyon-1, Villeurbanne, Rhone, France AIMS: The successful application of Patient-Reported Outcomes (PROs) in clinical research was expected to turn into their rapid dissemination in clinical practice as well. However, the adoption of existing tools by clinicians still is limited, despite efforts made by developers to shorten and simplify the instruments. Our assumption is that the expected properties of instruments may differ because they differ in their objectives. This paper explores the differences between PROs according to their intended use. METHODS: Original articles, review articles, research papers and abstracts in English and French published in peer-review journals were reviewed to support the discussion. RESULTS: Beyond brevity, simplicity of scoring procedure and clinically sound contents, major other characteristics differ according to the intended use of a questionnaire. The perspective, the operational objective, the main property of the score, the rationale for content selection and item reduction, have to be adapted. CONCLUSIONS: Important differences exist between measurement instruments for clinical research and tools for clinical practice. They are likely to have an impact on the development as well as on the validation of PROs for clinical practice.

156/Abstract 1618 Validation of A Priority Tool for Catarct Extraction Antonio Escobar, Unidad de Investigacion, Hospital de Basurto, Bilbao, Bizkaia, Spain, Jose M. Quintana, Unidad de Investigacion, Hospital de Galdakao, Galdakao, Bizkaia, Spain, Amaia Bilbao, Oficina de apoyo metodologico, Fundacion Vasca de Innovacion e investigaciones Sani, Sondika, Bizkaia, Spain, Susana García, Urko Aguirre, Unidad de Investigacion, Hospital de Galdakao, Galdakao, Bizkaia, Spain

158/Abstract 1044 AIMS: Different priority criteria for cataract extraction have been created in different countries, but none has included the appropriateness of the intervention. The aim of the study was to evaluate the validity of the developed priority tool using the RAND method (RAM) METHODS: We developed the priority criteria on a modification of the RAM. Priority in the context of cataract extraction was defined as the benefit that the patient may obtain from undergoing surgery. Eight variables were considered: appropriateness, ocular comorbidities, visual acuity in the cataractous eye, visual function, visual acuity in the contralateral eye, expected visual acuity after the intervention, type of cataract (laterality), and social dependence. 310 scenarios were obtained. A panel of 11 ophthalmologists was assembled and scored the scenarios on a 9point scale, with 9 indicating the highest priority and 1 the lowest. The 310 scenarios, were classified as high, intermediate or low priority. Priority scoring systems were developed by means of optimal scaling (OS) and general linear models (GLM). Weights for each variable category were apportioned among variables so that the scores range from 0 to 100. We evaluated the validity of the tool by means of the ANOVA, taken as dependent variable the score of each dimension of SF-36 and the VF -14 and the 3 groups of priority as independent. Regression models were also used to evaluate de relationship between priority score and each dimension of the questionnaires RESULTS: We studied 1899 patients. The mean age was 73.7 (SD=8.4) and 59.5% of sample was female. The mean priority scores were 66 (SD=15.1) for OS and 70.7 (SD=13.6) for GLM. There were differences among the 3 groups of priority in the physical function and role physical dimensions of the SF-36 questionnaire and in the VF-14 score. As expected, in SF-36 (no bodily pain) and VF-14 the regression coefficients were all negative and statistically significant CONCLUSIONS: Our results showed acceptable validity for prioritizing cataract extraction patients

Primary Care Physicians' Diagnosis, Self Reported Symptoms and Disease Severity to Test the Discriminative Validity of the Italian Version of the EQ-5D: A Cross Sectional Study in North Italy Elena Savoia, Maria P. Fantini, Laura Dallolio, Department of Medicine and Public Health, University of Bologna, Bologna, Italy, Paolo P. Pandolfi, Department of Public Health, Local Health Authority of Bologna, San Lazzaro di Savena, Italy, Natalina Collina, Department Public Health, Local Health Authorityof Bologna, San Lazzaro di Savena, Italy, Milena Belletti, Giuliana Fabbri, Department of Medicine and Public Health, University of Bologna, Bologna, Italy, Francesca Raggi, Department of Oncology, Local Health Authority of Bologna, Bologna, Italy

AIMS: The primary goal of our study was to assess the discriminative validity of the Italian version of the EQ-5D in a freeliving population of North Italy using socio-demographic factors and diagnostic sub-groups. Our secondary goal was to compare the discriminative validity of the EQ-5D to the SF-12_s and assess its variation according to disease severity and source of clinical data information: primary care physicians_ diagnosis versus self-reported clinical conditions. METHODS: The SF-12, the EQ-5D plus an additional questionnaire on socio-demographic characteristics, clinical conditions and symptoms was filled by 1,622 adults, randomly selected from the Registry of the Health Authorities of the city of Bologna, Italy. The primary care physician of each subject was contacted to report on the subject_s health status. RESULTS: We found the discriminative validity of the EQ-5D to be similar to the SF-12_s. Both instruments reported low scores in subjects of older age, females, widowed, unemployed and subjects with a low educational level. Results supported our hypothesis that the source of clinical data information as well as disease severity affects the instrument_s ability to detect differences between diagnostic subgroups. CONCLUSIONS: Both the Italian versions of EQ-5D and SF-12 can be used to assess and monitor QOL in the general population; the two instruments have similar discriminative validity. When a QOL study is conducted in the general population primary care physicians_ health status assessments may strongly improve the identification of the determinants of low quality of life. The effort of

157/Abstract 1689 A Comparison of the Characteristics of Patient-Reported Outcomes According to Their Use (Clinical Research or Clinical Practice) Benoit Arnould, Patient Scales for Clinical Practice, MAPI Values,

A-120

obtaining medical diagnosis pays back on the ability of detecting valid differences between diagnostic sub-groups.

The literature review found over 30 measures that have been used in diabetes studies. Some measures have been used extensively and were well developed, are practical for administration in clinical trials, and have been shown to display sound psychometric properties. Recommendations were provided for administration of measures to assess symptoms, well-being, vitality and depression, all of which are key outcomes in clinical trials and patient care. However, unlike other chronic diseases (e.g. COPD) the incorporation of HRQOL and PRO measures within clinical trial programmes has to date not resulted in associated label claims and the possible reasons for this were discussed. CONCLUSIONS: A number of PRO measures currently exist in diabetes, however, in view of the impact of diabetes treatment regimens on HRQOL and the rapidly evolving compounds in development, companies may need to adapt existing PRO measures or develop alternative measures to achieve label claims.

159/Abstract 1166 QoL of Mothers Who Take Care of Children With Infantile Autism Sachiko Emoto, Michiko Kobayashi, Siebold University of Nagasaki, Nagayo-cho,, Nagasaki, Japan, Takashi Mandai, Japanese Society of Quality of Life Research, Kobe-shi,, Hyogo, Japan AIMS: The purpose of this study was to evaluate the Quality of Life(QOL)of mothers who take care of children with infantile autism. METHODS: Forty-two mothers who take care of children with infantile autism participated in this study. Original selfadministered questionnaire including 44 questions divided into 11 categories was used. RESULTS: Conbach's alpha coefficients of our questionnaire were high enough to accept for clinical use : 0.97 in socialized function, 0.96 in well-being, 0.96 in dietary problem, 0.96 in sleep, 0.96 in control of childcare, 0.96 in environmental problem, 0.96 in economical problem, and 0.96 in mother-child interaction, respectively. Our questionnaire contained 12 main factors and cumulative contribution was 0.82. There were significant QOL differences in the categories of well-being, control of childcare, mental function, mother-child interaction, fundamental life style (P<0.05) between the excellent and the impaired QOL mothers. CONCLUSIONS: These findings indicate that our questionnaire has high enough reliability and potency of validity to use for mothers who take care of children with infantile autism. As the support for mothers who take care of children with infantile autism, right recognition for infantile autism, introduction of parent's association, and handling procedure for children are most important. We must pay more attention for the mothers who take care of children with infantile autism to establish the social support system without anxiety.

161/Abstract 1573 Lack of Differences Between Health-Related Quality of Life in a Representative Sample of Cancer Patients and in the General Population: Why? Mogens Groenvold, Camilla Pedersen, Cecilia R. Jensen, Mette T. Faber, Anna T. Johnsen, Palliative Medicine, Bispebjerg Hospital, Copenhagen, Denmark AIMS: To compare the HRQL of a representative sample of cancer patients against scores of a general population sample. METHODS: Study population: A random, representative sample of 2,245 persons who had been in contact with a hospital for cancer during the last 12 months. Measurement: the EORTC QLQ-C30 was sent by mail. Scores were compared against a Danish general population sample consisting of women (response rate 69%)(Klee M et al. 1997). As version 1 was used in the general population study and version 3 in the cancer study only scales that had not been changed were compared. The same translation of items was used. Scores were transformed to 0-100 scales. RESULTS: Of the cancer patients 1,518 (68%) responded. There were more women (64%) than men. The mean age was 64 years. Almost cancer diagnoses were represented; the largest subgroup was breast cancer (33%). 38% had advanced disease (stage 3/4). 40% had been diagnosed within two years. There were few and small differences between cancer patients and the general population sample. The differences (general population cancer patients) (a negative score representing lower quality of life in cancer patients) were: Emotional function 4, cognitive function -2, social function -4, fatigue -5, nausea/vomiting -2, pain -2, dyspnea -6, lack of sleep 0, lack of appetite -6, constipation -4, diarrhea -4, financial difficulties 0. CONCLUSIONS: Surprisingly few and small differences were observed. The lack of differences cannot be explained by the finding from other general population studies that women report lower QL or by differences in age. As previously reported for the Hospital Anxiety and Depression Scale (Groenvold M et al. 1999) the lack of differences between cancer patients and the general population is most likely due to response-shift, positive readaptation, or selective reporting in cancer patients. Such direct comparisons between persons in very different situations may not provide valid results.

160/Abstract 1702 Patient Reported Outcome in Diabetes and their Regulatory Review. Darren J. Clayson, Pharmaquest Ltd, Banbury, Oxfordshire, UK, Frank J. Snoek, Diabetes Psychology Research Group, VU University Medical Center, Amsterdam, The Netherlands, William H. Polonsky, Behavioral Diabetes Institute, University of California, San Diego, California, Melva T. Covington, Global Health Outcomes, Eli Lilly and Company, Indianapolis, Indiana AIMS: Regulatory authorities will approve a compound for the treatment of a disease if it can 1) be shown to be curative in a disease, 2) be curative for a fraction of those afflicted, 3) prolong disease-free survival or 4) improve quality of life while the therapy is administered. As the first two criteria are currently not applicable in diabetes, health-related quality of life (HRQOL) and other patient reported outcome (PRO) assessments have become critical measures in clinical trials of new treatments. This literature review was performed to provide recommendations for future clinical trials based on the adequacy of the existing portfolio of measures. METHODS: Within diabetes, research into HRQOL and other PRO assessments is well developed. A comprehensive literature review was conducted to examine the properties of PRO measures used in diabetes research. Measures were reviewed in terms of their development, practicality and psychometric properties. The review also included interviews with key opinion leaders in the area of diabetes to provide a clinical perspective on the adequacy of the existing measures. RESULTS:

162/Abstract 1399 The Expanded Prostate Cancer Index Composite (EPIC): Validity to Measure the Impact of Different Treatments for Localized Prostate Cancer Àngels Pont, Health Services Research Unit, Institut Municipal d'Investigació Mèdica (IMIM-IMAS), Barcelona, Spain, Joan

A-121

Status(POMS) fourtimes in 1,3,6,12 months after operation, respectively. RESULTS: From 6 to 12 months after operation, the significant improvements of QOLs were shown especially in physical and mental problems. There was significant correlation between QOL and occupation in one month after operation. In 3 and 6 months after operation, significant correlations between occupation, age, assistant therapy and QOL were indicated, but no significant correlation was demonstrated in 12 months after operation. There were strong correlations between the results by POMS and FACT-B in every time. There was also strong correlation between QOL and mood. Especially, anxiety, tension and depression were most important factors for QOL impairments. CONCLUSIONS: These findings indicate that we must pay more attention to support the patients with breast cancer operation by considering causes and timing on QOL deterioration continuously.

Mendivil, Olatz Garin, Health Services Research Unit, Institut Municipal d'Investigació Mèdica, Barcelona, Spain, Ferran Guedea, Oncologia Radioteràpica, Institut Català d'Oncologia, L'Hospitalet de Llobregat, Barcelona, Spain, Ferran Aguiló, Servei d'Urologia, C.S.U. Bellvitge, L'Hospitalet de Llobregat, Barcelona, Spain, Fernandez Pablo, Oncología Radioterápica, Instituto Oncológico de Guipuzkoa, San Sebastián, Spain, Victor Macias, Oncologia Radioteràpica, Hospital General de Catalunya, Sant Cugat del Vallès, Barcelona, Spain, Alfonso Mariño, Oncología Radioterápica, Centro Oncológico de Galicia, A Coruña, Spain, Javier Ponce de Leon, Urologia, Fundació Puigvert, Barcelona, Spain, Gemma Sancho, Oncologia Radioteràpica, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, Montse Ferrer, Health Services Research Unit, Institut Municipal d'Investigació Mèdica, Barcelona, Spain AIMS: Prostate cancer is asymptomatic during the earlier stages but treatment could have a considerable impact on patients HRQL. The EPIC was developed to measure the impact of different treatments but there is little evidence. It measures 4 summary scores (50 items): Urinary, Bowel, Sexual and Hormonal. The objective was to assess reliability and validity of the EPIC. METHODS: Patients came from a larger study; the first 50 patients receiving radical prostatectomy (RP), prostate brachytherapy (PB) and external radiotherapy (ER) were analyzed. The EPIC was administered before and one month after treatment using computer-assisted telephone administration. The Spanish version of the EPIC was obtained following a standard adaptation procedure which included forward and back-translation with an expert panel and pilot test. Patients receiving hormonotherapy were compared to patients who did not to assess cross-sectional validity. Longitudinal validity was evaluated by comparing the effect size (ES) of change pre-post treatment scores between the treatment options. RESULTS: Mean age was 74 years and 28% of the total sample was receiving hormonotherapy. Cronbach's alpha achieved the required standard of 0.7, except for Bowel. Patients receiving hormonotherapy reported worse HRQL on Sexual and Hormonal scores than patients who did not (p<0.05). Patients treated with RP reported the most sizeable deterioration in Urinary and Sexual scores (ES of 4.5 and 1.2, respectively, p<0.001); among patients treated with PB, deterioration was observed mainly on the Urinary and Bowel scores (ES of 3.1 and 1.0, respectively, p<0.001); and for patients receiving ER, the Bowel score reflected the most sizeable deterioration (ES of 1.7, p<0.001). CONCLUSIONS: These results suggest that the EPIC is a valid instrument for measuring the impact on HRQL of different treatments for prostate cancer. Funded by: DURSI-GENCAT (2005-SGR00491), AATRM 086/24/2000, FIS PI020668, FIS PI020078 and IRYSS Network (G03/202 and FIS PI052403).

164/Abstract 1185 The Concepts of Spirituality for Japanese Jatients with Terminal Lung Cancer Hideki Origasa, Biostatistics and Clinical Epidemiology, University of Toyama School of Medicine, Toyama, Japan, Akihiro Tokoro, Psychosomatic Internal Medicine, Kinki-chuo Chest Medical Center, Sakai, Japan, Miyako Tazaki, Psychology, Tokyo University of Science, Tokyo, Japan, Nobuko Yotsutani, Biostatistics and Clinical Epidemiology, University of Toyama School of Medicine, Toyama, Japan, Toshiko Goto, Nursing, Masaaki Kawahara, Internal Medicine, Kinki-chuo Chest Medical Center, Sakai, Japan AIMS: We conducted a qualitative study in order to develop a new questionnaire of spirituality for Japanese terminal lung cancer patients. This study would clarify any difference in the concepts of spirituality between Japan and western countries. METHODS: We recruited a focus group that could exhibit important items of spirituality. It consisted of three panels of 6 terminal lung cancer patients, 5 participants from the family of a patient with terminal lung cancer, and 3 nurses working at the lung cancer ward. A clinical psychologist moderated an interview at each panel for 1-2 hours. Interviews were recorded and summarized into items. We constructed facets of the spirituality and classified them into important domains. RESULTS: There were a total of 198 items regarding the spirituality generated by participants; 76 from patients, 67 from families, and 55 from nurses. We found that there were two different stages for each patient. First was a struggling stage against the situation and the second was an accepting stage of coming death. By qualitative analysis, the following 17 facets were appeared and those were classified into the above stages. In the struggling stage, seven facets of introspection, acceptance of disease, distrust to medicine, hopefulness, blockade in mind, anxiety of disease/work/family, and painful distress were chosen. In the accepting stage, nine facets of spirits in life, meaning of life, inner strength, peaceful mind, kindness, thanks to family/friends, regression to nature, acceptance/forgiveness of person and disease, joy/happiness were raised. These facets were further summarized into the domains of anxiety, distress, and distrust for struggling stage; and spirits in life, peaceful mind, joy/happiness for accepting stage. For Japanese, religiosity was lacked in general; however regression to nature seemed to be replaced by that. CONCLUSIONS: Spirituality was recognized from two different stages of struggle and acceptance. A total of 17 facets were important for Japanese terminal cancer patients. Religiosity seemed to be lacked for the Japanese.

163/Abstract 1186 Quality of Life Changes and Associated Factors After Breast Cancer Operation Fumiko Utsumi, Siebold University of Nagasaki, Nagayo-cho,, Nagasaki, Japan, Mitsuko Ikematsu, Kinuko Ahiru, Yuuka Sasaki, Shinji Ohno, Akira Ohshima, National Kyushu Cancer Center, Fukuoka-shi,, Fukuoka, Japan AIMS: The purpose of this study was to make clear the interventional timing for the patients with breast cancer operation in order to prevent the Quality of Life(QOL)deterioration by investigation on postoperative QOL changes and associated factors prospectively after operation. METHODS: One hundred nineteen patients with the first breast cancer operation participated in this study. QOLs were evaluated by using the Functional Assessment of Cancer Therapy-Breast(FACT-B)and the Profile of Moods

A-122

study. Participants completed the International Prostate Symptom Score before treatment, at treatment completion, and 6, 12 and 24 months after therapy. Responses were compared to the genitourinary quality of life using the Mantel-Haenszel Chi-Square test. RESULTS: A total of 1186 questionnaires were analyzed with 322 responses at baseline, 294 at treatment completion, 243 at 6 months, 201 at 12 months and 126 at 2 years. The specific urinary complaints; incomplete emptying, frequency, intermittency, urgency, weak stream, straining and nocturia were all significantly associated with the overall genitourinary quality of life at all time points (p<0.0001). CONCLUSIONS: Genitourinary quality of life was not dominated by any particular urinary complaint, however, it was strongly associated with all complaints across time. Therapeutic options should seek to minimize all urinary complaints and the overall quality of life rather than one at the expense of another.

165/Abstract 1754 Quality Of Life Of Retinoblastoma Survivors in The Netherlands Jennifer van Dijk, Dept. of Psychology and Dept. of Opthalmology, VU Medical Centre, Amsterdam, the Netherlands, Saskia Imhof, Dept. Opthalmology, Jaap Huisman, Dept. of Psychology, VU Medical Centre, Amsterdam, Amsterdam, Annette Moll, Dept. of Opthalmology, Peggy Cohen-Kettenis, Dept. of Psychology, VU medical centre, Amsterdam, the Netherlands, Peter Ringens, Dept. of Opthalmology, VU medical centre, Amsterdam, Amsterdam AIMS: Retinoblastoma (rb) is a highly malignant tumour of the eye occurring in very young children (0-6 yrs). In the Netherlands the incidence of rb is 1: 17.000 newborns. In the Dutch rb center 732 survivors are registered. Rb can have debilitating effects on vision, appearance and mental or social life. The aim of this study is to describe the quality of life of rb adult survivors in relation to the normal population and between two subgroups of hereditary and nonhereditary forms of rb. METHODS: Design: cross-sectional within subject design. From June 2005, all adult rb survivors (18-35 yrs) have been invited to participate in the study (N= 60). They administered the SF-36 quality of life questionnaire by self-report (35% males, 50% hereditary) and a semi-structured disease specific interview, during a home visit. The interview contains psychosocial questions, e.g. coping, living with a visual impairment, hereditary or rare disease, and was based on literature and focus group discussions with rb experts and survivors. RESULTS: Independent sample t-tests showed no differences between the adult rb survivors (mean: physical=53.3; mental=53.5) and the normal population (ph=48.8; me=51.9), on both SF-36 composite scores. Results from the hereditary group (ph=51.9; me=53.8) were comparable to the nonhereditary rb group (ph=54.3; me=53.2). However, the interview showed that 38% did experience several problems: fear of second primary tumours, visual impairment and having children with increased rb-risk, uncertainty, dependency and negative self-image, problems at school and work. Mann-Whitney tests showed lower scores on five SF-36 subscales (Role Physical functioning, Bodily pain, General health, Vitality and Mental health) compared to other rb survivors (p<0.05). CONCLUSIONS: We conclude that adult rb survivors as a group experience relatively good quality of life, however quite a large group of rb adult survivors do have considerable problems. The causes of differences in outcome might be due to treatment or coping-style differences and need to be further explored.

167/Abstract 1809 Waste Treatment by an Incinerator: What About Health-Related Quality of Life? M. Fátima Reis, Osvaldo R. Santos, Paula Gomes, Pedro Aniceto, Pedro Aguiar, Preventive Medicine, Faculty of Medicine, Lisbon, Portugal, Maurício Melim, Regional Direction of Public Health (RAM), Health Ministry, Funchal, Portugal, Liliana Cardoso, J. Pereira Miguel, Preventive Medicine, Faculty of Medicine, Lisbon, Portugal AIMS: Waste incineration impact on health-related quality of life and on self-perceived general health state is scarce. This paper derives from a longitudinal monitorization program of the impact of continuous exposition to waste incineration activity on self-perceived health status and on other health-related quality of live indicators. METHODS: The study follows a longitudinal and controlled design. Two different areas were defined according to the level of exposition of the incineration effects. In both areas, subjects were selected/recruited following a random route procedure. MOS SF-36 was used for assessing self-perception of general health and of other health-related quality of life dimensions. RESULTS: Collected data comes from the first assessment moment of this epidemiological monitorization program. Statistical significant differences were found between exposed and non-exposed groups regarding self-rated general health after controlling for gender, age, educational level and being or not professionally active. SF-36 mean scores regarding the dimensions of body pain, general health, vitality and mental health, as well as the mean score from SF-36 physical component, were also found to be significant higher (meaning better health-related quality of life) in the non-exposed group than in the exposed group. CONCLUSIONS: Data suggest an association between exposition to incineration activity and health-related quality of life. The continuation of the study will allow to understand better whether these results are mainly an effect of risk perception or if they result of eventual waste treatment adverse effects on health.

