Qual Life Res (2012) 21:1–132 DOI 10.1007/s11136-012-0248-x
ABSTRACTS
19th Annual Conference of the International Society for Quality of Life Research
International Society for Quality of Life Research 2012
Symposia
(2) Are generic preference-based measures valid in mental health? A mixed methods enquiry
(1) Implementing patient-reported outcomes in clinical practice: an expert panel discussion of current knowledge and unanswered questions
John Brazier, PhD, School of Health and Related Research, University of Sheffield, Sheffield, United Kingdom; Brendan Mulhern, MRes, School of Health and Related Research, University of Sheffield, Sheffield, United Kingdom; Jordi Alonso, PhD, Health Services Research Unit, IMIM-Institut de Recerca Hospital del Mar, Barcelona, Spain, and CIBER en Epidemiologı´a u Salud Pu´blica (CIBERESP), Barcelona, Spain; Janice Connell, BSc (Hons), School of Health and Related Research, University of Sheffield, Sheffield, United Kingdom; Dianna Papaioannou, MSc, School of Health and Related Research, University of Sheffield, Sheffield, United Kingdom; Gemma Vilagut, BSc (Hon), Health services Research Unit, IMIM-Institut de Recerca Hospital del Mar, Barcelona, Spain, and CIBER en Epidemiologı´a u Salud Pu´blica (CIBERESP), Barcelona, Spain
Claire F. Snyder, PhD, Johns Hopkins School of Medicine, Baltimore, MD, United States; Neil K. Aaronson, PhD, The Netherlands Cancer Institute & University of Amsterdam, Amsterdam, The Netherlands; Thomas E. Elliott, MD, Essentia Institute of Rural Health, Essentia Health, Duluth, Minnesota, United States; Joanne Greenhalgh, PhD, School of Sociology and Social Policy, University of Leeds, Leeds, United Kingdom; Michele Halyard, MD, Radiation Oncology, Mayo Clinic, Scottsdale, AZ, United States; Rachel Hess, MD, MS, Center for Research on Health Care, University of Pittsburgh, Pittsburgh, PA, United States; Deborah M. Miller, PhD, Mellen Neurology, Cleveland Clinic, Cleveland, OH, United States; Bryce B. Reeve, PhD, Lineberger Comprehensive Cancer Care & Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States; Maria-Jose Santana, PhD, University of Calgary, Calgary, AB, Canada To assist researchers and clinicians interested in using PROs in clinical practice, a working group of 10 experts developed a User’s Guide. This User’s Guide covers the goals and challenges of implementing PROs in clinical practice; determining the who, where, and when for PRO administration; identifying the methods and modes of administering PRO questionnaires; developing processes for reporting the results; providing information to support interpretation and actions in response to the PRO results; and evaluating the impact of the PRO intervention on the practice. A unifying framework was employed across all sections of the User’s Guide, with each section addressing (1) options for consideration, (2) resources required for each option, and (3) each option’s relative advantages and disadvantages. During this panel discussion, the User’s Guide authors will highlight the key considerations regarding each of the issues described above. Half the time will focus on interactive discussion, both among the panel members and between the panel and the audience. At the end of the session, attendees will have a clear understanding of the steps involved in implementing PROs in clinical practice, options for designing each step, the trade-offs among the options, and the key issues in the field requiring additional study.
The last decade has seen increasing use of generic preference-based measures such as EQ-5D and SF-6D. It is claimed that these are applicable to all patient groups, and this has support in many physical conditions. However research in mental health is more limited. This symposium describes four studies from a programme of mixed methods work to examine the validity of EQ-5D and SF-6D in mental health. The first is the European survey of mental health which investigates how these instruments describe the impact of mental health. The second is a systematic review investigating the psychometric performance of the measures across five conditions. The next session describes new psychometric analyses using seven data sets. This is followed by a report on qualitative evidence from mental health service users on how their condition impacts their quality of life. The final session brings together this evidence to address the validity of these measures in mental health. It concludes that these measures may be valid for use in mood disorders, but evidence suggests they are not suitable in psychosis. There may also be a place for a new generic measure focused on health related quality of life in people with mental health problems.