166/Abstract 1156 Genitourinary Quality of Life Compared to Radiation Therapy Induced Urinary Symptoms in Prostate Cancer Patients Maria M. Hardy, Radiation Oncology, Mamtha Balasubramaniam, Research Institute, John M. Robertson, Radiation Oncology, William Beaumont Hospital, Royal Oak, MI, Ann Calcaterra, Radiation Oncology, William Beaumont Hospital, Troy, MI AIMS: The radiation therapy treatment options for prostate cancer have different genitourinary side effects and complications. Knowledge of the relationship between these toxicities compared to the overall genitourinary quality of life may improve the ability to make individualized treatment recommendations. Based on our experience, we hypothesized that specific urinary symptoms, such as urgency, would demonstrate a stronger relationship with urinary quality of life then others, such as nocturia, and tested for this relationship. METHODS: 325 men with prostate cancer treated with 7 different radiation therapy and hormonal combinations participated in this Institutional Review Board approved longitudinal prospective

168/Abstract 1255 Quality of Life and Spiritual Religious Coping in Panic Disorder: Longitudinal Preliminary Data Raquel G. Panzini, Psychiatry, Denise R. Bandeira, Psychology, Universidade Federal do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, Brasil, Marcelo P. Fleck, Psychiatry, Hospital de Clínicas de Porto Alegre, Porto Alegre, Rio Grande do Sul, Brasil AIMS: To evaluate treatment sensibility to change of two spiritual/religious (S/R) measures: World Health Organization

A-123

Quality of Life Instrument-Spirituality, Religiosity and Personal Beliefs Module (WHOQOL-SRPB) and Brief Spiritual Religious Coping Scale (Brief SRCOPE), in a longitudinal study with Panic disorder patients. METHODS: Thirty-eight Panic disorder patients using either cognitive-behavioral psychotherapy or antidepressant (minalcipran) were assessed at baseline, middle term and just after treatment with the following instruments: Brief SRCOPE, WHOQOL-SRPB, WHOQOL-BREF, Panic Inventory, Clinical Global Impression (CGI), Hamilton Anxiety Scale, Beck Depression Inventory. RESULTS: After treatment, 39.5% patients achieved symptoms complete remission (0 crisis+CGI<2), 29% partial remission [0 crisis(13.2%) or CGI<2(15.8%)] and usually went from moderately to slightly symptomatic on CGI. Using Paired t-Tests, in average, throughout Panic disorder treatment, patient's anxiety (p<.004) and depression (p<.0001) scores reduced and quality of life (QoL) scores increased (p<.0001). Also, the use of negative spiritual religious coping (SRC) decreased (p<.02), improving their Negative SRC/Positive SRC Ratios (p<.05), and patient's General Spiritual QoL scores increased (p<.05), as well in the domains Experiences of Awe & Wonder (p<.03), Wholeness & Integration (p<.02) and Inner Peace (p<.01). CONCLUSIONS: Negative SRC reduction throughout treatment was associated with improvements on health scores and on use of Positive SRC in relation to Negative SRC (NSRC/PSRC Ratio). Also, General Spiritual QoL and others spiritual domains showed positive association with health status measures. Preliminary results attested treatment sensibility to change of WHOQOL-SRPB and Brief SRCOPE in Panic disorder patients, and highlighted the S/R based interventions value to promote health population. Regression analyses to evaluate S/R variables capacity to predict clinical outcomes in Panic disorder patients will be conduced after an adequate sample size estimation be achieved (N=60).

There were not statistically significant differences in the changes in HRQoL, measured by the two summary scores of the SF-36, among the 3 appropriateness categories nor in the complications rate postintervention. But, differences were found with the VF-14, having the appropriate an improvement of 25.0 points, the uncertain 20.1 and the inappropriate 16.4 points (p<0.01). Improvement on visual acuity (scale 0-1) was 0.52 for the appropriate , 0.45 for the uncertain, 0.36 for the inappropriate (p<0.01). CONCLUSIONS: This results support partially the predictive validity of our appropriateness tool. Also, they questioned the only use of generic HRQoL questionnaires for outcomes assessment purposes in this pathology since they seem unable to capture relevant changes in these patients.

170/Abstract 1350 The Problem of Purpose in Quality of Life Measurement Leah M. McClimans, Philosophy, London School of Economics, London, United Kingdom AIMS: Validity is an important property of HRQoL measures, but respondent bias threatens a measures validity. It's impossible to get rid of all respondent bias. The most we can hope to do is minimize it where possible and use auxillary hypotheses to explain poor validity when it occurs. But unfortunately responsent bias doesn't always show up in classical validity tests. In fact it only shows up when respondents answer contrary to established hypotheses as to how they should answer. But respondent bias also occurs when respondents answer in line with established hypotheses, but for different reasons. This paper argues that we can illuminate and reduce 'silent bias' only if we let respondents ask as well as answer questions. METHODS: To begin I look at two different studies of the NHP and the SIP which illustrate the problem of 'silent bias'. I argue that 'silent bias' occurs because respondents don't understand the intended purpose of the questions asked. But understanding a question's purpose is vital to understanding a question. I argue that if we want respondents to understand the purpose of our questions than the proper place to locate that purpose is in the measure's research construct. RESULTS: Finally I argue with respect to a recent study in the Canadian journal of Ophthalmology that the research constructs themselves have vague purposes, which only become more concrete insofar as we use them in different circumstances. But this argument suggests that researchers themselves aren't entirely clear about the meaning of a research construct prior to using it in a particular context and thus the individual questions in the questionnaires aren't completely clear. The indeterminacy of the research construct opens up space for respondents and researchers to negotiate the meaning of these questions and ultimately the meaning of the research construct. CONCLUSIONS: If the validity of HRQoL measures is important to ensuring its role as a tool for setting health policy and allocating heath care resources, then we need to allow respondents and researhers alike to ask as well as answer questions.

169/Abstract 1394 Validity Study of an Appropriateness Evaluation Tool by Measuring HRQoL Outcomes Jose M. Quintana, Unidad de Investigacion, Hospital de Galdakao, Galdakao, Vizcaya, Spain, Inmaculada Arostegui, Matematica Aplicada y Estadistica, The University of the Basque Country UPV/EHU, Leioa, Bizkaia, Spain, Susana García, Urko Aguirre, Nerea Gonzalez, Iratxe Lafuente, Research Unit - Clinic Epidemiology, Galdakao Hospital, Galdakao, Bizkaia, Spain, Antonio Escobar, Research Unit - Clinic Epidemiology, Basurto Hospital, Bizkaia, Bilbao, Spain AIMS: Appropriateness explicit criteria are useful, tools for utilization review purposes. But such tools must show their validity previously to be then used for management purposes. The goal of this study was to show the validity of cataract extraction appropriateness explicit criteria developed by this group using RAND methodology. METHODS: This is a prospective observational study of consecutive patients who undergone cataract extraction on 5 different hospitals. All patients completed the SF-36 and VF-14 before the intervention, and 2 months after it. Based on the information provided by the ophthalmologist of each center we were able to judge the appropriateness of each intervention. We also collected information about complications presented up to 2 months after the intervention. Appropriateness was classified in appropriate, uncertain or inappropriate intervention. We study if the HRQoL changes, complications rate, and visual acuity improvement, in each of these 3 categories were different, as a measure of validity of our tool. Statistical Analysis: we used Chi Square test, and ANOVA and Scheffé test for multiple comparisons. RESULTS: We were able to collect data about 2161 patients. Of them, 63.3% were considered appropriate interventions, 25.8 uncertain and 10.8 inappropriate.

171/Abstract 1182 Examining the Relationship Between Global and Domain Measures of Quality of Life by Three Factor Structure Models Chia-huei Wu, Psychology, National Taiwan University, Taipei City, Taiwan (R.O.C.), Grace Yao, Psychology, National Taiwan University, Taipei, Taiwan (R.O.C.) AIMS: Measurements for quality of life can be divided into two categories: global measurements from top-down perspective which regards evaluations in various life domains as resulting from evaluation of life as a whole, and domain-specific measurements

A-124

(t(188)=2.05;p<.05). CONCLUSIONS: Patients presenting binge eating disorder, those who continuously eat or presented night eating syndrome reported lower quality of life, but patients presenting the other specific eating behaviours did not distinguished themselves from patients without these behaviours concerning quality of life.

from bottom-up perspective which regards evaluations in various life domains as contributing to evaluation of life as a whole. Concerns of global measurements and domain-specific measurements are different, although both of them aim to measure individuals_ evaluation of life. The aim of this study was to examine the relationship between global and domain measures of quality of life by three confirmatory factor models. METHODS: 340 students at National Taiwan University participated in this study. They complete the Satisfaction with Life Scale (SWLS; a global measurement) and the WHOQOL-BREF (a domain-specific measurement). Three models were specified to examine the relationships among scores of the SWLS and the WHOQOL-BREF. The first model was a common factor model in which scores of the SWLS and the WHOQOL-BREF were all influenced by a single factor. The second model was a correlated two-factor model in which scores of the SWLS were influenced by a factor and scores of the WHOQOL-BREF were influenced by another factor, and these two factors were allowed to be correlated. The third model was a three-factor model, in which one factor (quality of life) influenced the scores of the SWLS and the WHOQOL-BREF, one factor (global approach) only influenced the scores of the SWLS, and another one factor (domain approach) only influenced the scores of the WHOQOL-BREF. RESULTS: The results showed that the third model was the best model, suggesting that global measures and domain measures did tap the same construct on quality of life, however, the global or domain approach also have substantial impacts on the meaning of scores. CONCLUSIONS: Comparisons of these three models suggested that global and domain measures did not have exactly the same meaning, however, they did tap the same construct.

173/Abstract 1450 Direct-to Patient Recruitment: A Novel Method for Conducting Patient-Reported Outcomes Studies Tiffany B. Miller, Shoshana S. Colman, Hany Zayed, Department of Analytics, Quintiles Strategic Research Services, San Francisco, CA AIMS: To best assess patient comprehensibility and satisfaction with medical device labeling, the FDA recommends pre-testing labeling among patients. Traditionally, doctors' offices served as investigational sites for these types of patient-reported outcomes (PRO) studies. However, PRO studies do not typically involve medication administration, so it is possible to conduct PRO research using alternative recruitment and operational strategies. The case study described here used market research facilities as study sites to measure lay-users' comprehension of and satisfaction with investigational medical device labeling. METHODS: Using participant databases maintained by the market research facilities, 95 diabetes subjects were recruited. Subjects were required to complete various assessments, including satisfaction with labeling questionnaires. The physical set-up of the market research facilities (e.g. observation rooms with one-way mirrors) facilitated the work of clinical monitors, who performed site visits at each facility. Nurse Educators were also on-site to provide medical assistance if necessary. All data were analyzed according to a Statistical Analysis Plan. RESULTS: The direct-to-patient approach offered several advantages such as targeted and rapid recruitment, easy monitoring, and cost efficiencies. Participants recruited by the facilities were committed and reliable, as they had participated in other studies previously. Dedicated study staff at each facility offered flexible appointment schedules for subjects, increasing convenience for participants and study compliance. Measurement of subject satisfaction with labeling met all acceptance criteria, demonstrating that the study design was valid. Clinical monitors also concluded that the study met all quality assurance and control criteria. CONCLUSIONS: This study utilized a novel method for direct-topatient recruitment. The statistical results show that running this type of study using a market research facility-based approach is a viable alternative to the physician, site-based approach.

172/Abstract 1665 Quality of Life and Eating Behaviour in Obese Patients Isabel L. Silva, Human and Social Sciences College, Universidade Fernando Pessoa, Porto, Portugal, Jose L. Pais-Ribeiro, Faculdade de Psicologia e de Ciencias da Educacao, Universidade do Porto, Porto,, Portugal, Helena Cardoso, Endocrinology Department, Hospital Geral de Santo António, Porto, Portugal AIMS: The aim of this study was to analyse the differences in quality of life between patients who present binge eating disorder, night eating syndrome, seasonal eating disorder, sweet eaters and patients without this eating behaviour characteristics. METHODS: Participants: A cohort of 198 obese patients was studied. 50.3% presented binge eating disorder; 14.8% night eating syndrome; 26% seasonal eating disorder; 36.9% eat continuously, 39.5% eating compulsion (sweet or grass eaters) and 78.1% reported yo-yo phenomena. Instruments: Participants answered to the Portuguese version of SF-36. Procedure: Participants answered to the questionnaires in the context of a personal interview, after their informed consent, and clinical data were collected from hospital records. RESULTS: Data analysis suggested that seasonal eating disorder, eating compulsion and patients who experienced yo-yo phenomena did not differ from patients without this eating behaviours concerning quality of life. The present study suggests that patients with binge eating report lower quality of life in the following domains: health transition (t(195)=-2.52;p<.05), role functioningphysical (t(193)=2.69;p<.01) and role functioning emotional (t(194)=3.06;p<.01). Furthermore, patients who continuously eat report lower quality of life in health transition (t(193)=-2.84;p<.01) and vitality (t(192)=2.40;p<.05). Finally, patients presenting night eating syndrome report lower quality of life in health transition (t(194)=-2.71;p<.05), physical functioning (t(192)=2.41;p<.05), vitality (t(193)=2.60;p<.05), social functioning (t(194)=2.88;p<.01), role functioning-emotional (t(193)=3.87;p<.0001) and mental health

174/Abstract 1663 Obesity:Is There a Relation Between Quality of Life and Negative Affect? Isabel L. Silva, Human and Social Sciences College, Universidade Fernando Pessoa, Porto, Portugal, Jose L. Pais-Ribeiro, Faculdade de Psicologia e de Ciencias da Educacao, Universidade do Porto, Porto, Portugal, Helena Cardoso, Endocrinology Department, Hospital Geral de Santo António, Porto, Portugal AIMS: The aim of this study was to analyse the relation between quality of life and negative affect (namely anxiety and depression) in patients with obesity diagnosis. METHODS: Participants: A cohort of 198 obese patients, 84.8% of which were females; aged between 15 and 65 (M=38.86; SD=11.47); and with a body mass index between 30.37 and 100.92 (M=49.03; SD=9.20) was studied. Instruments: Participants answered to the Portuguese versions of SF36 and of the Hospital Anxiety and Depression Scale. Procedure:

A-125

Participants answered to the questionnaires in the context of a personal interview, after their informed consent. RESULTS: Data analysis reveals that health transition, physical functioning, role functioning-physical, bodily pain, general health, social functioning and role functioning-emotional are quality of life domains not significantly related to anxiety and depression levels reported by obese patients. Nevertheless, vitality revealed to be significantly related to anxiety (r(197)=-.15; p<.05) and depression (r(197)=-.16; p<.05) levels, as well as mental health showed to be significantly related to anxiety (r(192)=-.21; p<.01) and depression (r(192)=-.21; p<.01) symptoms. CONCLUSIONS: Anxiety and depression levels proved to be negatively and weakly related to vitality and mental health quality of life domains, however they do not revealed to be significantly related to any other quality of life domain.

presenting for prenatal care at the University of Ghana completed a demographic questionnaire, the SF-12 and the LOT-R. Data were analyzed using SPSS. P<.01 was taken as significant. RESULTS: Data were collected on 101 pregnant women (28% nulliparious; mean age, 29.7; mean weeks gestation, 31.8). 93% were married, and 86.1% worked for money. Mean LOT-R scores were 15.6, ± 3.85. This was significantly higher than comparable U.S. women (p=.004). Overall LOT-R scores were associated with current treatment for emotional issues (p=.001). LOT-R scores were not associated with any other demographic and health-related variable, including age, marital status, pregnancy complications, self-rated difficulty of the pregnancy, or ongoing comorbid conditions. LOT-R scores were also not associated with SF-12 scores. Ghanaian women's SF-12 scores were significantly lower than U.S. norms on physical functioning (p=.001), role physical (p<.001), social functioning (p<.001), role emotional (p<.001), and PCS (p=.001). Ghanaian women's QOL scores were significantly higher than U.S. female population norms on general health perceptions (p<.001) and vitality (p<.001). Note that there were no significant differences on the mental health summary scores and the mental health subscales. CONCLUSIONS: Despite decrements in physical functioning during pregnancy, Ghanaian women are significantly more optimistic than U.S. women. Although LOT-R scores were associated with current treatment for emotional issues, SF-12 mental health subscale and summary scores were not significantly different across cultures. This suggests that the concept of O/P is independent from mental health status in Ghanaian women. Future research needs to explore the link between O/P and health outcomes in Ghanaian and U.S. women.

175/Abstract 1294 Factors Affecting HRQOL in Diabetic Outpatients Hisako Adachi, Adult and Gerotological Nursing, University of Gifu, Gifu, Gifu, Japan, Takaaki Oyamada, Early Childhood Studies, Gifu Women's University, Gifu, Gifu, Japan AIMS: THis study aims to investigate the factor that affects HealthRelated Quality of Life(HRQOL) in diabetic outpatients. METHODS: 112 outpatients with diabetics participated in this study after they signed the imformed consent form. Their average age was 55.7±13.7 years old, the mean HbA1c% was 7.0±1.0 and the mean duration of diabetes was 13.4±10.5 years. The questionnaire and time trade-off method were used. The questionnaire consisted of 14 items(5 point Likert-type scale) on the reason that outpatients with diabetes would trade their remaining life with disease for less short life without disease, or do not even if possible. RESULTS: (1)The principal-component factor analysis with Varimax rotation was applied to date of 112 patients for 14 items, and 4 factors separated out, accounting for 50.3% of the variance. (2)Cronbach's alpha coefficient value for each item was 0.65-0.75. (3)4 factors were named: loneliness factor from restricted social activities(factor 1), getting worse factor(factor 2), hopeless future factor from physical and psychological troubles in daily life with diabetes(factor 3), and negative feeling factor from severe diet control(factor 4). (4)On results of time trade-off method, 42 patients(37.5%)(Trade Group) wished to trade their remaining life with disease(diabetic state) for less short life without disease(fully healthy state), however, 70 patients(62.5%)(No Trade Group)did not wish. (5)Mean scores of factor 2 and factor 4 of Trade Group were significantly higher than those of No Trade Group(p<.01). CONCLUSIONS: Those results show that 4 factors affect HRQOL in diabetic outpatients, and are important to the psychological care of diabetes.

177/Abstract 1757 Sexual Health and Health-Related Quality of Life in Belgium, Italy, Portugal and Spain Osvaldo R. Santos, Statistics, Euroconsumers, Lisboa, Portugal, Guy Sermeus, Peter Kupers, Statistics, Euroconsumers, Bruxelles, Belgium, Marco Anelli, Statistics, Euroconsumers, Milano, Italy, Pedro Ortiz, Statistics, Euroconsumers, Madrid, Spain AIMS: Sexual health is a main dimension of health. However, association between gender-specific sexual health and health-related quality of life (HRQoL) is understudied. Main goals of this study were (a) to get representative and nationwide data on self-reported sexual health measures and on self-reported sexual dysfunction and (b) to evaluate the association between gender-specific sexual health and several dimensions of HRQoL. METHODS: Postal selfadministered questionnaire survey with probabilistic representative samples of four national populations. For the assessment of sexual health, it has been used (a) the Index of Sexual Satisfaction, (b) the Sexual Health Inventory for Men and (c) the Female Sexual Function Index. HRQoL was assessed through a self-perceived general health single-item measure (the first question from SF-36) and through the Nottingham Health Profile. RESULTS: In total, 10124 adult people (between 18 and 74 years of age) participated. Most prevalent sexual problems were: among women, lack of libido (30,6%), difficulties in lubricating (25,1%) and difficulties with orgasm (18,7%); among men, premature ejaculation (24,2%), erectile dysfunction (19,1%) and performance anxiety (15,5%). After adjusting for several demographic and life-style variables (through multiple regression analysis), significant associations were found, in every country, between all indicators of health-related quality of life and all indicators of sexual health for both men and women (stronger correlations were found among men than among women). CONCLUSIONS: Collected data allowed a multi-national representative study of the association between sexual health and HRQoL. Results support significant association between sexual

176/Abstract 1722 Optism/Pessimism and HRQoL IN Ghanaian Pregnant Women Cheryl A. Moyer, Global REACH, Geraldine Ekpo, UMMS, Cecilia Calhoun, MHIRT, Jonathan Greene, UMMS, Sujata Naik, OB-GYN, Emily Sippola, David T. Stern, Global REACH, University of Michigan, Ann Arbor, MI, Richard Adanu, Isaac Koranteng, E Y. Kwawukume, OB-GYN, University of Ghana, Accra, Ghana, Frank J. Anderson, OB-GYN, University of Michigan, Ann Arbor, MI AIMS: Cross-cultural variations in optimism/pessimism (O/P) and its potential influence on HRQOL are little studied. This research aimed to: 1) Determine O/P among Ghanaian pregnant women; 2) Compare O/P scores to U.S. norms; and 3) Determine if O/P is associated with current health status and/or HRQOL. METHODS: Pregnant women

A-126

(T1), 1 year later (T2) and 4-5 years after T2 (T3). So far, 50 subjects completed all follow-ups. QOL was assessed with WHOQOL-Bref and Social Functioning Scale (SFS). Symptoms were evaluated with PANSS at T1. Information on prehospital factors: age at onset of problems with functioning, duration of impaired functioning (DIF), length of psychotic symptoms (LPS), duration of untreated psychosis (DUP) was collected during interviews. In this paper we give attention to only relationships between early factors and QOL at T 3. Mann-Whitney test and Spearman's coefficient were used for bevariate analyses. Multiple regression was conducted to test predictive values of early factors on SFS-total and WHOQOL domains that showed significant correlations with independent variables. RESULTS: Significant correlations were found between SFS and WHOQOL domains and DIF (SFS: total score, recreation, independence-competence, employment; WHOQOL: environment), LPS (SFS: total, withdrawal, social, recreation, independencecompetence, employment; WHOQOL: social relationships), PANSS T1 (5 SFS domains and 3 WHOQOL domains). Lower scores were found for patients with longer DUP (SFS: total, withdrawal, social, recreation; WHOQOL: physical, environment). The regression analysis showed that for SFS, PANSS T1 and LPS contributed independently accounting for 24% of explained variance. WHOQOL: physical and environmental domains were influenced by PANSS T1 (9%) and social relationships by DUP (9%). CONCLUSIONS: Early factors are still noticeable determinants of OQL 5-6 years after a first hospitalization. A large proportion of variance in OQL was unexplained and likely influenced by other, mainly current factors.

health and physical, psychological and social dimensions of HRQoL. These findings reinforce the importance of a biopsychosocial assessment of sexual dysfunctions in both causal and maintaining variables. Likewise, HRQoL measures should be included in the assessment of sexual health and in the evaluation of sexual-related treatment effectiveness.