(3) Response shift detection in secondary data analysis: findings and implementation guidelines Carolyn E. Schwartz, Sc.D., DeltaQuest Foundation, Inc., Concord, MA, United States; and Tufts University School of Medicine, Boston,
123
2
Qual Life Res (2012) 21:1–132
MA, United States; Joel Finkelstein, M.Sc., M.D., FRCS(C), Division of Orthopaedics, Sunnybrook Health Sciences Center and The University of Toronto, Toronto, Ontario Canada; Nancy E. Mayo, BSc(PT), M.Sc., Ph.D., Department of Medicine, School of Physical and Occupational Therapy McGill University Division of Clinical Epidemiology Division of Geriatrics McGill University Health Center; Sara Ahmed, Ph.D., Faculty of Medicine, School of Physical Therapy, McGill University, Montreal, Quebec, Canada; Tolulope T. Sajobi, Ph.D. School of Public Health, University of Saskatchewan, Saskatoon, Saskatchewan, Canada
Hospital of Philadelphia, University of Pennsylvania, Philadelphia, PA, United States; Presenters: Katherine B. Bevans, PhD, Department of General Pediatrics, Children’s Hospital of Philadelphia and University of Pennsylvania, United States; Samantha J. Anthony PhD, MSW, RSW, McMaster University, Hamilton, Ontario, Canada; Christiane Otto, University Medical Center Hamburg-Eppendorf, Department of Child and Adolescent Psychiatry, Hamburg, Germany; Anne W. Riley, PhD, Department of Population, Family & Reproductive Health, Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD, United States
Response shift studies often draw on existing data sources and implement secondary data analysis. This symposium will discuss the theoretical and methodological challenges facing secondary analyses. Drawing from work done in spinal surgery patients, primary care patients, and post-stroke caregivers, we will discuss key findings from studies using the then-test, trajectory analysis, structural equation modeling, and relative importance methods. We will discuss the impact of missing data on response shift detection, and provide a guidelines checklist for the optimal implementation of secondary analysis.
This symposium is designed to characterize cutting edge research being conducted in the area of child and adolescent HRQOL and to report on an international Delphi process to identify the research priorities in child HRQOL. Four papers will be presented from an international group of researchers. Dr. Bevans will present on a newly developed mixed-method (qualitative-quantitative–qualitative) approach that will be illustrated using examples derived from the development and validation of 10 pediatric Patient Reported Outcome Measurement Information System (PROMIS) item banks. Dr. Anthony will report on a large qualitative study of children with cancer and child cancer survivors about their conceptualization of health. She will highlight where their concepts overlap with classic aspects of health assessed in HRQOL tools, as well as what is not captured by these instruments, including positive transformation experiences such as an enhanced sense of self, altered priorities and appreciation for life. Dr. Otto will speak about the process by which data from a number of large studies of child HRQOL in Europe were pooled to develop a new computer-adapted test (CAT) instrument, the KIDS-CAT, which has 5 domains: psychological well-being, physical well-being, autonomy and parent relations, social support and peers, and school. Finally, Dr. Riley will present the results of an international Delphi process by which research priorities were generated from members of the ISOQOL Child Health Interest Group as well as recently published authors of child HRQOL studies. Research topics were generated in round 1, rated by 69 respondents in round 2, and then refined. The top 20 research priorities will be discussed. Dr. Forrest, the discussant, will provide a brief integration and discuss next steps in stimulating new international collaborations and advancing research and clinical practice in child HRQOL.
(4) Patient-centered approach to adverse event reporting: the US National Cancer Institute’s Patient-Reported Outcomes Version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE; HHSN261201000043C and HHSN261201000063C) Sandra A. Mitchell, PhD, CRNP; Outcomes Research Branch, National Cancer Institute, Bethesda, MD, United States; Amylou C. Dueck, PhD; Mayo Clinic, Scottsdale, AZ, United States; Galina Velikova, MD, PhD; University of Leeds, St. James Institute of Oncology, Leeds, United Kingdom; Deepa Lalla, BPharm, PhD; Health Outcomes, Genentech Inc., South San Francisco, CA, United States; Ethan Basch, MD; Memorial Sloan-Kettering Cancer Center, New York, NY, United States The US National Cancer Institute’s Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) is an item bank of adverse symptom items for patient self-reporting in NCI-sponsored clinical trials. Items can be administered to patients using an online software system or an interactive voice response system. The items are intended to be complementary to the NCI’s CTCAE, a lexicon of clinician-reported adverse event items required for use in all NCI-sponsored trials. This symposium will summarize the context for development of the PRO-CTCAE and describe current and potential future uses of PRO-CTCAE within the US and internationally. Specifically, this symposium will provide an overview of the PRO-CTCAE measurement system and will describe the development and psychometric testing of PRO-CTCAE items. The symposium will also illustrate a collaboration with partners in the pharmaceutical industry who are using the PRO-CTCAE, as well as highlight an ongoing effort in the UK to integrate patient selfreporting of adverse events using PRO-CTCAE and other descriptive items into routine cancer care. Lastly, this symposium will describe the use of PRO-CTCAE in clinical research, regulatory decisionmaking, and policy in the US.