178/Abstract 1333 Life Quality in Psychotics Patients Maribel Peró, Metodologia de las Ciencias del Comportamiento, Universidad de Barcelona, Barcelona, Spain, Barcelona, Montserrat Freixa, Departament de Metodologia de les Ciències del Compo, Victoria Carreras, Metodologia de les Ciencies del Comportament, Universitat de Barcelona, Barcelona, Spain, Spain, Josep Maria Otin, Barcelona, Spain, Spain, Joan Pi, Hospital de dia Llúria _CPB - SSM, Barcelona, Spain, Spain AIMS: The objective of this paper is study the possible relationship between the quality of life with demographic variables, the attitudes of the relatives, the time of stay in the center and the antecedents of mental disorderes in the relatives. METHODS: We have worked with a sample of 188 psychotics, according to the criteria DSM-IV, from the Barcelona's LLúria hospital. The evaluation of the quality of life of these patients has been obtained with the QLS (Heinrichs, D.W.; Hanlon, T.E., Carpenter, W.T., 1984). The scale items have been extracted from the fundamental manifestations of the schizophrenia's deficit syndrome that, conceptually, are grouped in four categories: 1. intrapsychic basis, 2. interpersonal relations, 3. instrumental role and 4. objetcs and common activities. Heinrichs, D.W.; Hanlon, T.E., Carpenter, W.T., (1984). The Quality of Life Scale. Schizophrenia Bulletin, 10 (3), 388-398. RESULTS: The results point at the existence of statistically significant differences in the category intrapsychic basis of the QLS between the patients in whom the familiar attitude is with conscience of support with overinvolvement with respect to the group of familiar attitude with conscience of supporting without overinvolvement (t = 2,353; p = 0.021), being greater the score in which overinvolvement exists. On the other hand also can be found statistically significant differences in the instrumental category roll between the patients who have familiar mental upheaval record and those that not, in the QLS (t = 2,106; p = 0.037) and in the total score in QLS (t = 2,130; p = 0.035). In both cases the average is higher in the group with familiar mental upheaval record than in the group that does not have familiar antecedents. CONCLUSIONS: The higher average in the group with familiar mental upheaval record than in the group that does not have familiar antecedents in the intrapsychic basis and in the instrumental category roll is a fact that can be contemplated in the therapy with psicotic patients.

180/Abstract 1047 Time-Course of Health-Related Quality of Life of Schizophrenic Patients After a First Psychiatric Hospitalization. Preliminary Results. Krystyna Jaracz, Krystyna Gorna, Neurological and Psychiatric Nursing, University of Medical Sciences, Poznan, Poland AIMS: Longitudinal studies of quality of life (QOL) of schizophrenic patients are inconsistent. This inconsistency may be partly explained by a relatively short time of the observation. The aim of this study was to examine the time-course of subjective and objective QOL in first-episode schizophrenic patients and to investigate the relationship between changes in QOL over time and changes in psychopathological status. METHODS: A cohort of 86 patients was observed at 1 month after a first hospitalization (T1), 12 months later (T2) and 4 - 5 years after T2 (T3). At discharge all subjects met the diagnostic criteria for schizophrenia (ICD-10). A comparative group consisted of 86 healthy subjects matched according to sex and age. A total of 50 interviewed subjects (17 women, 33 men; mean age 26.3) completed the follow-up. Subjective QOL was assessed by 4-domain WHOQOL - Bref and objective QOL by 7-domain Social Functioning Scale (SFS). The psychopathological symptoms were evaluated with Positive and Negative Syndrome Scale (PANSS). Nonparametric tests were used for statistical analyses. RESULTS: At T3 objective and subjective QOL was low compared with the healthy group. The total sore in the SFS improved throughout the observation period (102, 104, 104.9, respectively; p<.05). However, higher mean scores were seen only in 2 domains: Independence-Performance (p<.05) and Interpersonal (p<.05). The lower scores were obtained for Employment (p<.05). The WHOQOL four domains remained stable. The psychopathology symptoms showed a significant deterioration, especially between T2 and T3 (p<.01). There was a negative relationship between T1 and T3 changes in SFS and PANSS (r=.51, p<.01). CONCLUSIONS: These results show that subjective QOL and most areas of the objective QOL did not increase over the observation period.

179/Abstract 1259 Early Predictors of Quality of Life in First Episode of Schizophrenia Krystyna Gorna, Krystyna Jaracz, Neurological and Psychiatric Nursing, Janusz Rybakowski, Adult Psychiatry, University of Medical Sciences, Poznan, Poland AIMS: Majority of studies concentrated on premorbid factors that may influence the outcome of schizophrenia have been retrospective. The aim of this study was to investigate pre-hospital and early clinical factors that could affect patients' quality of life (QOL) from a longitudinal perspective. METHODS: A group of 86 schizophrenic patients has been observed at 1 month after a first hospitalization

A-127

city of Gothenburg were included in the study group (50 patients) and those living outside the city were included in the control-group (63 patients). Nutritional status was assessed with weight, body mass index (BMI) and BIA (bioelectrical impedance). HRQL was assessed with the EORTC QLQ-C30 and QLQ-H&N35 at diagnosis, 1,2,3, 6, 12 and 36 months after treatments start. RESULTS: Results: Large individual variations regarding weight loss, BMI and FFM (fat-free mass) and HRQL were found at the different measurement points in both groups. However, no large differences were found when the study and control group were compared. HRQL was, as expected, worse after three months and thereafter gradually improved to the same levels as at diagnosis, except for problems with dry mouth that remained increased. A significant correlation between weight loss and loss of appetite and pain was found in the study group and between weight loss and problems to swallow food in the control group. The impact of counselling by a dietician was difficult to evaluate since the study was not randomized and it turned out that the study group contained more patients with advanced cancer treated with combined treatment compared to the control group. CONCLUSIONS: Conclusions: For both patients groups weight, FFM and HRQL scores decreased during treatment and then gradually improved to similar levels as at diagnosis. A correlation was found between weight loss and loss of appetite, pain and problems swallowing solid food.

Psychopathological status seems to be an important determinant of this lack of improvement. The results point to the role of early pharmacological and psychosocial interventions in schizophrenia.

181/Abstract 1212 Symptoms, Supportive Care Needs, and Function in Cancer Patients: How are They Related Claire F. Snyder, Medicine, Amanda L. Blackford, Elizabeth GarrettMayer, Julie R. Brahmer, Michael A. Carducci, Antonio C. Wolff, Oncology, Johns Hopkins School of Medicine, Baltimore, MD, Danetta E. Hendricks, Health Policy & Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, Kristina R. Weeks, Medicine, Johns Hopkins School of Medicine, Baltimore, MD, Sydney M. Dy, Albert W. Wu, Health Policy & Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD AIMS: To understand the relationships among symptoms, supportive care needs, and functional status. METHODS: 88 cancer patients (45 breast, 43 prostate, mean age 60, 77% White, 94% ECOG 0-1) completed the Supportive Care Needs Survey-34 and EORTC QLQC30. The associations among EORTC function (physical, role, emotional, cognitive, social, global health/QOL), EORTC symptoms (fatigue, nausea/vomiting, pain, dyspnea, insomnia, appetite loss, constipation, diarrhea), needs (physical & daily living, psychological, patient care & support, health system & information, sexual), age, tumor type, and ECOG status were tested. We constructed 2 multivariate item regression models: 1) function described by symptoms and patient characteristics and 2) function described by needs and patient characteristics. Needs were added to Model 1 and symptoms were added to Model 2 in step-wise fashion to determine their added explanatory value. A linear mixed effects model accounted for the correlation among the six function scores within patients. RESULTS: Age and ECOG were associated with function (p<.05). Fatigue, appetite loss, and pain were associated with lower function (p<.05). Psychological was the only needs domain to add information explaining function beyond the symptoms and patient characteristics (p<.05). Increased physical & daily living and health system & information needs were associated with lower function (p<.05). Fatigue, appetite loss, nausea/vomiting, dyspnea, pain, and diarrhea each added information beyond the needs and patient characteristics (p<.05). The needs domains were no longer significant when fatigue was added to the model, and only health system & information remained significant when pain was added to the model. This is likely due to the strong correlations between and among the symptoms and needs domains. CONCLUSIONS: Symptoms and supportive care needs of cancer patients are both significantly related to function. Research using larger samples can further dissect their main and interacting effects on function.

183/Abstract 1651 Quality of Life in Advanced Oral and Oropharyngeal Cancer Karin Petruson, Otorhinolaryngology Head and Neck Surgery, Claes Mercke, Oncology, Eva Hammerlid, Otorhinolaryngology Head and Neck Surgery, Sahlgrenska University Hospital, Gothenburg, Sweden AIMS: The purpose of this study was to longitudinally evaluate health-related quality of life (HRQL) in patients with advanced (stage III-IV) oral and oropharyngeal cancer treated with external radiotherapy (ERT) and brachytherapy (BT). METHODS: 73 patients with stage III-IV oral and oropharyngeal cancer were recruited to the study, and depending on chemotherapy response divided into two treatment groups; high ERT (64.6 Gy) + low BT (<20 Gy) or low ERT (40.8 Gy) + high BT (>22 Gy). HRQL was assessed with the EORTC QLQ-C30 and QLQ-H&N35, at diagnosis, 3,6,12 and 36 months after treatment start RESULTS: Differences in HRQL scores between the groups during the study were minor. After three years however 90% of patients treated with high ERT, and 75% of patients treated with low ERT reported clinically significant problems with dry mouth and 65% of patients treated with high ERT, and 40% of patients treated with low ERT reported clinically significant problems with swallowing solid food. CONCLUSIONS: Patients with advanced oral and oropharyngeal cancer responding well to neoadjuvant chemotherapy may be offered a low ERT dose with a BT dose exceeding 22 Gy with the probable advantage of smaller irradiated volume, without jeopardizing HRQL. However problems with dry mouth and swallowing solid food are still severe three years after treatment

182/Abstract 1643 Health-Related Quality of Life and Nutritional Support in Head and Neck Cancer Ewa Silander, Otorhinolaryngology Head and Neck Surgery, Sahlgrenska Unversity Hospital, Gothenburg, Sweden, Karin Petruson, Eva Hammerlid, Otorhinolaryngology Head and Neck Surgery, Sahlgrenska University Hospital, Gothenburg, Sweden

184/Abstract 1217 The Imact of Cosmetic Rehabilitatio Program on the Body Image and Quality of Life of Oral Cancer Patients Hsueh-Erh Liu, Nursing, Chang Gung University, Tao Yuan, Taiwan, ROC., Taiwan, ROC., Sharen Huang, Genomic medicine, National clinical core for genomic medicine, Taipei, Taiwan, ROC., Taiwan, ROC.

AIMS: Aims: To explore if individual nutritional support during and after treatment, might prevent weight loss and improve health-related quality of life (HRQL) in patients with head and neck cancer. METHODS: Methods: 113 patients with newly diagnosed, untreated head and neck cancer were recruited to the study. Those living in the

A-128

AIMS: The purposes of this study include: 1). To identify the impact of cosmetic rehabilitation on body image and quality of life of oral cancer patients who had completed flap reconstruction. 2). To identify the associated factors with the changes of body image and quality of life. METHODS: A quasi-experiment with longitudinal follow-up research design was conducted. Patients were recruited from the OPD at a medical center in the northern part of Taiwan. They were distributed into the experimental group or control group by their personal preference. The experimental group received instruction of cosmetic rehabilitation whereas the control group received regular nursing care. A total of 24 subjects completed 3 follow-ups (admission, 6-week, 12-week) in each group. Instruments included personal demographic data form, MBSRQ-AS, EORTCCore, and EORTC-H&N. RESULTS: The results showed that: 1). Compared with the control group, subjects in experimental group reported no significant changes of body image during each follow-up. Meanwhile, single item comparison found that the experimental group reported significant improvements in the items as "I like the way I look now" , " most people thinks that I am good-looking", and "face". 2). Compared with the control group, subjects in experimental group reported no significant changes of quality of life during each regular assessment. Meanwhile, single item comparison found that the experimental group reported significant improvements in item as "I feel distress by my appearance".. CONCLUSIONS: Results of present study indicated that cosmetic rehabilitation improved personal appearance in both the eyes of oral cancer patients adn the others'. In order to help patients improve their self-confidence of personal appearance, integrated this service into daily postoperative care is strongly suggested.

having higher levels of disability and symptoms. Proxies only provide valid and useful information about cancer patients QoL for QLQ-C30 domains where the correlations were good, however, should be used with caution particularly for physical, emotional, cognitive and social functioning, pain, insomnia, diarrhoea, global QoL and total QoL score, where de means differences were statistically significant. It will be necessary to increase our sample in order to perform this analysis by diagnosis.

186/Abstract 1092 Relationship Between Disease Specific Functions and Utility Scores on Prostate Cancer Fumitaka Shimizi, Katsuki Fujino, Yoichi M. Ito, Biostatistics, Takashi Fukuda, Drug Policy And Management, University Of Tokyo, Tokyo, Jp, Japan, Makoto Fujime, Urology, Juntendo University, Tokyo, Jp, Japan, Yasuo Ohashi, Biostatistics, University Of Tokyo, Tokyo, Jp, Japan AIMS: To assess the relation between the disease specific functions and utility scores before conducting decision analysis. METHODS: Between October 2004 and September 2005, 323 patients received radical prostatectomy, external beam radiotherapy, brachytherapy, hormonal therapy or a combination of these for localized prostate cancer at three institutions were asked to answer the several selfadministered questionnaires to assess generic QOL (EQ-5D and SF36), disease specific QOL (UCLA-Prostate Cancer Index, hormonal function of Expanded Prostate cancer Index Composite, and International Prostate Symptom Score), and co-morbidities (Index of Co-existent Disease). Community weighted utility scores were obtained from EQ-5D and SF-36, while patient_s directly elicited utility score was obtained using time trade-off (TTO). Univariate analysis was performed to examine a linear trend between the disease specific functions, age, co-morbidities, and utility scores derived from these three methods. In order to assess these relations simultaneously, we applied the general linear model accommodating for the intra-patient correlation of utility scores. RESULTS: Mean age was 71.5 years old (SD: 6.0). Mean utility score derived from EQ-5D, SF-36 and TTO was 0.90 (SD:0.15), 0.74 (SD:0.08) and 0.89 (SD:0.15) respectively. In the univariate analysis, all the variables except age and sexual function had a linear trend with each utility score. In the general linear model, bowel function (p<0.001) and hormonal function (p<0.001) were strongly related to the utility score derived from TTO as compared to EQ-5D and SF-36, while comorbidities were related to the utility scores derived from EQ-5D and SF-36 as compared to TTO. CONCLUSIONS: These results help us to grasp which disease specific function has an impact on overall QOL in patients with prostate cancer. We suggest that, among the three methods for measuring utility scores, TTO is the most responsive tool for reflecting the disease specific functions.

185/Abstract 1414 Cancer Patients Quality of Life, Measured by Themselves and Proxy Alexandra Oliveira, Medical Oncology, Hospital de São Sebastião, Santa Maria da Feira, Portugal, Francisco L. Pimentel, Secção Autónoma Ciências Saúde, Universidade de Aveiro, Aveiro, Portugal, Augusta Silveira, FCS, Universidade Fernando Pessoa, Porto, Portugal AIMS: Health Related Quality of Life (HRQoL) is one of the most important cancer patient outcomes. Most of the information that health care providers have is given by the patient, however proxy can also provide important data about the patient, so is necessary to knew the reliability of that. The aim of this study was to examine whether proxies can provide useful information on HRQoL of cancer patients. METHODS: We used EORTC QLQ-C30 in a sample of 330 cancer patients with different diagnosis and 330 proxys from oncology department of Hospital de São Sebastião. RESULTS: Sample was composed by: patients, 157 female (47,6%), 173 male (52,4%), age average of 62,9 (SD=12,5); proxys, 230 female (69,7%), 100 male (30,3%), age average of 50,6 (SD=14,4).The most frequent diagnosis, that are 78,4%, were colon, gastric, breast, prostate and lung cancer. We found moderate to good agreement (ICC and Pearson correlation ranged from 0.53 to 0.80) between responses from patients and those from their proxies for the QLQ-C30 scales. The results were moderate for emotional, cognitive and social functioning, fatigue, nausea and vomiting, dyspnoea, insomnia, constipation, diarrhoea, financial difficulties and global QoL, and were good for physical functioning and role, pain appetite loss and for total QoL score. Statistically significant differences were found for physical, emotional, cognitive and social functioning, pain, insomnia, diarrhoea, global QoL and total QoL score. CONCLUSIONS: In general, proxies rated the patients as having a more impaired QoL than the patients themselves. Significant other rated patients as

187/Abstract 1616 Utility Mapping or Neuropathic Pain Using the Health Utility Index Mark 3 (HUI 3): A Spanish Perspective Javier Rejas, Health Outcomes Research Department, Medical Unit, Pfizer España, Alcobendas (Madrid), Spain, Concepción Pérez, Pain Unit, Hospital de la Princesa, Madrid, Spain, Rafael Gálvez, Pain&Palliative Pain Unit, Hospital Virgen de las Nieves, Granada, Spain, Silvia Huelbes, Experimental Neurology, Hospital de Parapléjicos de Toledo, Toledo, Spain, Joaquín Insausti, Pain Unit, Hospital Severo Ochoa, Leganés (Madrid), Spain, Silvia Díaz, Health Outcomes Research Unit, Artac Bioestudios, Madrid, Spain

A-129

AIMS: This research was carried out to ascertain utilities for patients suffering from chronic pain of broad Neuropathic or Nociceptive origin. METHODS: We surveyed patients with chronic Neuropathic (NeP) or Nociceptive (NoP) pain. In addition to demographic data, patients completed the Short-Form McGill Pain Questionnaire (SFMPQ). Utilities were determined by means of HUI 3. Five-point ordinal Present-Pain-Intensity item (PPI) and the 0-100 mm-VisualAnalog-Scale (VAS) both from the SF-MPQ were used to classified pain severity at present and on average during the past week. According to PPI, pain was classified as mild, discomforting, distressing, horrible and excruciating, while the past week average pain was segmented as mild (<40), moderate (>40&<70) and severe (>70). Diagnosis was made by expert in pain management using standard clinical criteria. Descriptive statistics, ANOVA and ANCOVA were applied. RESULTS: A total of 322 patients (59.1% women, 57.8% NeP, 42.2% NoP) were surveyed [mean (SD) age; 61.4+15.5 years]. Most patients reported mild (31.3%), discomforting (32.6%) or distressing (24.4%) present pain, while 7.7% scored the pain as horrible and 3.8% excruciating. Mean pain severity during last week was mild in 7.4%, moderate in 39.2% and severe in 53.4%. Utility scores negatively correlated with pain severity, both at present and during last week; the higher de pain the lower the utility: r=-0.30, p<0.001; F=9.7; p<0.001 (0.476, 0.481, 0.315, 0.118 and 0.226) and r=-0.39, p<0.001; F=27.4, p<0.001 (0.705, 0.497 and 0.286) for PPI and last-weekaverage pain classifications respectively. After adjusting by age, sex and severity of pain, HUI 3 scores were similar in NeP and NoP patients: 0.353 (0.291-0.416) vs 0.334 (0.192-0.476); F=0.06, p=0.808. CONCLUSIONS: Lower utility scores were significantly associated with severity of pain both at present and during last week: the higher the severity of pain the lower the utility scores. No significant differences were observed between types of pain.

avoided because pain signifies damage were correlated with overall QoL (r=0.35,p=0.04), psychological domain (PSD)(r=0.4,p=0.03), level of independence (LID) (r=0.4,p=0.02),social relations(SRD)r=0.4,p=0.02), environment(ED)(r=0.4,p=0.03) and spirituality,religion and personal beliefs(r=0.4,p=0.03).PSD domain was correlated with PE(r=-0.4,p=0.01).LID was correlated with BAMMCP(r=-0.44,p=0.01) and with MCP(r=-0.4,p=0.02).ED was correlated with the beliefs that others should be solicitous when the patient experiences pain (r=-0.34,p=0.04).SRD was correlated with MCP(r=-0.4,p=0.02). CONCLUSIONS: Results suggest that some pain beliefs and attitudes can affect negatively the QoL of chronic pain patients. Thus, beliefs should also be treated so that one can improve QoL and pain.

189/Abstract 1373 Assessing Pain in the Elderly: Obstacles and Measurement Tools Elizabeth G. Cooperman, Buehler Center on Aging, Northwestern University Feinberg School of Medicine, Chicago, IL, Sheila M. Perry, Chih-Hung Chang, Buehler Center on Aging, Northwestern University Feinberg School of Medicine, Chicago, Illinois AIMS: Identifying the barriers to evaluating and treating pain in the elderly is central to the discussion of pain measurement tools. Cognitive and communication impairment, sensitivity to pharmacological treatment, and healthcare professionals' misconceptions about pain and aging can hinder adequate pain control in the geriatric population. This study explores pain assessment and management issues specific to the elderly through systematic literature review. It focuses on instruments that have been validated for use with elders and on those designed specifically to measure pain in this population. METHODS: Relevant articles were located using a key word search of electronic medical and nursing journal databases. The key words included: pain assessment or pain measurement AND aged, elders or aging. They were entered into databases Ovid Medline, CINAHL and PsychINFO and were then reviewed one-by-one for relevance. All articles were published between 1996 and 2005. RESULTS: The search returned 94 articles, 28 of which were selected as highly relevant to this study. Within these, 16 different pain measures were evaluated for use with the elderly, the most commonly cited being the McGill Pain Questionnaire, the Visual Analog Scale and the Verbal Descriptor Scale. Three measures in this subset were designed expressly for elders; the Geriatric Pain Measure addresses the need for multidimensional pain assessment, while the Checklist of Nonverbal Pain Indicators and the Discomfort Scale for Dementia of the Alzheimer's Type rely on observable behavior/gesture, rather than self-report, to assess pain in cognitively impaired elders. CONCLUSIONS: Inadequate pain management persists among the elderly and can have serious consequences on their functioning, quality of life, mood and cognitive condition. Though some pain assessment tools have been validated for use with the elderly and a handful have been developed to overcome specific barriers to managing elders' pain, more work is still needed to confront the challenges of accurate and timely assessment of pain and optimal pain management in the elderly.

188/Abstract 1275 Attitudes Towards and Beliefs AboutPain and Quality of Llife of Chronic Pain Patients Karine A. São Leão Ferreira, School of Nursing, University of São Paulo, São Paulo, São Paulo, Brazil, Analu B. Shiota, Multidisciplinary Center of Pain, General Hospital of University of São Paulo, São Paulo, São Paulo, Brazil, Juliano d. Santos, School of Nursing, University of São Paulo -Brazil, São Paulo, São Paulo, Brazil, Manoel J. Teixeira, School of Medicine, University of São Paulo, São Paulo, São Paulo, Brazil AIMS: Although reports suggest that beliefs about pain play an important role in the adjustment of chronic pain patients,the impact of beliefs on the quality of life(QoL) has not been described yet.We aimed to assess the correlation between attitudes towards and beliefs about pain and QOL of chronic pain oupatients. METHODS: We conducted 36 interviews with patients attending in a Multidisciplinary Center of Pain,Brazil,using the brief version of the Survey of Pain Attitudes(SOPA-B) and the WHOQOL-100 to assess QoL. RESULTS: Mean age was 50.8 years(sd=9,8),92.7% were female,30.6% were married and 55.6% had as many as 8 years of education.The mean scores of SOPA-B domains were: feelings about pain control =1.84(sd=1.0,range=0-3.8),solicitude=2.12(04),medication=3.04(0.3-4.0),pain-related disability=2.5(1.0-4.0),pain and emotions(PE)=3.31(0.25-4.00),medical cures for pain(MCP)=3.04(0.33-4),and pain-related harm= 2.02(0.0-3.7).The most adjusted attitude domain was PE and the most unadjusted was the MCP.There were negative correlations between the physical domain of QoL(PYD) and PE(r=-0.42,p=0.01), belief in the appropriateness of medications for managing chronic pain(BAMMCP)(r=-0.40,p=0.02) and that there was a MCP(r=-0.42, p=0.01).The beliefs that one was disabled and that activity should be

190/Abstract 1237 Impact of Chronic Back Pain on Health-Related Quality of Life in a Representative Sample of Adults in Germany Measured with the SF-8 Ute Ellert, Hannelore K. Neuhauser, Epidemiology and Health reporting, Robert Koch Institute, Berlin, Germany

A-130

pain and its management, most of which is geared towards patients. Patients should be taught how to distinguish important information in order to be able to effectively manage their pain.

AIMS: To examine the association between chronic back pain in the last 12 months and health-related quality of life (HRQOL) measured with the 8-item short form questionnaire (SF-8) in a representative sample of the adult population in Germany. The SF-8 is a new generic quality of life instrument, which provides results comparable to those of the widely used SF-36, but is shorter and therefore more suitable for general health surveys. The German version of the SF-8 has not been used in population health surveys and the impact of chronic back pain on HRQOL has not been investigated with the SF8 so far. METHODS: The SF-8 standard version (4-week recall) has been used to measure HRQOL in a computer-assisted telephone interview (CATI) health survey of a representative sample of the adult residential population in Germany (n=8318, age 18-96 years). Chronic back pain was defined as "three months and more continuous back pain, namely almost daily". The influence of chronic back pain on health-related quality of life was analyzed using linear regression models with age, sex, social class and 17 illnesses as independent variables. Co-morbidities were included as self-reported medical diagnoses (only exception: depression as self-reported illness). RESULTS: The 12-month-prevalence of chronic back pain was 22% in women and 15% in men. Participants with chronic back pain reported significantly lower HRQOL in all domains compared to those without chronic back pain. Men with chronic back pain rated their HRQOL in the pain scale and in all mental scales significantly higher than women with chronic back pain. In multivariate analysis the negative association between chronic back pain and HRQOL was significant in all scales and in both summary measures (standardised ß between -.1 and -.3). Sex had no significant effects in this model. CONCLUSIONS: The ascertained associations are of a similar order of magnitude as in other studies. These results suggest that the SF-8 is a useful tool for measuring HRQOL in women and men with chronic back pain.

192/Abstract 1406 Validation of the Taiwan Version of the WHOQoL-BREF Instrument in Patients with COPD Ya-Lin Hsu, Ya-Wei Shiao, Institute of Environmental Medicine, China Medical University, Taichung, Taiwan, R.O.C., Jian-Jung Chen, Chinese Medicine, China Medical University Hospital, Taichung, Taiwan, R.O.C., Wen-Miin Liang, Institute of Environmental Medicine, China Medical University, Taichung, Taiwan, R.O.C. AIMS: Chronic obstructive pulmonary disease (COPD) has become one of the leading causes of morbidity and mortality in industrialized and developing countries. The measurement of quality of life in this population has become an important issue. This study aimed to assess the validity of the WHOQOL-BREF questionnaire in patients with COPD. METHODS: The study group comprised 146 male patients with COPD whose ages ranged from 50.4 to 87.6 years. Healthrelated Quality of life (HRQOL) of patients was assessed by the Taiwan version of the WHOQOL-BREF questionnaire. The itemscale correlations were used to assess the convergent and discriminant validity. Cronbach's alpha, exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were used to evaluate the internal consistency and construct validity. RESULTS: All domains showed no ceiling and floor effect. Internal consistency revealed by the Cronbach's alpha coefficient ranged from 0.75 to 0.81 in the four domains. The success rate for convergent validity in the physical, psychological, social, and environmental domains were 100%, 100%, 75%, 89%, respectively. The success rates for discriminant validity in the physical, psychological, social, and environmental domains were 96%, 100%, 100%, 100%. In EFA, 4 factors emerged with eigenvalues above 1, accounting for 84% of the variance. Testing of the four domain structure of the WHOQOL theoretical constructs by CFA showed that there was not a good fit. The modified model was suggested based on the modification indices. CONCLUSIONS: Overall, the WHOQOL-BREF is suitable for measuring HRQOL among patients with COPD. EFA and CFA were performed not only to explore and confirm the factor constructs of WHOQOL model but also to help us better understand the factor constructs of our data.

191/Abstract 1425 Printed and On-Line Resources for Patients in Pain Management Sheila M. Perry, Elizabeth G. Cooperman, Chih-Hung Chang, Buehler Center on Aging, Northwestern University, Chicago, IL AIMS: Chronic pain is costly in both human and economic terms. The American Pain Foundation (APF) reports over 50 million people suffer from chronic pain and it costs the government about $100 billion annually. Helping patients learn to better manage their pain by using printed and/or electronic materials can reduce this cost. METHODS: On-line searches were conducted using the keywords pain management and video. We identified nine other related Websites from references or related links of three primary sources. Paper materials were obtained from the APF. One CD-ROM and three videos were found at Websites of Jones & Bartlett Publishers, Med Films Inc., Pain.com, City of Hope, and C-Health. Internet materials were gathered from the Websites of APF, StopPain.org, University of Michigan Health System, City of Hope, National Cancer Institute, CancerSupportiveCare.com, WebMD, and HealthTouch.com. RESULTS: Paper materials include a newsletter; pain care bill of rights; four brochures; and information on pain medicine. The CD-ROM and videos address problems with pain relievers, pain's impact on sleep, pain management, and gives patients a guide to pain management websites. Webpages from StopPain.org, the APF, and City of Hope were found to be the most inclusive sources. StopPain.org covers pain types, assessments, treatments, management, and recommendations. The APF stresses the use of a pain diary and describes the benefits of chair yoga. City of Hope's website lists educational tools including kits, books, instructional guides, and informational sheets. CONCLUSIONS: There is a wealth and diversity of information available regarding

193/Abstract 1412 Quality of Life Predicts Frequency of Exacerbations in Patients with COPD Jian-Jung Chen, Chinese Medicine, China Medical University Hospital, Taichung, Taiwan, R.O.C., Yao-Ching Wang, Institute of Environmental Medicine, China Medical University, Taichung, Taiwan, R.O.C., Liang-Wen Hang, Te-Chen Hsia, Chest Medicine, China Medical University Hospital, Taichung, Taiwan, R.O.C., WenMiin Liang, Institute of Environmental Medicine, China Medical University, Taichung, Taiwan, R.O.C. AIMS: Exacerbations occur commonly in patients with moderate or severe chronic obstructive pulmonary disease (COPD) but if there is association between quality of life and their frequencies were not clear. We sought to determine whether quality of life could predict the frequency of exacerbations for patients with COPD. METHODS: A total of 148 male patients aged 50 to 88 years diagnosed with COPD were interviewed in person using both the WHOQOL-BREF

A-131

for comparisons among various populations. Prior to using a generic HRQL instrument, the measure should be validated within the target population. The purpose of this study was to evaluate two generic HRQL measures in a sample of patients with type 2 diabetes. METHODS: Patients with type 2 diabetes in Scotland and England completed two generic HRQL measures: the EQ-5D and the Psychological General Well-Being Index (PGWB). Two diabetesspecific measures were administered: the Appraisal of Diabetes Symptoms (ADS) and the Diabetes Symptom Checklist-Revised (DSC-R). Psychometric analysis examined the performance of the EQ-5D and PGWB. RESULTS: There were 130 participants (53 Scotland; 77 England; 64% male; mean age = 55.7 years). Responses on the EQ-5D and PGWB reflected a moderate degree of impairment. Mean scores were consistent with previous diabetes samples: EQ-5D Index score, 0.75; EQ-5D VAS, 68.8; PGWB global score, 67.9. All scales of the PGWB demonstrated good internal consistency reliability (Cronbach's alpha=0.77 to 0.97). The EQ-5D and PGWB demonstrated convergent validity through significant correlations with the ADS (r=0.48 to 0.61), DSC-R scales (0.33 to 0.81 except ophthalmology subscale), and patients' Body Mass Index (0.15 to 0.38). The EQ-5D Index score and PGWB global score discriminated between groups of patients where expected (e.g., patients who wanted to lose weight vs. patients who did not; patients who had experienced hypoglycemia during the day vs. those who did not). CONCLUSIONS: The PGWB demonstrated adequate internal consistency reliability, and both generic instruments had good convergent and discriminant validity. Results support the use of the EQ-5D and PGWB among patients with type 2 diabetes, possibly in combination with condition-specific measures.

and SGRQ. And they were followed for 12 months. The frequencies of exacerbations due to COPD were assessed at clinic visit, emergency visit and hospitalization based on their medication treatment for each visit. We excluded the cases of death and loss-tofollow-up from our analysis. Poisson regression was employed for our analysis. RESULTS: During the follow-up period, 10 patients died, and 24 patients lost to follow-up. After adjusting for age, smoking history, education level, co-morbidity, stage of disease, and current smoking status, the physical domain, psychological domain, and social domain of the WHOQOL-BREF, the activity domain, impact domain and total domain of the SGRQ were significantly associated with the frequency of exacerbations. CONCLUSIONS: Quality of life is an important predictor of frequency of exacerbations. The QOL can be used to provide important information about targets for intervention.

194/Abstract 1358 An Evaluation of the Measurement Structure and Test for Gender-Based Differential Item Functioning in the Full and Mini Versions of the Asthma Quality of Life Questionnaire Carlo A. Marra, Clinical Epidemiology and Evaluation, University of British Columbia, Vancouver, British Columbia, Canada, Lynd D. Larry, Clinical Epidemiology and Evaluation, University of British Columbia, Vancouver, B.C., Canada, Chris Richardson, Nursing, University of British Columbia, Vancouver, British Columbia, Canada AIMS: To evaluate measurement models of the full and mini versions of the Asthma Quality of Life Questionnaire (AQLQ) and confirm the presence of gender-based differential item functioning (DIF) previously reported for the following two AQLQ items: _Feel bothered by or have to avoid cigarette smoke in the environment and _Feel bothered by or have to avoid going outside because of weather or air pollution . METHODS: Confirmatory factor analysis was used to evaluate the measurement models of the full and mini versions of the AQLQ using data from 202 asthma patients. Ordinal logistic regression was then used to test for the presence of gender-based differential item functioning. RESULTS: The results of the factor analyses indicated that the fit of the mini AQLQ was good and substantially better than the fit of the full AQLQ (which was found to have only moderate fit). The results did not confirm the presence of gender-based DIF for the two AQLQ items. CONCLUSIONS: Overall, the results of this study support the use of the mini AQLQ over the full length AQLQ. Further research is needed to validate the lack of gender-based DIF among different populations of asthma patients.

196/Abstract 1585 Development and Validation of a Generic QoL Questionnaire with an Integrated Physical Activity Component (QoL-Feedback) Anke Hanssen-Doose, Klaus Schüle, Institute of Rehabilitation & Sport for the Disabled, German Sport University Cologne, Cologne, Germany AIMS: Comparing the particular importance attributed to the physical, psychological and social components of Physical Activity (PA) by individuals and experts and their actual representation in QoL-instruments, a divergence becomes obvious. Current generic QoL instruments reflect PA from a risk oriented perspective, e.g. the limitation of physical functioning. The aim of this study was to develop and validate a resource oriented QoL-questionnaire with an integrated PA component. METHODS: The item pool was generated on the basis of literature analysis and qualitative interviews (Hanssen-Doose & Schüle, 2002). The pilot version is developed in German language and consists of 58 items in 23 QoL scales, which are associated with 4 main domains (physical, social, psychological and a context domain). The validation study was realised in cooperation with a statutory health insurance. The questionnaire was sent to N=1950 insured persons aged between 40 and 70. The sample was random, however, it was adjusted to two-thirds affected by type 2 diabetes and one third not affected. RESULTS: A total of n=518 individuals responded, which equals a return rate of 26%, for the retests 50%. 314 individuals were affected by type 2 diabetes (66%), 174 were not affected (34%). The sample with type 2 diabetes was characterised by 50% males, aged 62 years (SD 6.7), and BMI 30 kg/m² (SD 7.1). The sample without type 2 diabetes was characterised by 22% males, aged 55 years (SD 9.2), and BMI 27 kg/m² (SD 5.3). The average missing data rate was 1.5% per item (SD 1.8). The test of convergent validity showed high correlations with corresponding subscales of the SF-36. The internal consistency was a=0.92 for the physical domain, a=0.87 for the psychological

195/Abstract 1104 Validation of Two Generic Health-Related Quality of Life Measures in Patients with Type 2 Diabets: The PGWB and EQ5D Nicole Yurgin, Global Strategic Pricing and Health Outcomes, Eli Lilly and Company, Indianapolis, IN, Louis S. Matza, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Kristina Secnik, Global Health Outcomes, Eli Lilly and Company, Indianapolis, IN, Karen G. Malley, Malley Research Programming, Inc., Rockville, MD AIMS: Although condition-specific health-related quality of life (HRQL) measures may be more sensitive to change, generic measures are also useful because they can assess impact of disease and treatment on a broad range of health dimensions, while allowing

A-132

gathered: relevant clinical and treatment information, included tolerability and efficacy, SF-12 (HRQoL, TSQM (treatment satisfaction), and Morisky-Green (adherence) questionnaires. Feasibility, reliability, accuracy, content validity, discriminant validity, construct validity, and concurrent validity were assessed. RESULTS: Final reduced questionnaire was composed by 6 dimensions: Treatment Efficacy, Convenience of use, Impact on Daily Activities, Medical Care, General Satisfaction, and Undesirable Side-effects. Unidimensional scales (Cronbach_s alpha ranging 0.813- 0.912) correlated, and allow computation of a summary score (alpha = 0.890). SATMED-Q dimensions moderately correlated with TSQM dimensions (.577 to .680). Treatment Satisfaction interacted with HRQoL depending on pathology (max r = .485). Differences between tolerability and efficacy groups are found depending on dimension and whether the clinician or the patient informs. Adherence groups show differences in some treatment satisfaction dimensions. CONCLUSIONS: The SATMED-Q is a reliable and valid measure of treatment satisfaction, structured in 6 dimensions, and a summary score.

domain, a=0.86 for the social domain and a=0.70 for the context domain. CONCLUSIONS: The results of the QoL-Feedback validation showed very satisfying outcomes for reliability and validity. It was well accepted by the study sample and seems to be a promising alternative to common generic QoL-instruments.

197/Abstract 1136 Assessing Quality of Nursing Home Care to Enhance Quality of Life Among Chinese Nursing Home Residents: A Review of the Existing Instruments Ka Fai Wong, Linda Yin King Lee, Joseph Kok Long Lee, Connie Yuen Han Chiu, Nursing, The Open University of Hong Kong, Hong Kong, China AIMS: To accommodate the increasing elderly population who cannot be cared for at home, nursing homes are growing in size and number. The quality of care delivered by nursing homes thus becomes an important concern nowadays. As quality of nursing home care can have direct impact on residents_ quality of life, valid assessment is necessary. Yet there is not a well-established Chinese tool to address this specific issue. This study aimed to identify the crucial content and essential features of a quality of nursing home care instrument for the Chinese context. METHODS: A systematic computerized search was performed and16 instruments that had been used to assess quality of nursing home care over the last 10 years were identified. The instruments were assessed with regards to construct definition and dimension coverage. Quality indicators were compared and essential features of the instruments were identified. RESULTS: Results indicated that a desirable instrument should be theoretically based and able to assess all the structure, process and outcome perspectives. Further, it should highlight the multidimensional aspect of the nursing home care and incorporate the viewpoint from both care providers and residents. Lastly, residents_ specific needs such as language ability, medical condition and dependency level should also be addressed. CONCLUSIONS: Findings provide impetus for developing a Chinese instrument for assessing quality of nursing home care.

199/Abstract 1167 New Quality of Life (QoL) Questionnaire on Death Education for Children Nozomi Kinoshita, Michiko Kobayashi, Siebold University of Nagasaki, Nagayo-cho,, Nagasaki, Japan, Takashi Mandai, Japanese Society of Quality of Life Research, Kobe-shi,, Hyogo, Japan AIMS: The purpose of this study was to develop the desirable QOL questionnaire of death education for children. METHODS: Fortyfive children(from 7 years old to 9 years old) participated in this study. Our new original self-administered questionnaire consisted of 34 questions divided into 6 categories. RESULTS: Cronbach's alpha coefficients of our questionnaire were high enough to accept for clinical use : 0.76 in selfcare, 0.74 in friend, 0.73 in family, and 0.71 in physical problems, etc., respectively. Our questionnaire contained 9 main factors and cumulative contribution was 0.83. We examined the five-times narrative death education (Kamishibai) effects on QOL for children. Compared with the children who could not accept the death, the children who could accept the one showed the significantly excellent QOLs not only in total QOL, in the category of friend, and in the one of family before death education, but also in total QOL and in the category of friend after death education(P<0.05). CONCLUSIONS: These findings indicate that our new QOL questionnaire has high enough reliability and potency of validity to use on the evaluation of death education for children. We believe that we must pay more attention to the necessity of narrative death education for children and the new QOL questionnaire for evaluation, in order to improve the QOL of daily living for children.

198/Abstract 1634 Development and Psychometric Properties of a New Generic Questionnaire Measuring Satisfaction With Treatment With Medicines: The SATMED-Q Miguel A. Ruiz, Antonio Pardo, Methodology, School of Psychology, Universidad Autónoma de Madrid, Madrid, Spain, Javier Rejas, Javier Soto, Department of Health Outcomes Research, Medical Unit, Pfizer España, Madrid, Spain, Fernando Villasante, Centro de Salud Horcasitas, Madrid, Spain, Jose Aranguren, Clínica Madrid, Fuenlabrada, Spain

200/Abstract 1216 AIMS: To develop a multidimensional generic questionnaire measuring Satisfaction with Treatment with Medicines. The questionnaire was designed to be used in chronic patients undergoing a long term pharmacological treatment of any disease. METHODS: Six general practitioners in the Community of Madrid collected 4 samples: 1) 21 chronic patients for content focus groups; 2) 12 patients to assess feasibility; 3) 151 patients for item reductions; and 4) 455 patients for psychometric assessment. The two later stratified by gender, age and main disease (Diabetes II, Hypertension, Osteoarthritis, Prostate, EPOC/Asthma, Depression, and Migraine). Relevant issues were converted into 36 items around 7 dimensions by an expert panel. Factor Analysis item reduction resulted in a 6 dimensional, 17 item questionnaire. Additional measures were

Evaluation of WHO Quality of Life Questionnaires in the Uruguayan Elderly Population Adriana Suárez, Juan J. Dapueto, Laura Schwartzmann, Medical Psychology, Facultad de Medicina, Universidad de la República, Montevideo, Uruguay AIMS: The WHOQOL-BREF and WHOQOL-OLD questionnaires, designed to measure quality of life (QL) respectively in general population and in the elderly, and the spirituality module SRBP, have gone through many validation studies in its multilingual versions. However, little research has been conducted in South American Spanish speaking populations. This study aimed to evaluate the

A-133

performance of these instruments in the Uruguayan elderly population. METHODS: The data were collected from 425 people over 60 years of age, in four locations: a YMCA senior group, an institution for social and educational activities (UNI3), a convenience sample of elderly people not attending programmed social or educational activities, and inpatients from a public nursing home. Reliability was evaluated using Cronbach´s alpha. Spearman rho correlations were used to study associations among instruments. RESULTS: Cronbach´s alpha showed acceptable to good internal consistency, with values ranging from .67 to .73 for the BREF subscales, from .67 to .83 for the WHOQOL-OLD subscales, and from .63 to .95 for the SRPB. A correlation matrix showed high significant correlations among the BREF, OLD, and SRPB subscales (Spearman rho higher than 0.30) with the exception of the WHOQOL-OLD Death subscale, which solely correlated with the BREF Physical Well-Being subscale. Both BREF and OLD subscales showed significant mean differences among known groups (in vs. outpatients; perception of health) when tested by t-test (p<.001). While Death subscale scores were lower in inpatients (more concern about death) (p<.001), they do not differ when it comes to people who see themselves as healthy or sick. CONCLUSIONS: The results showed that the WHOQOL-BREF, WHOQOL-OLD, and SRPB have good reliability and validity to be used as research instruments in the Uruguayan elderly population. The Death subscale needs further analysis to determine its relationship with the other QL measures.