(5) Conceptual and methodological advances and research priorities for Child HRQOL Organizer and Corresponding Author: Anne W. Riley, PhD, Department of Population, Family & Reproductive Health, Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD, United States; Moderator and Discussant: Christopher B. Forrest, MD, PhD, Department of General Pediatrics, Children’s
123
Concurrent Oral Sessions Thursday, 25 October 2012 101: Societal Health Outcome Evaluation (101.1) From deindustrialization to individual health-related quality of life: multilevel evidence of contextual predictors, mediators and modulators across French regions, 2003 Etienne Audureau, Assistance Publique-Hoˆpitaux de Paris, Universite´ Paris-Descartes, France; Ste´phane Rican, Space, Health and Territories Laboratory, University of Paris Ouest-Nanterre, France; Joe¨l Coste, Assistance Publique-Hoˆpitaux de Paris, Universite´ Paris-Descartes, France AIMS Although small area effects on health-related quality of life (HRQoL) have been extensively studied, less is known at the regional level, particularly in France where no multilevel evidence is available. This work aims at confirming the existence of regional disparities in HRQoL across French regions and exploring the contextual predictors at play. METHODS This study used data from a large representative cross-sectional survey conducted in 2003 (N = 16,732), exploring individual and regional determinants of the SF-36 Physical
Qual Life Res (2012) 21:1–132 Functioning (PF) and Mental Health (MH) subscales. We considered a causal pathway leading from deindustrialization to HRQoL and assessed the roles of net migratory flows, deprivation, and the social and physical environment. Multilevel linear mixed models were fitted to account for the hierarchical clustering of individuals (level 1) nested in the following levels considered for our study: household (level 2), urban unit (level 3) and region (level 4). RESULTS After adjustment on socio-demographic individual and household characteristics, worse HRQoL results were found in regions most affected by deindustrialization (e.g. North: -0.084 Standard Deviations [PF] and -0.048 SD [MH] in men; -0.159 SD [PF] and -0.115 SD [MH] in women). Multilevel analysis showed that regional deindustrialization, assessed as the proportion of industrial employment lost since the 80’s, was significantly associated with poorer MH scores (-0.15 per % in men, p = 0.01; -0.27 in women, p \ 0.001), while non significant trends were observed for PF (-0.12 in men, p = 0.22; -0.16 in women, p = 0.13). There was evidence for mediating effects of migration, abstention rate and individual health-related behaviors. Intraregional heterogeneity and cross-level interactions were also found, indicating larger negative effects on HRQoL of widowhood and divorce with increasing regional level of alcoholrelated mortality [PF and MH in both genders] and of long-term individual unemployment with increasing regional unemployment rate [MH in men]. CONCLUSIONS Using a multilevel framework accounting for several individual and group-level characteristics, this study confirms the existence of disparities in HRQoL between French regions and provides new evidence for the complexity of place effects on individual health. Our results highlight the need for carefully conceptualized multilevel analyses to guide health policies effectively.
(101.2) Individual and household characteristics associated with adult longitudinal health status assessment: Regional Rural Injury Study-III (RRIS-III) Colleen M. Renier, BS, Essentia Institute of Rural Health; Susan G. Gerberich, Environmental Health Sciences, University of Minnesota, Minneapolis, MN; Andrew D. Ryan, Environmental Health Sciences, University of Minnesota, Minneapolis, MN; Bruce H. Alexander, Environmental Health Sciences, University of Minnesota, Minneapolis, MN; Timothy R. Church, Environmental Health Sciences, University of Minnesota, Minneapolis, MN; Patricia M. McGovern, University of Minnesota, United States; Ann S. Masten, United States; Steven J. Mongin, Environmental Health Sciences, University of Minnesota, Minneapolis, MN AIMS To evaluate the association between longitudinal Health Utilities Index-3 (HUI-3) information, fixed demographic covariates and time-varying covariates: individual demographic and health information, and aggregate household (HH) youth health information, among adults residing in agricultural HHs with youth (\20 years of age), in a five state Midwest United States (U.S.) region. METHODS The U.S. Department of Agriculture National Agricultural Statistics Service randomly sampled 32,000 agricultural operations from their Master ListFrame, in equal numbers from five states. Using computer-assisted telephone interviews, operations were screened for eligibility: agricultural HHs with youth, actively farming/ranching as of January 1st 2007. Baseline (BL) interviews, for the last 6 months of 2006, were completed by 1,459 of 3,459 eligible operations; injury interviews collected data for the two 6-month periods of 2007. HHs with youth injury (cases) were matched 3:1 with control HHs. Case and control HHs were interviewed one and 2 years following the injury interviews. Analysis included all adults (20+ years) with valid HUI-3 multi-attribute utility scores. Mixed models analysis, adjusted for within-HH and within-subject correlation, evaluated the longitudinal association