202/Abstract 1516 Development and Validation of a Questionnaire to Assess NiacinInduced Cutaneous Flushing Josephine M. Norquist, Epidemiology, Merck & Co., Inc., Upper Gwyneth, PA, Qinfen Yu, John F. Paolini, Clinical and Quantitative Sciences, Merck & Co., Inc., Rayway, NJ, Kelly McQuarrie, Douglas J. Watson, Epidemiology, Merck & Co., Inc., Upper Gwyneth, PA AIMS: HDL level is an independent risk factor for coronary heart disease (CHD). Reducing LDL and raising HDL are methods for reducing risk of CHD. Niacin is the most effective agent approved for raising HDL. However, niacin-induced flushing (NIF) significantly limits patient acceptance. To further characterize NIF, a patientreported Flushing Symptom Questionnaire (FSQ) which assesses the frequency, severity, duration and bother of NIF was developed and validated. METHODS: A draft FSQ was pilot tested and refined through subject interviews and health-care providers who prescribed niacin. The final FSQ was validated in an 8-week, randomized, double-blind, placebo-controlled study of niacin and placebo. Patients were randomly allocated to 1 of 4 treatment groups (combinations of placebo, niacin 1g and 2g). Patients completed the FSQ daily using an e-diary. The primary endpoint of the study was based on Global Flushing Severity Score (GFSS), an item of the FSQ that assesses overall flushing severity on a 0-10 scale with labels of None(0), Mild(1-3), Moderate(4-6), Severe(7-9) and Extreme(10). RESULTS: 175 patients were enrolled (placebo n=33, niacin n=142, at week one). GFSS test-retest reliability and reproducibility coefficients were all above 0.75. Construct validity of the GFSS was supported by moderate to strong correlations (r > 0.5) with other FSQ scores. The GFSS and other FSQ scores discriminated between treatment groups and demonstrated expected relationships with known groups. The GFSS demonstrated a high responsiveness in patients who started on niacin and switched to placebo during the study. Similar results were seen for FSQ items related to flushing bother, and to individual symptom items of redness, warmth, itching and tingling; however, the GFSS performed as well or better than any of the individual symptom items. CONCLUSIONS: Results support the measurement properties and validity of the FSQ, specifically the GFSS, the individual flushing symptom scores, and the flushing bother score as endpoints in clinical trials to assess niacin-induced flushing.

201/Abstract 1218 The Translation and Linguistic Validation of the Impact of Weight on Quality of Life (IWQoL-LITE) Questionnare Caroline Houchin, Diane Wild, Oxford Outcomes Ltd, Oxford, United Kingdom AIMS: The IWQOL-Lite was developed in 2001, in the United States and has been translated into 41 different languages. The IWQOL-Lite is a 31-item short form of the original IWQOL. It is designed to assess the effect of obesity on quality of life. The objective of this study was produce translations that are conceptually equivalent to the original and to other language versions ensuring the relevance of the translation within the target cultures. METHODS: A standard methodology was employed: 2 forward translations, a reconciliation of the forward translations or an in-country review, 2 back translations, back translation review, linguistic validation interviews with 5 obese patients in each country and/or a review by clinicians in the country and 2 proofreadings. RESULTS: Numerous cultural and linguistic issues became apparent throughout the translation process. In Czech, Hungarian and Spanish, it was necessary to clarify that because of the patient's weight, they had problems picking up objects _'from the floor ' - otherwise this could be misunderstood as lifting any object. The term 'mobility' was considered to be too high a register and was modified to _'getting around ' in UK English and _'moving around ' in Danish. There were few direct equivalents for the term _self-conscious and it had to be explained in everyday language as _'embarrassed / awkward / overly aware of myself' . _Sexual performance could not be translated literally as it had a negative connotation in most languages. In UK English, it was modified to _'I experience physical difficulties during sexual activity '. The wording _'turnstiles' was found to be culturally inappropriate for Norway and Spain. The wording was retained in Norway, as it was understood; in Spain, it was reworded as '_revolving doors '. CONCLUSIONS: The IWQOL-Lite has been translated and linguistically validated in 26 languages using a rigorous translation process. A number of cultural and linguistic issues became apparent and were resolved. The measure is now appropriate for use in multinational trials.

203/Abstract 1727 Testing a Conceptual Model of Health-Related Quality of Life in Overactive Bladder Karin S. Coyne, Chris Thompson, Donald Stull, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Zhanna Jumadilova, Outcomes Research, Pfizer, Inc, New York, NY AIMS: Overactive bladder (OAB) is a symptomatic condition; therefore, treatment goals are to reduce symptom bother and improve health-related quality of life (HRQL). Conceptual models, or frameworks of expected findings, are useful to hypothesize treatment outcomes. We evaluated the validity of a conceptual model that OAB treatment would improve urinary symptoms which would then lead to improvements in HRQL and symptom bother. METHODS: Secondary analyses of data from a 12-week, single-arm trial of OAB patients treated with tolterodine were performed. Patients completed the Overactive Bladder Questionnaire (OAB-q) and bladder diaries assessing urinary frequency, urgency, and incontinence at baseline, 4, and 12 weeks. The OAB-q includes a Symptom Bother scale and 4 HRQL subscales (Coping, Concern, Sleep, and Social Interaction). Latent growth modeling (LGM) was used to evaluate the

A-134

simultaneous effects of changes in bladder diary variables on Symptom Bother and HRQL. LGM estimates intercepts and slopes of change as latent variables in a structural equation modeling context to assess the relationship of one growth curve with another while modeling measurement error. Slopes of change for all variables were calculated using baseline, 4-, and 12-week data. RESULTS: 583 patients had complete data for all time points. Mean age was 61 years; 70% were women and 90% were white. Overall model fit was quite good (CFI=0.97, SRMR=0.03). Improving urinary symptoms improved HRQL and reduced Symptom Bother, with some domains improving more than others. Reductions in urinary frequency appear to have the greatest effect on the Symptom Bother (Beta= 0.6), Sleep (Beta= -0.65), and Coping (Beta= -0.42) domains, whereas reductions in urgency appear to have the greatest effect on Concern (Beta= -0.42) and Social Interaction (Beta= -0.27). CONCLUSIONS: A conceptual model of treatment-related improvements in HRQL and symptom bother was validated using LGM. This demonstrates the usefulness and applicability of LGM methods for evaluating longitudinal conceptual models of patientreported outcomes.

205/Abstract 1171 Quality of Life is not Differnet in Kidney Transplanted Patients Treated with Cyclosporine Versus Tacrolimus Bases Immunosuppression Miklos Z. Molnar, Agnes Z. Kovacs, Andras Szentkiralyi, Institute of Behavioural Sciences, Csaba Ambrus, Fresenius MC, Eszter Vamos, Agnes Koczy, Lilla Szeifert, Anett Lindner, Maria Czira, Marta Novak, Istvan Mucsi, Institute of Behavioural Sciences, Semmelweis University, Budapest, Hungary AIMS: Earlier reports suggested that tacrolimus base immunosuppression may be associated with better quality of life in kidney transplanted (Tx) patients. In this analysis we assessed if health related quality of life (HRQOL) was associated with immunosuppressive medications in Tx patients. METHODS: All patients 18 years or older (n=1067) who were followed at our university outpatient transplant clinic were approached to participate in the study. Baseline demographic characteristics and laboratory results and a list of immunosuppressive medication were recorded at enrollment. The patients completed a battery of validated, selfadministered questionnaires including the Kidney Disease Quality of Life Questionnaire (KDQOL). Statistical analysis was carried out using the SPSS software. RESULTS: Mean age of patients was 49±12 years, 60% were male, the prevalence of diabetes was 17%. 67% of patients were on cyclosporine based; 17% of patients were on tacrolimus based immunosuppressive therapy. 13% of patients have been talking azathioprine and 63% took mycophenolate mofetil (MMF). Patients who were on cyclosporine had similar QoL scores for most of the HRQOL domains to patients on tacrolimus (median (interquartiles range, IQR) 75 (38) vs 75 (38) for "burden of kidney disease"; 80 (33) vs 80 (32) for "emotional wellbeing"; 75 (35) vs 80 (38) for "physical functioning"; p=NS, for all). We did not find any significant difference in the HRQOL of patients on azathioprine versus MMF. CONCLUSIONS: None of the most freq uently used immunosuppressive medications is associated with an advantage in terms of HRQOL in kidney transplanted patients.

204/Abstract 1733 Validation of the Overactive Bladder Family Impact Measure Louis Matza, Karin S. Coyne, Jessica Brewster-Jordan, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Evan Goldfischer, Research, Hudson Valley Urology, P.C., Poughkeepsie, NY AIMS: Overactive bladder (OAB) impairs patients' health-related quality of life; however, there are no validated instruments to assess the impact of OAB on patients' family members. We assessed the psychometric properties of a new instrument, the Overactive Bladder Family Impact Measure (OAB-FIM). METHODS: Family members of patients with or without OAB (spouses/significant others and daughters) completed the 32-item draft OAB-FIM. OAB and control patients completed the Overactive Bladder Questionnaire (OAB-q) and the Patient Perception of Bladder Condition (PPBC). Validated relationship quality measures were also administered. Subscale structure was determined with item and exploratory factor analyses; reliability and validity were assessed. RESULTS: 193 patient-family member dyads (163 OAB, 30 control) were included. Because of poor item performance (eg, floor effects, inadequate factor loading), 13 items were deleted from the OAB-FIM, leaving 19 items in the final OAB-FIM, with 6 subscales. Only 4 subscales (Irritation, Activities, Travel, Concern) were relevant to all family members; 2 additional subscales (Sleep, Sex) were relevant only to spouses and significant others. The OAB-FIM was highly discriminating between OAB and control family members, with OAB family members indicating greater impact (all P<0.0001). Internal consistency reliability (Cronbach's alpha >0.71) and test-retest reliability (intraclass correlation coefficients >0.73) were high for all subscales. Concurrent validity of the OAB-FIM was supported by significant (P<0.001) Spearman's correlations with the OAB-q (0.35-0.58) and PPBC (0.31-0.56). OAB-FIM scores discriminated by family member perceptions of patient OAB severity. Correlational analyses of the OAB-FIM and relationship quality measures suggest that greater symptom impact on family members is associated with poorer relationships. CONCLUSIONS: This study supports the reliability and validity of a new measure to assess the impact of OAB on family members and demonstrates that family members are affected by the disease whether or not they live with the patient.

206/Abstract 1172 Health-Related Quality of Life of Chronic Dialysis Patients Treated with Different Dialysis Modalities Eszter Vamos, Miklos Z. Molnar, Institute of Behavioural Sciences, Csaba Ambrus, Fresenius MC, Andras Szentkiralyi, Agnes Z. Kovacs, Agnes Koczy, Lilla Szeifert, Maria Czira, Anett Lindner, Marta Novak, Istvan Mucsi, Institute of Behavioural Sciences, Semmelweis University, Budapest, Hungary AIMS: Health related quality of life (HRQoL) is an important measure of treatment effectiveness in chronic diseases. Furthermore, HRQoL scores are reportedly associated with hospitalization and mortality in chronic dialysis patients. It is not known, however, if different dialysis modalities (hemodialysis (HD) vs continuous ambulatory peritoneal dialysis (CAPD)) are associated with different level of quality of life. In this cross-sectional study we assess the association of HRQOL with the type of dialysis in chronically dialyzed patients on the transplantation waiting list. METHODS: Patients (n=214) from nine dialysis center in Budapest were approached to participate in the study. Thirty nine patients refused participation therefore our population consists of 175 patients. Baseline demographic characteristics and laboratory results were collected. The patients completed a battery of validated, selfadministered, questionnaires including the Kidney Disease Quality of Life Questionnaire (KDQOL). Statistical analysis was carried out using the SPSS software. RESULTS: Mean age of patients was

A-135

AIMS: To measure quality of life in a large sample of patients with rheumatoid arthritis (RA) treated by a representative number of wellknown Argentine rheumatologists. METHODS: Subjects with an adequate literacy level and who signed consent were included. The questionnaire gained information regarding demographics, diagnosis, disease duration, co-morbidity, treatment, pain, disability, and validated versions for Argentina of SF-36, HAQ, RADAI, EuroQoL and BASFI. RESULTS: 31 sites and 64 investigators and coinvestigators took part in this trial, with 1059 evaluable patients (pts): 1024 (96.7%) subjects with RA, and 33 (3.1%) pts with SA;83% of the total sample were female; the mean age was 54 (17-88) years; time with symptoms was 11,8 years, and time with diagnosis was 10.5 years; 186 pts (17.6%) had symptoms during £2 years and the rest (82.4%) during >2 years. The population with SA was not statistically significant and was not taken into ac-count. Results for the RA pts were: SF-36:Total values per sub-group/ SD: PF:52/ 28; PR:36/ 41; BP:51/ 25; GH.:48/ 21; V:53/ 22; SF:65/ 27; RE:53/ 42; MH:61/ 22. Euro-QoL:Respondents with the option of "No problems" were 34% in Mobility; 62.3% in Personal care; 33.8% in Daily activities; 56.7% referred anxiety and/or depression and 74% moderate pain (only 7.7% chose _no pain ). RADAI: Median (interquartile range) was 3.8 (Q1=2.1 and Q3=5.2); 25% of the patient showed a score greater than 5.2, which expressed a great activity of the disease. HAQ: Median score was 0.87 (mild incapacity) and 11.6% of the pts had severe incapacity. There was a significant correla-tion (P<0.001) between doctors and pts at categorizing health status and pain, both evaluated dur-ing the survey day. CONCLUSIONS: This was the first national survey on quality of life of rheumatic diseases. This can be used to make national and international comparisons about quality of life in generic and disease specific instruments, and to develop additional surveys focused in definite groups of pts.

48±13 years, 62% were male, the prevalence of diabetes was 16%. 25 patients were on CAPD and 150 were on HD. The demographic characteristics were similar in the two patient populations, except the total end stage renal disease time (median (interquartiles range (IQR)): HD: 40 (43) months vs CAPD: 12 (22) months; p<0.001) and serum hemoglobin (HD: 111±15 g/l vs CAPD: 119±17 g/l; p=0.018). Quality of life scores were similar for the two dialysis modalities in all HRQoL domains. QoL scores for the "physical functioning" domain seemed to be somewhat lower for patients on CAPD (median (IQR); CAPD vs HD: 55 (40) vs 75 (35); p=0.065) but the difference did not reach statistical significance. CONCLUSIONS: Quality of life seems to be overall similar for comparable patient groups on different dialysis modalities.

207/Abstract 1173 Obstructive Sleep APNEA Syndrome is Independently Associated ith Worse Quality of Life in Transplated Patients Agnes Koczy, Miklos Z. Molnar, Eszter Vamos, Agnes Z. Kovacs, Andras Szentkiralyi, Lilla Szeifert, Rezso Zoller, Andrea Dunai, Istvan Mucsi, Marta Novak, Institute of Behavioural Sciences, Semmelweis University, Budapest, Hungary AIMS: A high prevalence of Sleep Apnea Syndrome (SAS) have been reported in patients with chronic renal disease. Previous studies suggested that the presence of SAS had a negative impact on healthrelated quality of life (QoL) in dialysed patients. It is not known whether the presence of SAS has a negative impact on QoL in transplanted patients. The aim of our study was to analyze the association of SAS and QoL in renal transplant recipients. METHODS: We report data obtained in 841 patients followed in a single transplant center in Budapest. The patients completed a battery of questionnaires that included the Berlin Questionnaire to estimate the risk for SAS. Also included in the battery was the Kidney Disease Quality of Life Questionnaire (KDQoL). We also recorded laboratory data, socio-demographic and clinical parameters. RESULTS: Mean age of Tx patients was 49±13 years (18-75 years), 59% male, 17% diabetics and the calculated creatinine clearance (abbreviated MDRD formula) was 49±19 ml/min. The prevalence of patients with high risk for SAS was 27.5%. Quality of life of patients with high risk for SAS was significantly worse along all of the QoL domains assessed by the questionnaire than QoL of patients with low risk for SAS. For example patients with high risk had significantly worse self perceived physical functionality (median (interquartile range (IQR)): 60 (45) vs 84 (35); p<0.001), vitality (median (IQR): 58 (40) vs 70 (30); p<0.001), emotional well being (median (IQR): 68 (36) vs 80 (28); p<0.001), burden of kidney disease (median (IQR): 69 (50) vs 81 (31); p<0.001) and sleep (median (IQR): 70 (33) vs 78 (23); p<0.001). Finally, in multivariate linear regression analysis, high risk for SAS remained a significant, independent predictor of worse QoL along almost all of the generic and disease targeted domains after adjusting for socio-demographic and clinical co-variables. CONCLUSIONS: The presence of obstructive sleep apnea syndrome is independently associated with poorer QoL in transplanted patients.

209/Abstract 1052 Quality of Life (QoL) of Chemical Devotees Involved in Pulmonary Complications of Sulfur Mustard Sedigheh Khanjari, H,A Jamali,., nursing, Iran University of Medical Sciences & Health Service, Tehran, Iran AIMS: Sulfur mustard has a great negative impact on physical, psychological, social and spiritual aspects of the QoL.Now, people of Sardasht infected with sulfur mustard still suffering from various damages of organs particularly pulmonary system after 18 years (2004) .The Specific objects were to (1) describe the multidimensional QoL of people;(2) identify the influence of demographic, medical, and other psychological factors on people QoL. METHODS: To determine the QoL of chemical devotees with pulmonary complications of sulfur mustard .This was a descriptive study in which data were obtained through questionnaires 260 chemical devotees, all live in sardasht at time of chemical bombardment of sardasht, suffering from pulmonary complications of sulfur mustard contributed in this study. We chose Sardasht because UN confirmed the use of mustard gas as well as nerve agent against Iranian troops. Also, samples were selected by random sampling. RESULTS: Findings showed that QoL in 71.5% of the sample group was moderate , only in 5.8% it can be considered good , and regarding physical , psychological , social well being belonged respectively to 11.9% , 8.1% , 1.9% whereas 22.8% of the participants reckoned that their spiritual aspect is good . CONCLUSIONS: People infected with sulfur mustard are in need of programs to assist them . Programs need to include responsible organizations to help counteract negative effect of the damage on their mental health, and to enable chemical devotees to continue with a better quality of life. Based on the findings, the QoL in the social ,

208/Abstract 1088 Cooperative Quality of Life Study of RA in (COP QoL Argentina) Maria C. Lunic, Rheumatology, Hospital De Clinicas Jose De San Martin, Buenos Aires, Buenos Aires, Argentina, Gustavo Citera, Rheumatology Section, Instituto De Rehabilitacion Psicofisica, Buenos Aires, Buenos Aires, Argentina

A-136

physical and psychological domains was not good, so responsible organizations by providing drugs with low price , general car transportation particularly for chemical devotees , improving insurance services and developing city services of sardasht , should payback just a little bit of their sacrificing act . The researchers suggest a qualitative study on the each domains of quality of life in chemical devotees.

range of 16 and 28 years from the southern part of the country, who had worn spectacle lenses and contact lenses for the past six months were used for the study. They comprised of 66 myopes and 70 hyperopes who had no ocular disease. NEI-RQL-42 instrument was administered to these subjects who indicated responses for the use of both spectacle and contact lenses. The principal outcome examined was difference in quality of life scores in the different scales. Results were presented in percentile values,unpaired student t-test was used to determine statistical significance of differences between the two forms of correction. RESULTS: For myopes and hyperopes combined, the differences in scores between spectacle and contact lens corrections ranging from 0.5 to 38.8 on the 0 to 100 possible score range were seen in all 13 scales. Appearance showed the highest difference of 38.8 with 59.4 for spectacle and 98.2 for contact lenses,this difference was statistically significant. (P < 0.05). Worry showed the least difference of 0.5.Contact lenses showed a higher quality of life score in 11 of the 13 scales, while correction with spectacle lenses showed higher scores in 2 scales namely, symptoms and suboptimal correction.Analysis of myopes showed that 7 scales had a statistically significant difference between the two forms of correction. (P < 0.05). While the analysis of hyperopes showed that 5 scales had a statistically significant differences.(P < 0.05). CONCLUSIONS: It is therefore concluded that patients who have their refractive errors corrected with contact lenses will experience better quality of life than those corrected with spectacle lenses.

210/Abstract 1787 Literature Review of Electronic Patient Reported Outcome (ePRO) Measures Including Feedback Reports for Clinical Practice Janine Becker, Matthias Rose, Science Team, QualityMetric Incorporated and Health Assessment Lab, Waltham, Massachusetts, Barbara Gandek, Science Team, Health Assessment Lab, Waltham, Massachusetts, Jakob B. Bjorner, Science Team, QualityMetric Incorporated and Health Assessment Lab, Waltham, Massachusetts AIMS: Patient reported outcome (PRO) measures are important for assessing and monitoring disease impact and treatment outcomes. Electronic PRO (ePRO) measures that provide feedback reports may improve the efficiency of PRO measurement and facilitate its integration into clinical practice. However, the acceptance of ePRO measures in clinical practice is still very low. A literature review was carried out to summarize the current use of ePRO tools that provide feedback reports, investigate possible obstacles to the widespread use of ePRO, and describe successful ePRO implementations. METHODS: We searched the literature using PubMed and PsychInfo for articles published from 1996 to 2005 in English and German. Articles describing the use of ePRO measures in real world clinical applications were retrieved, and reviewed by three independent researchers for relevance. RESULTS: 21 articles were found describing the use of ePRO tools that included feedback reports. In 10 articles, PRO measures were administered by computer devices, in 5 articles using scanning technology of paper surveys. Most articles indicated that an instant PRO report was useful for clinical practice, but reports of permanent integration of ePRO into daily office routines were found only episodically. Several key obstacles to widespread use were found. Examples of successful ePRO integrations in clinical practice with at least 90% of patients routinely administered were found for oncology, nephrology, endocrinology, and psychosomatic settings. CONCLUSIONS: Implementing ePRO measures seems to be a feasible approach to assess PRO in clinical practice, but it requires well-designed software solutions and a sustainable support infrastructure to overcome obstacles in daily routines. Computerized Adaptive Tests (CATs) could be one method to achieve the high measurement precision needed for individual diagnostics with acceptable respondent burden.

212/Abstract 1430 Self-Esteem and Quality of Life in American and Brazilian Cardiac Patients Rosana Aparecida S. Dantas, School of Nursing of Ribeirão Preto, University of São Paulo, Ribeirão Preto, São Paulo, Brazil, Márcia A. Ciol, Rehabilitation Medicine, School of Medicine, University of Washington, Seattle, Washington AIMS: Self-esteem (SE) has been shown to be an important factor in the rehabilitation process and quality of life (QoL) after cardiac events in American subjects. We compared SE and QoL between American and Brazilian patients after Coronary Artery Bypass Graft (CABG). The initial hypothesis was that SE and QoL instruments would yield higher values for the American subjects. METHODS: Participants were 97 Brazilian and 84 American patients who underwent CABG surgery in a university hospital in their respective country. American subjects responded to a mailed questionnaire and Brazilians were interviewed in person. We applied the modified Flanagan Quality of Life Scale and the Rosenberg Self-Esteem instruments. Validated translations of both instruments were used in Brazil. Descriptive, t-tests, and linear regressions analysis were used. RESULTS: American patients tended to be older, had less time since surgery, lower proportion of females, higher number of bypasses, and fewer comorbidities than Brazilians (all p<0.02). Married status was similar in both groups. The modified Flanagan's QoL differed between American and Brazilian patients [mean 73.5 (sd=13.4) vs. 85.8 (13.0), respectively, p<0.001], while Rosenberg Self-Esteem Scale did not [33(4.9) vs. 32.2 (4.4), p = .15]. In a linear regression analysis, country of origin was statistically significant for QoL (p<.001) but not for self-esteem (p=.95), after adjusting for age, sex, marital status, time since surgery, presence of comorbidities, and number of bypasses. CONCLUSIONS: Brazilians reported higher QoL and similar SE to their American counterparts. Conclusions were unchanged after adjusting for other demographic and clinical variables. We expected that the poorer economic situation and somewhat lesser access to care in Brazil would translate into lower QoL and SE values. We speculate that the Brazilian versions of the QoL and SE instruments may not capture the same constructs as the

211/Abstract 1767 Quality of Life with Different Forms of Vision Correstion: A Comparative Study of Spectacle Lenses and Contact Lenses Faustina K. Idu, Eghosasere F. Iyamu, Optometry, University Of Benin, Benin City, Edo State, Nigeria, Uduak C. Udom, Optometry, Ceejay Vision Centre, Uyo, Akwa Ibom State, Nigeria AIMS: Spectacle lenses is the most accepted form of correction for refractive errors in Nigeria. However in recent years, the use of contact lenses is gaining more patronage especially amongst the younger generation. The aim of this study was to compare the quality of life reported with the use of spectacle lenses and contact lenses. METHODS: One hundred and thirty six subjects within the age

A-137

NYHA class I, II, III/IV analysis of variance (ANOVA) was applied. The convergent and divergent validity was evaluated through Spearman correlation coefficient (r) between the dimensions of the LHFQ and the generic instrument The Medical Outcomes Study (MOS) 36-item Short-Form Health Survey (SF-36) . RESULTS: The LHFQ total score and the physical dimension sub-scale discriminated the three sub-groups of NYHA classes I, II and III/IV (p< 0.0001). The emotional dimension discriminated the NYHA classes I and III/IV (p< 0.0034). The convergent validity of the LHFQ was supported, as the physical and emotional dimensions of the LHFQ and the SF-36 were significantly correlated (coefficients from 0.46 to 0.70 for physical dimension and from 0.31 to 0.65 for emotional dimension). Nevertheless the divergent validity between the LHFQ and the SF-36 was not well supported. CONCLUSIONS: The LHFQ seems to be a useful and valid tool to measure HRQL in the elderly with HF but new studies are still necessary to better understand its divergent validity, as well as its capacity to discriminate efficiently the patients more severely sick.

original instruments, possibly because of the cultural differences influencing the perception of SE and QoL in each population.

213/Abstract 1715 Depression is a Major Determinant of Quality of Life in Patients with Ischemic Heart Disease and End-Stage Renal Disease Luciane N. Cruz, Epidemiology, Marcelo P. Fleck, Psychiatry, Universidade Federal do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, Brazil, Carisi A. Polanczyk, Cardiology, Hospital de Clínicas de Porto Alegre, Porto Alegre, Rio Grande do Sul, Brazil AIMS: To evaluate the impact of depression on several domains of QOL in ischemic heart disease and end-stage renal disease and to compare the prevalence of depression among these patients. METHODS: Cross-sectional survey conducted in 176 patients, 105 with ischemic heart disease and 71 in hemodialysis assisted in an university hospital in Porto Alegre, Brazil. Depression was diagnosed according to the Mini International Neuropsychiatric Interview 5.0 and depressive symptoms were measured by the Beck Depression Inventory. QOL was assessed through the translated and validated versions of the Short Form-36 (SF-36) and World Health Organization Quality of Life Instrument-bref (WHOQOL-bref). Linear and logistic multivariate analysis were performed to estimate the independent effect of depression on QOL indices, after adjusting for socio-economic and clinical variables. RESULTS: Depression prevalence was 14.3% in the ischemic heart disease cohort and 9.9% in the hemodialysis group; depressive symptoms were present in 39% and 36% of each cohort, respectively. There was no significant difference in the QOL scores between the two groups. Patients with depression reported scores significantly lower than those nondepressed in all domains. The most affected were emotional role (24±34 vs. 66±42; p <0.01) and mental health (44±19 vs. 80±17; p <0.01)of the SF-36 and psychological domain (46±17 vs. 63±11; p <0.01) of the WHOQOL-bref. Multivariate analysis showed that depressive symptoms were predictors of lower QOL in all domains of the SF-36, with the highest standardized beta coefficients (ranging from -0.26 to -0.64). The R² coefficient was lower than 0.3 in most domains of SF-36, and between 0.3 and 0.42 of the WHOQOL-bref. CONCLUSIONS: Depression is an independent factor associated to worse QOL in patients with ischemic heart disease, and of similar magnitude to those with end-stage renal disease. Strategies aiming at improving QOL of these patients must consider the evaluation and management of depressive symptoms.

215/Abstract 1441 Predictors of Health Related Quality of Life in Elderly With Heart Failure Maria José D. Diogo, Fátima A. A. Scattolin, Roberta C. R. Colombo, Nursing, State University of Campinas, Campinas, São Paulo, Brazil AIMS: The purpose of this study was to identify predictors of healthrelated quality of life (HRQOL) in elderly with heart failure (HF). METHODS: An exploratory correlational design was adopted for the study. One hundred forty-six elderly outpatients with HF completed the Socio-demographic Characteristics Questionnaire and the Minnesota Living with Heart Failure Questionnaire (MLHFQ). RESULTS: Mean scores of MLHFQ were 41,2 (±17,8). Comparisons between the MLHFQ scores and the New York Heart Association (NYHA) classification for HF showed that patients in NYHA class I and II had significantly lower scores than those in NYHA class III and IV. Patients who reported dyspnea, fatigue and edema, as well as those who had not the habit of regular physical exercise, showed a greater compromise of HRQOL. Multiple regression analysis showed that NYHA classification and dyspnea were predictors of HRQOL in the elderly, explaining 42% of the HRQOL variance. CONCLUSIONS: Our findings point to the importance of interventions that enable the early detection and/or control of HF development, thus avoiding further progression of symptoms and impairment of functional status. Prospective studies should be conducted in order to assess whether control of dyspnea improves HRQOL in elderly people with HF.

214/Abstract 1449 Psychometric Properties of the Minnesota Living With Heart Failure - Brazilian Version in the Elderly Fernanda A. Cintra, Izabel C. R. S. Saccomann, Maria Cecília B. J. Gallani, Nursing, State University of Campinas, Campinas, São Paulo, Brazil

216/Abstract 1131 Quality of Life and Study Habits in Children of Alcoholics and Non Alcoholics Selwyn Stanley, Aswathy P. Rajan, Social Work, Bishop Heber College, Tiruchirappalli, Tamilnadu, India

AIMS: Among the specific instruments for heart failure (HF), Minnesota Living with Heart Failure (LHFQ) stands out but its psychometric properties have not been extensively studied in the elderly. The objective of this study is to evaluate the discriminative, convergent and divergent validity of the LHFQ, Brazilian version, applied to the elderly with HF. METHODS: Convenience sample of 170 elderly people, age > or = 60 years, with medical diagnosis of HF and in outpatient clinic was gathered from two university hospitals. The sample size was determined upon the correlation values considering as minimum significant correlation, r > or = 0.30. To evaluate the validity discriminative the LHFQ among patients with

AIMS: Children of alcoholics are often viewed as having major psychological problems resulting from their interactions within disturbed family structures. Studies indicate that they are at increased risk for a variety of negative outcomes such as anxiety disorders, low self-esteem, difficulties in family relationships, and generalized distress and maladjustment. Besides, they constitute a population atrisk for poor academic performance, school absenteeism and drop out. This study was done to assess and compare the nature of study

A-138

despite adjustment for a comprehensive list of clinical and health status measures. Such findings suggest that psychosocial characteristics may rival more common clinical status measures in predicting overall health-related quality of life among older adults receiving services in the community.

habits and QOL of children from families having an alcoholic parent with children of non-alcoholic parents and to determine the extent of correlation between their QOL and study habit dimensions. METHODS: Fifty children of alcoholics were contacted at a deaddiction center in Tiruchy, India, where their father was registered for in-patient treatment. A group of children of non-alcoholic parents,identified from the community and matched on age, family income and ordinal position served as the reference group. The Study Habit Inventory which assesses seven dimensions (home environment and planning, reading and note taking, subject planning, concentration, exam preparation, school environment and general habits) was administered along with the Perceived QOL scale by Patrick & Davis which assesses QOL on social, physical and cognitive dimensions. RESULTS: The results indicate that children of alcoholics showed a statistically significant difference from the control group on the overall QOL score and all its components. Further, their study habits were poorer than the control group subjects particularly with regard to their home and school environment, concentration, preparation for examination and general habits and attitudes. CONCLUSIONS: It is concluded that parental alcoholism results in a poor QOL in children and adversely impacts their academic habits. The finding is in consonance with the stress perspective of alcoholism which contends that the stress of living with an alcoholic can have adverse consequences for other family members.

218/Abstract 1761 Quality of Life and Physical Functioning After Short-Term Inpatient Rehabilitation for Older Adults with Severe COPD Joan J. Branin, Health Services Management and Gerontology, University of La Verne, La Verne, CA AIMS: Chronic obstructive pulmonary diseases (COPD) are a major cause of death and chronic disability. COPD has a slow onset during which lowered vital capacity and dyspnea with exertion lead to restrictions of daily activities. Quality of life deteriorates with increasing disability. This research evaluated pulmonary function and endurance, physical functioning, quality of life, and hospital and healthcare resource utilization before and after a three-week inpatient pulmonary rehabilitation program for older patients with severe COPD. Few studies have examined these measures at admission, discharge, and over multiple time periods post-hospital discharge for patients with severe COPD. METHODS: Ninety-six older patients with severe COPD (mean percent predicted FEV1 = 30%) were interviewed at admission and discharge and at two, four, and six months after discharge using the SF-36. Pulmonary function and walk scores were obtained from patient records. Hospital and healthcare resource utilization was based on self-reported follow-up data. RESULTS: Physical functioning significantly improved from admission to discharge and was maintained at six-month follow-up. Despite severity, these patients experienced significant improvement in psychological well-being from admission to discharge. All three aspects of quality of life (physical, emotional, and social role functioning) improved significantly and were maintained at sixmonth follow-up. Six months after discharge, these patients were performing their physical, emotional, and social roles at the same levels as the normal population. These patients were hospitalized significantly less frequently and for fewer days. Significant reductions in emergency visits were reported. CONCLUSIONS: These findings support the use and benefits of short-term inpatient pulmonary rehabilitation for older adults with severe COPD in improving physical and psychological outcomes while reducing healthcare resource utilization particularly relevant in today's managed care environment.

217/Abstract 1786 Social and Clinical Correlates of Health-Related Quality of Life in Frail Community-Based Seniors Colleen J. Maxwell, Jian Kang, Jennifer D. Walker, Community Health Sciences, Jenny Zhang, Institute of Health Economics, David B. Hogan, Medicine, University of Calgary, Calgary, Alberta, Canada, David Feeny, Public Health Sciences & Economics, University of Alberta, Edmonton, Alberta, Canada, Walter Wodchis, Health Policy Management & Evaluation, University of Toronto, Toronto, Ontario, Canada AIMS: To examine the relative contributions of social and clinical factors to health-related quality of life (HRQL), as measured by the Health Utilities Index Mark 2 (HUI2) among older home care clients living in the community. METHODS: Clients from two older home care populations in Calgary Alberta (n=264) and Wayne County Michigan (n=250) responded to interviewer administered HUI2 questionnaires and were then assessed using the Minimum Data Set Home Care (MDS-HC) assessment tool. HUI2 scores were compared across clients_ demographic, social and health characteristics, including prevalent chronic conditions. The relative importance of the selected social and clinical/health variables to HUI2 scores were examined using multivariable linear regression models. RESULTS: The overall mean HUI2 was 0.49 (sd=0.18). Mean HUI2 scores were significantly lower among clients with caregivers expressing high levels of stress/burden (0.41, sd=0.16), clients reporting poor selfrated health (0.43, sd=0.18), and those with arthritis (0.46, sd=0.18), congestive heart failure (0.46, sd=0.17), and urinary tract infections (0.40, sd=0.14). At the multivariate level, the strongest associations with HUI2 scores were observed for caregiver stress and self-rated health. Significant associations were also observed for the presence of arthritis, congestive heart failure and urinary tract infections. Controlling for several determinants of health, factors not associated with HUI2 scores included the presence of diabetes, coronary artery disease, cerebrovascular disease, chronic obstructive pulmonary disease, marital status, living arrangements, age, sex and age/sex interactions. CONCLUSIONS: Caregiver stress remained a particularly strong and independent correlate of lower HRQL scores,

219/Abstract 1565 Aging With Chronic Diseases Josiane P. Alberte, Psychology Medicates and Psychiatry, State university of Campinas - UNICAMP, Campinas, São Paulo, Brazil, Maria Elena Guariento, Medical clinic, State university of Campinas - UNICAMP, Campinas - Distrito de Barão Geraldo, São Paulo, Brazil AIMS: This study intended to evaluate the effect of common chronic diseases (CD) in elder_s quality of life (QOL). METHODS: Two groups of elders (age between 60 _ 80 years old) were evaluated: 46 old people from the Geriatric Ambulatory (GA) of the State University Hospital in Campinas (SP); 38 old members from the Third Age Group (TAG) of SESC in Campinas (SP). The evaluation of the eighty-four elders was based in an interview about the known co-morbidities and the used medications, besides the WHOQOL _ bref questionnaire. RESULTS: In matter of CD, the study showed

A-139

that 18.4% from TAG and 28.2% from GA were diabetic; 56.2% from TAG and 63.0% from GA presented a cardiac disease; 18.4% from TAG and 43.4% from GA presented arterial hypertension (p=0.014); 21.5% from TAG and 34.7% from GA complained of generalized pain; 34.2% from TAG and 69.5% from GA were depressive (p<0.001); and 23.6% from TAG and 34.7% GA presented any other kind of CD. In terms of medicine, 34.2% from TAG and 8.7% from GA took one medicine per day (p=0,004); 28.9% from TAG and 30.4% from GA took three medicines per day; and 21.0% from TAG and 58.7% from GA took six or more medicines per day (p<0.001). The WHOQOL-bref showed in the physical domain: 69.64 points to TAG elders and 48.21 points to GA elders (p<0.001); in the psychological domain: 70.83 points to TAG elders and 58.33 points to GA elders (p<0.001); in the social domain: 75.00 points to TAG elders and 66.67 points to GA elders; in the environmental domain: 68.75% points to TAG elders and 53.13 points to GA elders. CONCLUSIONS: Although TAG elders and GA elders differed significantly only in the diagnosis of arterial hypertension and depression, the second group took much more medicines and demonstrated an evident worse QOL in all the domains except the social one. The present evaluation is a confirmation of the importance that Public Politics, points to TAG elders and health services and professionals of health area must develop strategies for the prevention of CD and their complications among old people, trying to change the unhealthy life style what can bring the improvement of their QOL.

all SF-36 subscales, except social functioning. CONCLUSIONS: Results of this study suggest that nursing home residents have lower levels of HRQOL than the general population except for bodily pain and role limitation: physical and emotional. Health care professionals should routinely assess nursing home residents` HRQOL and design interventions that may improve HRQOL in this vulnerable population.

221/Abstract 1326 Quality of Life (QoL0 in Mothers Taking Care of Infants ~Comaprison of Japan and North Korea~ Rika Hayashida, Makiko Yamasaki, Michiko Kobayashi, Yoshiko Naka, Siebold University of Nagasaki, Nagayocho, Nagasaki, Japan, Park Eun-Sook, Im Yeo-Jin, Im Hye-Sang, Korea University, Seongbuk-Gu, Seoul, Korea, Takashi Mandai, Japanese Society of Quality of Life Research, Kobe, Hyogo, Japan AIMS: The purpose of this study was to investigate the difference of QOL in mothers taking care of infants between Japan and Korea. METHODS: 377 mothers who take care of infants under 12 months (200 mothers in Japan and 177 ones in Korea) participated in this study. Our original self-administered questionnaire consisted of 42 questions divided into 9 categories. RESULTS: The Cronbach's alpha coefficients of our questionnaire were high enough to accept for clinical use: over 0.77 in Japan and over 0.66 in Korea in all categories. Our questionnaire contained 11 main factors which matched the 9 categories quite precisely both in Japan and Korea. There were significantly positive correlations between mother-child interaction and psychological conditions (p<0.01), well-being and childcare function and control (p<0.01) etc., respectively both in Japan and Korea. Average QOL levels on 8 categories were more excellent in Japan compared with the ones in Korea (p<0.01) except for the category of financial circumstances. In 8 categories, there were little average differences in the categories of psychological conditions and mother-child interaction between in Japan and Korea. CONCLUSIONS: These findings indicate that our original QOL questionnaire has a sufficient reliability and validity to use for childrearing mothers in all over the world. In this study, we recognize that there are no significant differences in the categories in mother-child interaction and psychological conditions which are most important not only in Japan and Korea but also the countries in all over the world. For the excellent QOL child care, mother-child interaction is most important for the child-rearing mothers from the viewpoint of their psychological conditions, childcare function and control. We must pay more attention to many kinds of problems child-rearing mothers have at the various developmental stages of children.

220/Abstract 1687 Health-Related Quality of Life Among Norwegian Nursing Home Residents Compared with the General Population Jorunn Drageset, Health and Social Sciences, Bergen University College, Bergen, Norway, Gerd Karin Natvig, Geir Egil Eide, Public Health and Primary Health Care, University of Bergen, Bergen, Norway, Elisabeth C. Clipp, Duke University School of Nursing, Duke University Medical Center, Durham, USA, Harald A. Nygaard, Margareth Bondevik, Public Health and Primary Health Care, University of Bergen, Bergen, Norway, Monica W. Nortvedt, Health and Social Sciences, Bergen University College, Bergen, Norway AIMS: Relatively little is known about the health-related quality of life (HRQOL) of nursing home residents versus that of the general population. This study compared HRQOL among elderly nursing home residents in Bergen, Norway with norms from the general population of Norway. METHODS: This study used a two-group cross-sectional comparative design. The samples comprised nursing home residents (n = 227, response rate 90%) from all nursing homes in Bergen, Norway and a representative sample of the Norwegian population (n = 1137). The inclusion criteria for nursing home residents were: >65 years old; >6 months` residence in a nursing home; Clinical Dementia Rating <0.5; and ability to communicate. The researcher administered the SF-36 Health Survey and several sociodemographic items to each respondent (65-102 years old). Trained nurses assessed participants` mental capacity using the Clinical Dementia Rating Scale. Data for the general population derived from Statistics Norway. A representative sample of 1137 Norwegian citizens aged 65-102 years returned postal questionnaires. The general linear model univariate analysis of variance procedure was conducted to identify possible differences in HRQOL between the nursing home residents and the general population, controlled for age, sex, marital status and education. RESULTS: The nursing home residents reported significantly lower HRQOL scores than the general population on all SF-36 subscales except bodily pain and role limitation: physical and emotional. After adjustment for age group, sex, marital status and education, the groups differed significantly on

222/Abstract 1306 Relationship of Health-Related Quality of Life (HRQOL) and Population-Based Surveillance (Healthy Japan 21) for Japanese Mothers Yuko Sasaski, Internal Medicine and Rehabilitation Science, Tohoku University Graduate School of Medicine, Sendai, MIyagi, Japan, Mieko Henmi, Keiko Ito, Ishinomaki Public Health Center, Isinomaki, Miyagi, Japan, Masayuki Kanazawa, Yoko Goto, Hitomi Kataoka, Masahiro Kohzuki, Internal Medicine and Rehabilitation Science, Tohoku University Graduate School of Medicine, Sendai, Miyagi, Japan AIMS: Population-based surveillance of health-related quality of life (HRQOL) is needed to promote the health and quality of life(QOL) of Japanese residents and to monitor progress in achieving the overall

A-140

Healthy Japan 21 goals. This study aims to examine the correlation between Health-Related Quality of Life_HRQOL and public heath action in Japanese mothers and children. METHODS: A total of 1517 (18-49 years old) mothers in Ishinomaki City in Japan participated in the study. HRQOL measures assessed self-rated health; (the HRQOL-4). 1) Would you say that in general your health is excellent, very good, good, fair, or poor? (Self-rated health), 2) Now thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good? (Physically unhealthy days), 3) Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good? (Mentally unhealthy days), 4) During the past 30 days, for about how many days did poor physical or mental health keep you from doing your usual activities, such as self-care, work, school, or recreation? (Activity limitation days). RESULTS: A total of 20.9% of older mothers (aged >40 years) reported fair or poor health, whereas only 9.2% of young adults (aged 20-29 years) reported fair or poor health. Adults aged 20-29 years reported the highest number of mentally unhealthy days. As age increased, mean number of physically unhealthy days, overall unhealthy days, and activity limitation days increased, and mentally unhealthy days decreased. Mothers feel that bringing up a child is pleasant ware 87.7%. Especially older mothers reported more physically unhealthy days and activity limitation days, whereas younger mothers reported more mentally unhealthy days. CONCLUSIONS: Findings indicated that over half of mothers screened positive for significant psychological distress and that this was associated with low levels of family support and with bringing up a child with higher levels of challenging behavior. In addition, population health surveillance for Healthy Japan 21 should continue to refine and validate HRQOL, mental status, and self-reported health measures.

RESULTS: Of 87 dyads, the majority of patients (71%) and caregivers (70%) were Afro-American. Caregivers were 83% female, 58% spouse/partner, 69% spent time daily with patient, and 46% had limited literacy. Significant mean difference scores (SD) between perspectives included: EQ-VAS, 2.5 (SD 12.8, p<0.05), EORTC emotional functioning, 6.0 (SD 12.8, p<0.001), and EORTC physical functioning, 3.1 (SD 13.4, p<0.05). A general mean score trend was patient > proxy-patient > proxy-caregiver among proxies who were functionally literate, while proxies with limited literacy tended to perceive less difference between proxy perspectives. CONCLUSIONS: The viewpoint of a proxy rating of HRQL can significantly affect the HRQL assessment provided by informal caregivers, but may depend upon domain and disease state. However, results suggest that caregivers with low literacy may less likely to rate HRQL differently based on proxy viewpoint than functionally literate caregivers, a finding that requires further exploration.

224/Abstract 1531 Responsiveness and Minimal Important Differences of the GSRS Total Score in Renal Transplant Patients Leah Kleinman, Center for Health Outcomes Research, United BioSource Corporation, Seattle, WA, Wen-Hung Chen, Dennis Revicki, Center for Health Outcomes Research, United BioSource Corporation, Bethesda, MD, Anne Kilburg, PH4325, HE & OR Transplant, Novartis Pharma AG, Basel, Switzerland, Ari Gnanasakthy, Health Economics and Outcomes Research, Novartis Pharmaceuticals, East Hanover, NJ AIMS: Renal transplant (RT) recipients often experience various gastrointestinal side effects associated with immunosuppressive therapy. Symptom burden and its changes due to innovative treatment approaches can be measured by patient-reported outcomes. This study evaluated the responsiveness and clinical significance of the Gastrointestinal Symptom Rating Scale (GSRS) total score in RT patients. METHODS: A secondary analysis was performed using 2 longitudinal study samples of RT recipients with and without GI complaints (PROGIS with 278 subjects followed over 4-6 weeks and US02 Study with an interim 305 subjects followed over 12 weeks). The GSRS contains 15 gastrointestinal symptoms rated on 7 point Likert scales; the total score equals the average of the items. All subjects completed the GSRS at baseline and follow-up. Patient and clinician-rated global change in symptoms was assessed at end of study. Responsiveness was evaluated using effect sizes. Minimal important difference (MID) was determined by comparing mean baseline to follow-up changes in total scores for the stable group compared with a small improvement group and using standard error of measurement (SEM). RESULTS: Based on patient and physician ratings, effect sizes in responders were 1.06 and 1.02, respectively. For PROGIS differences in mean change of total score from baseline to end-of-study was 0.9 and 0.7 points for patient and clinician global ratings between the stable and small improvement groups, respectively. For US02 differences were 0.1 and 0.3 for patient and clinician global ratings, respectively. Differences in mean change scores between the small and moderate improvement groups ranged from 0.2 to 0.6 points. SEMs ranged from 0.33 to 0.36 across the two studies. The majority of the estimates ranged between 0.30 and 0.40 points for the MID. CONCLUSIONS: The GSRS total score demonstrated high responsiveness in changes of the overall GI symptom burden of RT patients. The MID for the GSRS total score is 0.3 to 0.4 points, and further research is needed to confirm these results.

223/Abstract 1586 Proxy Assessment of HRQL: Does Perspective Matter and Is Literacy Important? A. Simon Pickard, Hsiang-Wen Lin, Maria C. De Leon, Center for Pharmacoeconomic Research, University of Illinois, Chicago, IL, Roohollah Sharifi, Urology, University of Illinois at Chicago, Chicago, IL, Shih-Ying Hung, Pharmacy, University of Wisconsin Madison, Madison, WI, Zhigang Wu, Urology, UIC and Jesse Brown VA, Chicago, IL, Whitney Witt, Buehler Center for Aging, Northwestern University, Chicago, IL, Sara Knight, Urology and Psychology, UCSF - San Francisco VA, San Francisco, California, Charles Bennett, Jesse Brown VA, Northwestern University, Chicago, IL AIMS: The objectives of this study were: (1) to determine whether informal caregivers provide different ratings of patient HRQL according to the proxy viewpoint elicited (either from the proxypatient or proxy-caregiver view); (2) to investigate factors predictive of differences in assessments by proxy based on viewpoint (i.e. the intra-proxy gap). METHODS: A cross-sectional study design was used to recruit prostate cancer patients and their informal caregivers (dyads) from the Jesse Brown VA system in Chicago, Illinois. Dyads assessed HRQL using the EQ-5D and EORTC QLQ-C30+3, and each caregiver provided an assessment both from the proxy-patient and proxy-caregiver perspectives. Health literacy was assessed as limited or functional literacy based on a cutoff score of 60 on the rapid estimation of adult literacy in medicine (REALM) measure. Analysis focused on systematic differences between perspectives and identification of significant predictors of those differences, such as caregiver relationship to patient, depressive symptoms and literacy.

A-141

preference. Meta-analysis techniques will then be used to synthesize the actual scores with these judgements and estimate large, moderate, small and trivial differences on each QOL scale. RESULTS: 910 papers were identified through the literature search. 392 papers contain relevant QLQC30 score comparisons. These papers cover 13 cancer sites and a wide range of treatment/specialist areas within cancer. An expert panel consisting of approximately 45 reviewers is currently being convened. This paper will report concordance between experts reviewing the same papers, the approach for incorporating uncertainty in expert judgements and an update on the development of methods to incorporate patient opinion into the guidelines. The novel application of meta analysis for interpretation of QOL scores will also be detailed. CONCLUSIONS: The guidelines will inform sample size calculations for QOL studies using the QLQ-C30 and aid interpretation of group differences from the QLQ-C30 scores. The methodology being developed in this project will also be applicable to the development of guidelines for other widely used and published QOL measures.

225/Abstract 1578 Weight-Related Quality of Life Improvements with Rimonabant in Dyslipidemic Patients are Clinically Meaningful Ronette L. Kolotkin, Community and Family Medicine, Duke University Health Systems, Durham, NC, Chantal Poinard, Global Health Outcomes, Sanofi-Synthélabo Recherche, Paris, France, Ross D. Crosby, Biomedical Statistics, Neuropsychiatric Research Institute, Fargo, ND AIMS: Rimonabant is the first selective CB1-receptor blocker developed to manage cardiometabolic risk factors. In the RIO-Lipids trial, 1-year rimonabant 20 mg administration showed statistically significant improvement in weight-related quality of life compared with placebo. Translation of health-related quality of life results into clinical meaningful terms is becoming mandatory. This study compares 4 methods of evaluating meaningful change on IWQOLLite total score in a randomized clinical trial of rimonabant 20 mg. METHODS: Meaningful change in IWQOL-Lite total score was evaluated in 605 (58.8% female; mean age = 47.9) overweight/obese dyslipidemic patients (mean BMI = 33.1 kg/m2) with 3 distributionbased methods:1.0 SEM (standard error of measurement), 1.96 SEM, 0.5 effect size, and an integrated method which combines distribution-based and anchor-based methods and takes baseline severity and regression to the mean into account (Crosby et. al, 2004). Patients (ITT population) were classified as improved/unchanged/deteriorated for each method based upon 1-year IWQOL-Lite total score changes from baseline. Receiver Operating Characteristics (ROC) analysis was used to determine which of these methods was most closely associated with 1-year weight loss. RESULTS: All methods of determining meaningful change in weight-related quality of life produced statistically significant and similar results. The Crosby algorithm and .5 Effect Size methods were most strongly associated with weight loss (p=.001 and p<.001, respectively; area under the curve for weight loss = .657 [SE = .022.024] for both methods). CONCLUSIONS: Patients on rimonabant 20 mg exhibited a greater percentage of meaningful improvement than patients on placebo at 1-year.

227/Abstract 1677 Psychiatric Morbidity Among Patients with Carcinoma of the Oesophagus or the Gastro-Oesophageal Junction -A Prospective, Longitudinal Evaluation Henrik Bergquist, Magnus Ruth, Eva Hammerlid, Otorhinolaryngology, Sahlgrenska University Hospital, Göteborg, Sweden AIMS: Carcinoma of the oesophagus is often diagnosed at a late stage and is related to severe morbidity and a low 5-year survival rate. Previous studies have reported low health-related quality of life and high suicide rates for these patients. The present study aimed to prospectively evaluate the occurrence of psychiatric morbidity in patients with newly diagnosed carcinoma of the oesophagus or gastro-oesophageal junction. METHODS: 94 consecutive patients (median age 66, range 45-88 years) with all stages of the disease were included. The prevalence of psychiatric morbidity was evaluated with the Hospital Anxiety and Depression Scale (HADS) questionnaire at inclusion and 1, 2, 3, 6 and 12 months later. RESULTS: HADS- scores indicating anxiety disorder and/or depression at inclusion were reported by 42% of the patients. Lower proportions of patients with scores indicating anxiety disorder were found at the 1- (p<0.001), 2- (p<0.01), 6- (p<0.01) and 12 months (p<0.01) follow-ups as compared to baseline. Larger proportions of patients with scores indicating anxiety disorder were found at the 1(p<0.01) and 6 months (p<0.05) follow-ups among those with a duration of symptoms exceeding 6 months before diagnosis. Apart from this, the proportion of patients with HADS scores indicating psychiatric morbidity was stable over time with no apparent influence by sociodemographic or clinical data. CONCLUSIONS: Psychiatric morbidity is common among oesophageal cancer patients, both at inclusion and over time regardless of the cancer-therapy given. The findings stress the importance of monitoring the patients' mental health and offer adequate psychological care when needed.

226/Abstract 1519 Evidence-Based Interpretation Guidelines: A Novel Approach to Making Quality of Life Data Count Kim Hawkins, Clinical Trials Research Unit (CTRU), University of Leeds, Leeds, UK, Madeleine King, Centre for Health Economics Research and Evaluation, University of Technology, Sydney, Australia, Peter Fayers, Public Health, University of Aberdeen, Aberdeen, UK, Galina Velikova, Cancer Research UK Clinical Centre, Marrissa Martyn St-James, Julia Brown, Clinical Trials Research Unit (CTRU), University of Leeds, Leeds, UK AIMS: This project aims to produce evidence-based interpretation guidelines for the QLQ-C30 questionnaire using existing published data from the questionnaire. The guidelines will estimate small, moderate and large differences in scores from the QLQ-C30. METHODS: A literature search has been carried out to identify all sources of QOL scores from the QLQ-C30. Comparisons of scores over time or cross-sectional comparisons by clinical anchors have been extracted. An expert panel consisting of health care professionals with knowledge of cancer, its treatment and the QLQC30 questionnaire will be used in the next stage of the process. The expert panel will review the papers with the QOL results masked, and will judge the expected size of differences in QOL measured by the QLQ-C30 subscales. Patient opinion will also be incorporated into the interpretation guidelines, possibly by developing scenarios and eliciting patient

228/Abstract 1591 Health- Related Quality Of Life Predictors In Breast Cancer Berta Varela, Laura Schwartzmann, Medical Psychology, Universidad de la República, Montevideo, Montevideo, Uruguay, Bettina Caporale, Medical Psychology, Universidad de la República, Montevideo, Uruguay, Ana galain, Maria Abreu, Medical Psychology, Universidad de la República, Montevideo, Montevideo, Uruguay, Rafael Alonso, Quantitative Methods, Universidad de la República, Montevideo, Uruguay

A-142

(p=0.001), had worse performance scores (p<0.01) and were more likely to have squamous cell tumours (p<0.001). Baseline compliance was 73% and it improved thereafter. At baseline, patients selected for chemoradiotherapy reported worse HRQL than surgery patients, with significantly worse dysphagia scores (58 vs. 32 p=0.001). After surgery, most domains of HRQL were significantly worse than during chemotherapy. In the non-surgical group, however, HRQL scores deteriorated during concurrent radiotherapy to similar levels as experienced after surgery. HRQL recovery was achieved by the follow-up assessment except for persistent problems with diarrhoea and reflux after surgery. CONCLUSIONS: Where clinical outcomes are similar, the role of HRQL is critical to decision-making and, for poor prognostic conditions such as esophageal cancer, such information is required for informed consent. This exploratory nonrandomised study provides important data to inform this debate in esophageal cancer.

AIMS: The process of adjustment to breast cancer requires coping efforts to face diagnosis and treatments. Present longitudinal study aims to identify impact of illness and treatment on Health Related Quality of life (HRQL)as well as influence of coping strategies and emotional state over HRQL in breast cancer women. METHODS: Sample included 86 patients, mean age 57. Study was prospective and Patients self-reported through a battery of scales applied after diagnosis, surgery, chemotherapy and radiotherapy . Applied scales were Hospital Anxiety and Depression Scale (HADS), Profile of mood states ( POMS), Mental Adjustment to Cancer (MAC) and SF36 Nonparametric correlations were used to analyse the data. RESULTS: No significant anxiety or depression were detected by the HADS in group applications, in spite of individual differences, showing higher rates of anxiety after diagnosis. POMS scores varied during process maintaining close correlation with coping strategies evaluated by MINIMAC (MM). Tension anxiety(.65, p.000),depression-dejection (.55, p.000) and total POMS scores) correlated positively with ANXIOUS PREOCCUPATION (MM) (.55 p.000) and helplessness-hopelessness(.50, p-000) in all treatments, increasing during chemotherapy. Fighting spirit (MM) exhibited a negative correlation with POMS total scores (-.40, p.006) and a positive correlation with vigor-activity (.30, p.0.1). Emotional states (POMS) and coping (MM) had a strong impact on HRQL evaluated by SF36. TA and DD ( POMS)AND AP y HH (MM) were significant negative predictors of HRQL. Vitality (poms) AND fighting spirit were positive predictors correlating with MH; GH; SF, V ( .40 to.70, p.000) CONCLUSIONS: Significant positive and negative predictors of HRQL were found which can be addressed in psychosocial interventions. Emotional states and coping mechanisms allowed a better identification than anxiety or depression as psychopathology. Results point to the importance of emotional states and coping mechanisms for HRQL reinforcing health concept as not merely absence of psychopathology but positive wellbeing.

230/Abstract 1502 Comparison Between Cancer-Specific Instruments in Head and Neck Portuguese Cancer Patients: Preliminary Data Augusta Silveira, FCS, Universidade Fernando Pessoa, Porto, Portugal, Alexandra Oliveira, Medical Oncology, Hospital de Santa Maria da Feira, Santa Maria da Feira, Portugal, Francisco L. Pimentel, Secção Autónoma Ciências Saúde, Universidade de Aveiro, Aveiro, Portugal AIMS: Quality of life assessment in head and neck cancer patients is not a common practice in Portugal. Using portuguese version of specific instruments already validated, the aim of this study is a comparison of two major oncology questionnaires: QLQ-C30 from EORTC and its specific version for head and neck QLQ- H&N35 and FACT- H&N. METHODS: Two hundred and two head and neck cancer patients from Oncology Portuguese Institute of Porto (IPOFGCROP) completed the QLQ-C30, QLQ- H&N35 and FACT- H&N . All patients asked accepted to participate in the research. It was evaluated the agreement level between FACT-G total score and QLQ-C30 total score and between HNC total score from FACTH&N and QLQ- H&N 35 total score. Internal consistency was evaluated using Cronbach´s ± coefficient. RESULTS: Population characteristics: Two hundred and two head and neck cancer patients with an age average of 59,4 years (SD-12,1). 83,3% were males and 16,7% females. It was observed an excellent agreement between FACT-G total score and QLQ-C30 total score (r = 0,80; ICC= 0,82) and a good agreement between FACT- H&N total score and QLQH&N 35 total score (r = 0,73; ICC= 0,70). We found a good internal consistency for QLQ-C30 total score(±= 0,87), QLQ- H&N 35 total score(±= 0,90), FACT-G total score (±= 0,76) and a worse internal consistency for FACT- H&N total score. CONCLUSIONS: The data support the psychometric properties described in literature. Both are good evaluating tools and its selection should take in account administration simplicity and data interpretation.

229/Abstract 1510 Clinical Equipoise and the Role of Health-Realted Quality of Life (HRQL) in Defining Treatment Choice: An Example of HQRL Comparisons During Potentially Curative Treatments for Locally Advanced Esophageal Cancer Pernilla Viklund, Molecular Medicine and Surgery, Karolinska Institute, Stockholm, Sweden, Kerry N. Avery, Social Medicine, University of Bristol, Clifton, Bristol, UK, C Paul Barham, Surgery, United Bristol Healthcare Trust, Bristol, UK, Chris Metcalfe, Social Medicine, University of Bristol, Bristol, UK, Jo Nicklin, Surgery, United Bristol Healthcare Trust, Bristol, UK, Derek Alderson, Clinical Sciences at South Bristol, Jenny L. Donovan, Social Medicine, University of Bristol, Bristol, UK, Stephen J. Falk, Jane M. Blazeby, Surgery, United Bristol Healthcare Trust, Bristol, UK AIMS: This study explored the pivotal role of health-related quality of life (HRQL) in defining treatment choice between two clinically equivalent, but different, potentially curative treatments for esophageal cancer. METHODS: Consecutive patients with locally advanced esophageal cancer undergoing potentially curative treatment completed the European Organization for Research and Treatment of Cancer (EORTC) QLQ-C30 and QLQ-OES18 questionnaires to assess HRQL and esophageal symptoms. Patients underwent definitive chemoradiotherapy or neoadjuvant treatment and major surgery. Linear regression, adjusting for age, gender, performance status, tumour stage and tumour type, compared HRQL before, during and after treatment. RESULTS: 132 patients, 51 chemoradiotherapy and 81 neoadjuvant and surgery, underwent treatment. Those undergoing chemoradiotherapy were older

231/Abstract 1676 The Responsiveness of the Beck Depression Index, SF-36, SF-6D and HUI2/3 in active and latent Tuberculosis Fawziah Marra, British Columbia Centre for Disease Control, Carlo A. Marra, Lindsey Colley, Pharmaceutical Sciences, University of British Columbia, Vancouver, B.C., Canada, Susanne Moadebi, British Columbia Centre for Disease Control, University of British Columbia, Vancouver, BC, Canada, J. Mark Fitzgerald, Centre for Clinical Epidemiology and Evaluation, Vancouver Coastal Health Research Institute, Vancouver, BC, Canada, R. Kevin Elwood, British

A-143

substantial burdens in pain, cognition (problem solving, memory), emotion, and ambulation. Among the 8 attributes, 3% had no attributes affected; 18% had only 1 attribute affected, 16% had 2 attributes affected, 19% had 3 attributes affected, and 44% had 4 to 6 attributes affected. At baseline the mean overall HUI3 score was 0.22. At the 4-week follow-up improvements were seen for emotion and pain; the mean overall score was 0.27. The gain in overall HRQL (paired t-test, p=0.08) was clearly clinically important. CONCLUSIONS: Conclusions. HUI confirmed the high burden of illness in patients assigned to CCCM. Burdens in ambulation, emotion, and pain were well recognized but the quantification of the burdens was useful. The burden in cognition, however, had not been recognized and has important implications for the organization of care and management of patients. The gain in overall HRQL seen over the initial 4-week period in CCCM provides preliminary evidence that reductions in utilization associated with CCCM are apparently not at the expense of patient outcome.

Columbia Centre for Disease Control, University of British Columbia, Vancouver, BC, Canada, John Brazier, Health Economics Group, ScHARR, University of Sheffield, Sheffield, South Yorkshire, UK AIMS: Tuberculosis (TB) remains a public health threat worldwide. Treatment of active TB infection involves the administration of highly effective, yet toxic medications over a period of several months. Treatment of latent TB infection is more benign with only one drug administered. Patients_ quality of life (QoL) and the responsiveness of QoL instruments throughout TB treatment have not been previously characterized. METHODS: Active and latent TB patients attending an urban treatment centre were enrolled. Patients completed the Health Utilities Index 2 and 3 (HUI2/3), the Short Form 36 (SF-36), and the Beck Depression Index at baseline and 3 and 6 months thereafter. External criteria for change were a patient transition question, categories of the patient global assessment of disease activity VAS, and a dichotomous (Y/N) response variable regarding drug toxicity. Responsiveness was characterized by the effect size (ES), the standardized response mean (SRM), and an approach using general additive modelling. RESULTS: : 61 active and 51 latent TB patients completed all questionnaires at baseline and at 6 months. Most (68%) active TB patients reported improvement using the external criteria whereas most (71%) with latent TB reported remaining status quo. Few patients reported worsening. The SF-36 Physical and Mental Component Scores and the Beck Depression Index were the most responsive of instruments. The SF6D and HUI3 were more responsive than the HUI2 and the absolute magnitude of change in those who improved was highest in for the HUI3 (0.12 vs. 0.07 for the SF-6D and 0.04 for the HUI2). No instrument appeared to be responsive to the occurrence of drug toxicity. CONCLUSIONS: The preference-based measures did not appear to be as responsive as the SF-36 or the Beck Depression Index. The SF-6D and HUI3 had similar responsiveness but the differences in the magnitude of the absolute change scores may have important implications for cost-effectiveness analyses.

233/Abstract 1362 Comparison of Patient-Based Outcomes in Clubfoot Treatment: A Long Term Follow-Up Study Hiroko Matsumoto, Joshua E. Hyman, Samuel Dellenbaugh, Pediatric Orthopaedic Surgery, Columbia University, New York, New York, Jose Morcuendu, Orthopaedic Surgery, University of Iowa, New York, New York, Michael G. Vitale, Pediatric Orthopaedic Surgery, Columbia University, New York, New York AIMS: The treatment of patients with Talipes Equinovarus, commonly known as clubfoot, has been controversial due to the difficulty of assessing their long-term outcomes. The purpose of this study is to compare patient-based measures of long term outcomes in two cohorts of patients, one treated with Ponseti method, and the other treated with the open surgical method. METHODS: The study population consisted of 24 patients who underwent open surgical repair and 16 patients who were treated with Ponseti method for their idiopathic clubfoot deformity. The patients with the open surgical method were treated at the Division of Pediatric Orthopaedics at Columbia University. The patients with Ponseti method were treated at the Department of Orthopaedics at University of Iowa. The Child version of the Child Health Questionnaire (CHQ CF-87) 17 was utilized in this study. The means of all domains of the CHQ in the two patient cohorts were compared using independent samples t tests. RESULTS: There were no significant differences between two groups in gender, race and age at treatment; however, there was significant difference in age at follow up (p<0.001). The mean age for the surgical patients were 16.4 ± 1.5 years old while Ponseti patients were 26.6 ± 5.1 years old in average. The patients treated with the Ponseti method patients had significantly higher pain than the patients with open surgical method (p=0.036). The patients treated with the Ponseti method also scored significantly higher in General Health and Family Cohesion domains (ps=0.027) and significantly lower in Mental Health domain (p=0.047) on the CHQ than the patients treated with the open surgical method. CONCLUSIONS: Surprisingly, the patients who were treated with Ponseti method reported higher pain than the patients who had open surgery. This difference may be due to the fact that the patients with Ponseti method had the follow-up 10 years longer on average than the patients with open surgical method. Psychosocial outcomes might attribute to environmental factors rather than the treatment methods.

232/Abstract 1244 Using Health-Realated Quality of Life Measures to Improve Patient Care David H. Feeny, Kaiser Northwest Center for Health Research, Health Utilities Incorporated, Portland, OR, Richard Meenan, Center for Health Research, Kaiser Permanente Northwest, Portland, OR, David Labby, CareOregon, CareOregon, Portland, OR, Mark Spofford, David Mosen, Center for Health Research, Kaiser Permanente Northwest, Portland, OR, Rebecca Ramsay, CareOregon, CareOregon, Portland, OR AIMS: Aim. To examine the usefulness of a health-related quality of life measure in an assessment of a patient-management program. METHODS: Methods. CareOregon, a Medicaid provider in Oregon, USA, is using the Health Utilities Index Mark 3(HUI3) in a comprehensive assessment of their complex care case management (CCCM) program. CCCM has lowered plan-supported utilization and cost but data on patient-reported outcomes are lacking. A consecutive series of 63 new enrollees in CCCM were assessed in telephone interviews using HUI at baseline and 4 weeks later. HUI3 includes 8 attributes (vision, hearing, speech, ambulation, dexterity, emotion, cognition, pain) with 5 or 6 levels per attribute. Overall HUI3 scores are on a scale in which dead = 0 and perfect health = 1. Differences (changes) of 0.03 or more in overall score are regarded as clearly clinically important (CID). Enrollment in the study is ongoing. RESULTS: Results. The response rate was 100%. 68% of patients were female; the average age was 61. At baseline HUI3 identified

A-144

population; and to study the influence of socio-demographic (gender; age) and clinical variables (treatment regimen; disease duration) in patients perceptions of QoL. METHODS: Methods: A sample of 289 patients (n=59-Psychiatry; n=31-Gynaecology; n=73-Rheumatology; n=72-Oncology; and n=54-Health Centres) and of 315 individuals of the general population was assessed with the Portuguese version of WHOQOL-100. RESULTS: Results: The Portuguese version has acceptable internal consistency across patient groups (alpha range=.80-.94 across domains). Globally, psychiatric patients reported worst QoL in 4 of 6 domains, and Rheumatology patients reported worst QoL in 2 domains. Two domains showed greater discriminative power in comparison with the controls (Physical: F=30,09;p<.001; Independence: F=78,89;p<.001). Amongst patients of different health/hospital services 4 domains show discriminative power: Psychological; Independence; Social and Environmental. Women reported worst physical and psychological well-being and better spirituality. Older people (>45) revealed poorer QoL physically (p<.01) and in the level of independence (p<.001). In-patients reported better QoL, although statistically non significant, except in Independence domain (p<.001). Patients with a disease duration of less than a year reported better QoL in all domains. CONCLUSIONS: Conclusions: WHOQOL-100 showed a good ability to distinguish between patients and non-patients, and differentiate between the diverse samples of patients. Results also show that in-patients reported better QoL suggesting a positive availability and quality of health care in hospital. They also have important implications in the type of care that is being provided, suggesting that women may benefit from additional psychosocial support.

234/Abstract 1408 Patient-Reported Satisfaction with Prescription and Over-theCounter Treatments For Restless Legs Syndrome (RLS) Michael O. Calloway, USP Health Outcomes, GlaxoSmithKline, Research Triangle Park, North Carolina, Martin O'Leary, Global Health Outcomes, GlaxoSmithKline, Susan C. Bolge, Outcomes Research, Consumer Health Sciences, Preston, New Jersey AIMS: RLS is a chronic neurological disorder characterized by an irresistible urge to move the legs. This US survey assessed patient satisfaction with prescription treatments for RLS. METHODS: In April 2006 US adults from the National Health and Wellness US studies and Late Speed Research ailment panel diagnosed with moderate-to-severe primary RLS and one or more prescription or OTC treatments for RLS for 2+ months, were invited to complete a web-based survey. Patients were grouped by their current RLS treatment: Requip® (ropinirole) only, other prescription medication(s) (Rx) only, or OTC medication only. Patient satisfaction ratings on: overall treatment, speed of relief, consistency of relief, and convenience of dosing were measured on 7-point Likert scales (1=extremely satisfied/convenient, 7=extremely dissatisfied/inconvenient). Difference of means t-tests were used to compare satisfaction scores by current treatment and regression models to assess the effects of current treatment on satisfaction. RESULTS: Of the 348 respondents, 58 (17%) used ropinirole only, 65 (19%) used another Rx only, and 1 person used an OTC only. Patients using ropinirole only, reported a higher level of overall satisfaction than those using other Rx only treatment (48.3% vs 35.4%); speed of relief (39.7% vs 29.2%), consistency of relief (39.7% vs 29.2%), and convenience of dosing (67.24% vs 66.15%). Controlling for confounders, ropinirole users reported significant overall satisfaction with treatment (Odds Ratio [OR]: -.853; 95% CI: -1.648, -.573; p=.036), greater satisfaction with speed of relief [OR]: -.986; 95% CI: -1.773, -1.991; p=.014), and greater satisfaction with consistency of relief [OR]: -.93; 95% CI: -1.724, -.136; p=.022). There was no difference with satisfaction with convenience of dosing [OR]: -.508; 95% CI: -1.293, 0.277; p=.205). CONCLUSIONS: Patients with RLS using ropinirole (FDA-approved for moderate-tosevere primary RLS) reported a higher degree of satisfaction with treatment across several aspects of medication use, compared to those taking other Rx medications.

236/Abstract 1563 Development, Validation, and Norming of the SF-10 for Children of Health Survey Renee N. Saris-Baglama, Michael A. DeRosa, Anastasia E. Raczek, Jakob B. Bjorner, John E. Ware, Research and Development, QualityMetric Incorporated, Lincoln, RI AIMS: To develop, validate, and provide 2006 norms for the SF-10 for Children Health Survey (SF-10), a 10-item, caregiver-completed, alternative short form of the Child Health Questionnaire (CHQ). METHODS: Sample 1: In 1994, a subset of participants in the 1990 U.S. National Survey of Functional Health Status that were parents/guardians of a child aged 5-18 (N=369) completed a child health mail survey that included the CHQ PF-50. Sample 2: In 2006, a representative U.S. general population sample of parents/guardians of a child aged 5-17, with or without a chronic condition or disability, was recruited through an online panel and completed a child health survey that included the SF-10. Participants were randomly assigned to internet (N= 2,353) or mail survey administration mode (N=1,288). Analyses: Sample 1 data were analyzed to derive the SF10, develop norm-based scoring of two summary scores [Physical (PHS) and Psychosocial (PSS)], and provide preliminary tests of reliability and validity. Sample 2 data were analyzed to further validate the tool and develop 2006 norms for children with and without chronic conditions and disability. RESULTS: Internal consistency reliabilities were satisfactory (PHS: Į =.70; PSS: Į =.78) and correlations between SF-10 and CHQ PF-50 summary scores were high (PHS: r =.91; PSS: r = .90). In tests of relative validity (RV) in discriminating among known-groups differing in physical and mental conditions, SF-10 summaries performed well in comparison with corresponding CHQ PF-50 scores (median RV = .99). The SF-10 detected significant impairment in health-related quality of life (HRQOL) for children with disabilities. CONCLUSIONS: The SF-10 for Children is an empirically sound

235/Abstract 1657 The World Health Organization Quality of Life Instrument (WHOQOL-100) Across Diseases: Results from the Portuguese Validation Study Marco D. Pereira, Faculdade de Psicologia e Ciências da Educação, Universidade de Coimbra, Portugal, Maria C. Canavarro, Faculdade de Psicologia e de Ciências da Educação, Adriano Vaz Serra, Faculdade de Medicina, Mário R. Simões, Daniel M. Rijo, Faculdade de Psicologia e Ciências da Educação, Manuel J. Quartilho, Faculdade de Medicina, Carla S. Gameiro, Faculdade de Psicologia e Ciências da Educação, Universidade de Coimbra, Carlos F. Carona, Núcleo Regional do Centro, Associação Portuguesa de Paralisia Cerebral, Tiago F. Paredes, Faculdade de Psicologia e Ciências da Educação, Universidade de Coimbra AIMS: The WHOQOL-100 was conceptualised as a generic measure of quality of life (QoL), but particularly useful for use with patients across varying medical conditions. The aim of the current study is to test the internal consistency of WHOQOL across diseases; compare the QoL of patients within disease groups and with the general

A-145

and practical alternative to longer forms for measuring HRQOL. Development of 2006 norms based on a large, diverse sample enhances the interpretation of scores. Internet-based administration of the SF-10 is feasible and with just 10 items, the SF-10 can be easily integrated into many settings.

between pre- and post-operation groups of patients showed significantly higher (worse) scores for concern, physical/social functioning and overall scores in the pre-operating patients. Almost all subscales showed desirable inter-scale correlations. CONCLUSIONS: The Iranian version of the RSVP was found to be a reliable and valid measure of vision related quality of life in patients with refractive error and now it can be used in epidemiological and clinical studies of health related quality life in this group of patients in Iran.

237/Abstract 1056 Measurement Invariance of a Subjective Well-Being Model Vítor Bertoquini, José Pais Ribeiro, Psychology, Faculdade de Psicologia e de Ciências da Educação UP, porto, Portugal

239/Abstract 1146

AIMS: The assumption of measurement invariance is often made yet rarely is tested. The objective of this study is to analyse the measurement invariance of a Subjective Well-Being model across two samples. METHODS: The data were obtained as part of an ongoing longitudinal study on the Psychological determinants of Subjective Well-Being. The samples consists of 1334 college students aged between 17 and 52 (M=22.83 DP=4.85) and of 227 college students aged between 17 and 42 (M=23 DP=3.77). Life Satisfaction was measured with the SWLS and PANAS was used to assess the frequency of Positive and Negative Affect. To assess Subjective Well-Being a composite index was created from these three components [NA _ (PA + SWLS)]. Confirmatory factor analysis (AMOS 5.0) was used to analyse the Subjective Well-Being model. The fit of each model was compared to the previous, less constrained model to assess invariance. To establish configural invariance, a freely specified model is tested for each dataset. To establish metric invariance, a model with equal factor loadings is tested. RESULTS: The results of this study indicate that the configural (equivalent structure) and metric invariance (equivalent factor loadings) were achieved. Strict invariance was rejected. CONCLUSIONS: The findings support the view of higher-order construct of Subjective Well-Being with three factors. The three factors of Subjective Well-Being seem to be validly measured and the partial factorial invariance across groups lends support for some measurement equivalence.

The Japanese Version of Sickness Impact Profile (SIP) in Patients with COPD Yoko Goto, Occupational Therapy, School of Health Sciences, Sapporo Medical University, Sapporo, Japan, Masahiro Kohzuki, Yuko Sasaki, Hajime Kurosawa, Internal Medicine and Rehabilitation Science, Tohoku University Graduate School of Medicine, Sendai, Japan, Tokutaro Sato, National Rehabilitation Center for Persons with Disa, Tokorozawa, Japan AIMS: Chronic obstructive pulmonary disease (COPD) is characterized by progressive airflow limitation which results in exertional dyspnea and physical disability, subsequently, causing difficulty in performing activities of daily living (ADL) and affecting their health-related quality of life (HRQOL). Therefore, it is important to see HRQOL in patients with COPD in movable or behavioral disability. Sickness Impact Profile (SIP) is a sensitive and behaviorally based measure of sickness-related dysfunction. METHODS: We made the Japanese version of SIP using backtranslation methods, and assessed reliability and validity by 134 patients with COPD. RESULTS: Homogeneity of the Japanese version of SIP was demonstrated by a high Cronbach , s alpha (SIP-Overall: 0.96, SIPPhysical: 0.95, SIP-Psychosocial: 0.93).The strong correlation coefficient by test-retest method identified the high reliability of the Japanese version of SIP (SIP-Overall: r=0.92, SIP-Physical: r=0.91, SIP-Psychosocial: r=0.90). In the criterion based validity research, significant correlation was observed between eight subscales of the MOS 36-item short-form health survey (SF-36) and SIP scores. CONCLUSIONS: We conclude that the Japanese version of SIP is a feasible and clinimetrically sound measure to assess HRQOL for patients with COPD. Furthermore, we need to strive to make the abbreviation of the Japanese version of SIP.

238/Abstract 1125 The Refractive Status and Vision Profile (RSVP) Translation and Validation Study of the Iranian Version Arezoo Khadkhoda, Iraj Ahadzadegan Ahani, Eye Research Centre, Farabi Eye Hospital, Tehran University of Medical Sciences, Tehran, Iran, Ali Montazeri, Mental Health, Iranian Institute for Health Sciences Research, Tehran, Iran

240/Abstract 1650

AIMS: To translate and test the reliability and validity of the Refractive Status and Vision Profile (RSVP), a vision related quality of life questionnaire in Iran. METHODS: Forward-backward translation method, committee review and pilot testing were performed to develop the final Iranian version of the RSVP. Seventythree consecutive patients with refractive error before or after refractive surgery at the LASIK ward of Farabi Eye Hospital in Tehran, Iran completed the questionnaire. Reliability was measured by internal consistency (Cronbach's alpha) and test-retest reliability (intra-class correlation coefficient). Validity was evaluated by correlation between the different RSVP subscales, known groups comparison analysis and correlation between the subscales versus global items and conventional clinical measures. RESULTS: Internal consistency was high (Cronbach's alpha: 0.71-0.92; except for the subscale expectation, alpha: 0.6). Test-retest reliability of subscales and overall scores, as estimated by intra-class correlation coefficient, was high except for optical problems and glare. Comparisons

Assessment of the Quality of Life in the Portuguese Population: Reliability and Validity Studies of the WHOQOL Maria C. Canavarro, Faculdade de Psicologia e de C. da Ed., Universidade de Coimbra, Portugal, Marco D. Pereira, Faculdade de Psicologia e de C. da Ed., Daniel B. Rijo, Faculdade de Psicologia e de C. Ed, Sofia C. Gameiro, Faculdade de Psicologia e de C. da Ed., Adriano S. Vaz-Serra, Faculdade de Medicina, Mário R. Simões, Faculdade de Psicologia e de C. da Ed., Manuel J. Quartilho, Faculdade de Medicina, Tiago M. Paredes, Faculdade de Psicologia e de C. da Ed., Universidade de Coimbra, Carlos F. Carona, Núcleo Regional do Centro, Associação Portuguesa de Paralisia Cerebral AIMS: Aims: Investigate the psychometric properties of the Portuguese from Portugal version of the World Health Organization´s Quality of Life Instrument (WHOQOL-100)to Know if this version is a good assessment measure of the Quality of Life of the Portuguese

A-146

disabilities. Moreover, the results will shed light on sustainability, practicability, and usefulness of sensitivity based quality development.

Population. METHODS: Methods: The directives of the OMS for the validation of the WHOQOL in other international Centres were followed, in as far as the different phases and procedures. The application of the instrument, in the study of the quantitative field, was done on a sample of 604 subjects: 315 from the general public and 289 patients, from various hospital services and subject to different types of treatment. RESULTS: Results: The Portuguese (from Portugal) version has an acceptable internal consistency (alpha range=.84-.94 across domains) and time stability (r range from. .73.86, to p < .001), showing therefore good and reliable results. Also in the validity studies, the instrument showed good results, concerning the discriminative validation (all domains, with the exception of Domain 6 allow to differentiate patients from healthy subjects), validity of the construction (the correlations between the different domains are all statistically significant for p<.001; all the domains, with the exception of Domain 6, showing high and significant correlations with the general facet of the Quality of Life: r range from .61 (Physical) to .69 (Psychological and Level of Independence). A multiple linear regression in relation the general facet of the Quality of Life revealed that the Psychological, Level of Independence, Social Relationships and Environmen Domains appear to be better predictors of the quality required in general, according to a linear explanatory model of 67,6% of the variance. CONCLUSIONS: Conclusions: The reliability and validity studies done for Portuguese of Portugal of the original version of the instrument of the quality of life of the OMS (WHOQOL-100) to which four items were added, relating to the Portuguese facet, revealed good psychometric characteristics permitting the use of this version of WHOQOL in our country.

242/Abstract 1587 An International Field Study of Psychometrics of the EORTC QLQ-PR25 Questionnaire in Assessing the Health-Related Quality of Life (HRQL) of Patients with Prostate Cancer. George v. Andel, Urology, Onze Lieve Vrouwe Gasthuis, Amsterdam, The Netherlands, Andrew Bottomley, Data Center, EORTC, Brussels, Belgium, Sophie Fossa, Oncology, NORWEGIAN RADIUM HOSPITAL, Oslo, Norway, Fabio Efficace, Corneel Coens, Data Center, EORTC, Brussels, Belgium, Neil K. Aaronson, Psychosocial Research, THE NETHERLANDS CANCER INSTITUTE / AVL, Amsterdam, North Holland, The Netherlands AIMS: The aim of this study was to evaluate the psychometric properties of the EORTC QLQ-PR25, a questionnaire module (PR25) for assessing the HRQL of patients with prostate cancer. METHODS: The QLQ-PR25 and the core questionnaire, the QLQC30 were administered to 642 prostate cancer patients from 13 countries (11 languages) treated with curative (surgery, external radiation therapy, with or without hormonal therapy) or palliative (hormonal therapy only) intent. Assessments were performed before treatment, after 3 months and, in the case of palliative treatment, also after 6 months. The QLQ-PR25 assesses urinary-, bowel-, and sexual functioning as well as side-effects of hormonal treatment. A standard set of psychometric analyses was employed to evaluate the hypothesized scale structure of the PR25, its internal consistency reliability, and its responsiveness to change over time. RESULTS: Of the 642 patients 86 % met all eligibility criteria. Of the remaining 555 patients the compliance at baseline was 85%, at 3 months 77% and at 6 months 57%. The mean age of the sample was 67 year. Over 85% of patients completed the QLQ-C30 and PR25 within 15 minutes, 14% required assistance in completing the questionnaires, primarily due to vision problems, 20% reported that one or more questions was difficult or confusing, and 9% were upset by one or more questions (primarily related to sexuality). Multi-trait scaling analyses confirmed the hypothesized scale structure of the QLQPR25. Internal consistency reliability was satisfactory (Cronbach_s ± ranging from 0.70 to 0.86) except for the scale assessing side effects of hormonal treatment. Responsiveness to change over time was evidenced for all scales. Additional analyses are ongoing and will be presented. CONCLUSIONS: In general, the QLQ-PR25 demonstrates acceptable psychometric properties and can be used in conjunction with the QLQ-C30 in assessing the HRQL of patients with prostate cancer.

241/Abstract 1137 Sustainable Quality Development in Care and Service Institutions for People With Disabilities David Oberholzer, René Stalder, Institute for Special Education, University of Zurich, Zurich, Switzerland AIMS: In this paper, we outline a new perspective towards evaluating the quality of life for people with disabilities in the field of special education and health research. A research project at the University of Zurich is concerned with the evaluation of quality of life in people with disabilities, and conducted in close collaboration with institutions providing care and services for people with disabilities. The main purpose of the study is to develop a new instrument which efficiently and effectively guides the evaluation and operationalisation of quality of life in disabled people who live and work at those institutions, thereby adding an informative and sustainable element to existing measures of quality development and quality management. Consequently, the study uses a decidedly institutional focus. METHODS: As a first step, the concept of quality of life is transformed and adopted into a set of variables reflecting the specific requirements and conditions of the respective institutions. Secondly, the variables are weighed empirically and moulded to a sensitive network, using kybernetic sensitivity analysis. In the third and final step, this sensitive network is used to hypothetically identify efficient and effective measures for improving the quality of life in a given institution, and the resulting hypotheses are then tested in collaboration with the institution. RESULTS: A preliminary study showed the approach to be applicable and informative. An institution-specific framework for quality of life in people with disabilities was acquired and empirically verified. Sensitivity analysis allowed to infer institution-specific hypotheses indicating feasible improvements with respect to quality of life. CONCLUSIONS: The prospect of our research is to offer new possibilities for the evaluation of quality of life in people with

243/Abstract 1541 Does Presence of Meaningful QoL Data Influence Dermatology Outpatient Consultation? Sam Salek, Welsh School of Pharmacy, Cardiff Univeristy, Cardiff, UK, Alaw Roberts, Welsh School of Pharmacy, Andrew Finlay, Department of Dermatology, Cardiff University, Cardiff, UK AIMS: Previous studies have identified differences in patient and clinician perception in dermatology with regard to QoL. There is a need to improve clinician understanding and aid treatment decisionmaking. The aims were to determine whether clinicians used formal QoL information if provided by patients completing the Dermatology Life Quality Index (DLQI) questionnaire and if so, whether the information influenced treatment decision-making. METHODS:

A-147

Patients were recruited from a general dermatology secondary care clinic. The DLQI questionnaire was handed to adult patients to complete as they arrived for their appointment and were returned to the clinician. Clinicians had the opportunity to use DLQI information if they wished; they recorded whether DLQI information was used and whether it influenced treatment decision-making. RESULTS: 417 patients attended clinic and 242 (58%) questionnaires were completed. The mean score for DLQI was 7.0 (median=4.0, range 0_30). A further 26 were unevaluable. In 37 (15.3%) consultations DLQI information influenced the clinicians_ treatment decisionmaking. Patients with inflammatory skin disease had more impaired HRQoL in all areas measured by DLQI (p<0.01). The clinicians used DLQI information more with inflammatory skin patients. 98 consultations were observed to monitor QoL discussion. QoL discussion occurred in 85 (86.7%) consultations. In the observed consultations the domain of QoL most frequently discussed regarded symptoms of disease - question 1 (16 of 23 consultations by the specialist nurses (70%), 23 of 34 by the GPs with a special interest in dermatology (68%), 12 of 20 consultations by the clinical lecturer (60%), and 12 of 21 consultations by the registrar (57%)). CONCLUSIONS: The findings of this study indicate that clinicians may use formal HRQoL information if available. The data suggest that treatment discussions are influenced for those with high HRQoL impairment. Systematic patient-reported HRQoL outcomes may therefore be of benefit in routine clinics.

A-148

13th Annual Conference of the International Society for Quality of Life Research

Recommend Documents