3 between HUI-3 and interview, age at BL (AGE), sex, racial/ethnic minority (RACE), spouse in HH (SPS), asthma, heart disease (HRT), diabetes, hearing problems (HEAR), vision problems, other health problems (HLTH), limitations/impairments (LIM), 6-month period injury (INJ), and number of children with—health problems (CHLTH), 6-month period injuries (CINJ) and limitations/impairments (CLIM). RESULTS A total of 1,719 HH 9 Interview Set 9 Person combinations, for 727 HHs and 1,539 adults, were included. Mean BL HUI-3 multi-attribute utility score was 0.865. Statistically significant differences (p \ 0.05) in HUI-3 were associated with AGE, RACE, SPS, asthma, HRT, HEAR, HLTH, LIM, INJ, CINJ, and CLIM. On the 0.0 = Dead to 1.0 = Perfect Health scale, a difference of 0.03 is regarded as a ‘‘minimum clinically important difference’’. Clinically important differences in HUI-3 of -0.091, -0.031, -0.035, -0.055, -0.055, -0.154 and -0.041, were associated with RACE, asthma, HRT, HEAR, HLTH, LIM and INJ, respectively. CONCLUSIONS Demographics, time-varying covariates: health conditions, limitations/impairments and injuries, among adults residing in agricultural households are associated with clinically significant decreases in HUI-3, both short and long-term.
(101.3) Using time to symptomatic recovery as an outcome measure for para-operative pain study Qiuling Shi, PhD, University of Texas M.D. Anderson Cancer Center; Xin Shelley Wang, MD MPH, UT MD Anderson Cancer Center; Elizabeth A. David, UT MD Anderson Cancer Center; Ara A. Vaporciyan, UT MD Anderson Cancer Center; Charles S. Cleeland, PhD, UT-MD Anderson Cancer Center AIMS In clinical trials assessing para-operative pain management no single efficacy measure is ideal for all purposes. We propose the time to post-operative pain recovery as one way of describing outcomes in such studies. METHODS This is a secondary analysis on data from a post-operative symptom study in lung cancer patients with thoracic surgery. Pain severity was captured with the MD Anderson Symptom Inventory Lung Cancer Module, administered pre-surgery, day 3, day 5, 1 week, then every week till month 3 post-surgery. Time of pain recovery was defined as when patient’s pain level decreased to 3 or lower (0–10 scale) after surgery. Abilities to detect the differences among sex, surgery type, and postsurgery epidural analgesic use were compared between time to recovery and responder analysis (30 % reduction from peak pain severity). RESULTS Seventy-one non-small-cell lung cancer patients were included in current analysis. All patients reported 3 or lower pain pre-surgery. The median day of pain recovery was 16 (95 % CI 8–21). Females demonstrated longer time of recovery from post-operative pain than did males (median day: 22 vs. 10, p = 0.012). Time to recovery was associated with surgery types (Open vs. VATS: HR = 2.7, 95 % CI = 1.4–5.1), epidural analgesic use (No vs. Yes: HR = 3.5, 95 % CI = 1.4–9.1), and gender (female vs. male: HR = 2.5, 95 % CI = 1.3–4.8). Responder analysis demonstrated that more female reported 30 % or greater pain reduction than did males (79 vs. 55 %, p = 0.037) at month 2. No other factors were related to responder rate at months 1, 2 or 3. CONCLUSIONS Our analysis shows that time to recovery from post-operative pain is able to detect differences among multiple factors after surgery in lung cancer patients. When compared with the commonly used responder analysis, time to pain recovery was more sensitive to reveal those differences. Considering both severity and time of pain improvement, this outcome measure provides a comprehensive understanding of post-operative pain. Whether as a primary or secondary approach in a post-surgery pain study, the display of data using time to recovery may be useful in the interpretation of results and applicability to patient care.
123
4 (101.4) Hopelessness and health-related quality of life: results from Japanese general social surveys 2010 Misa Takegami, National Cerebral and Cardiovascular Center, Japan; Yosuke Yamamoto, Kyoto university, Japan; Noriko Iwai, Osaka University of Commerce, Japan AIMS Hopelessness is associated with mortality, cardiovascular disease, and depression in Europe; however, its correlated factors remain unknown in Japan. In addition, no studies have examined whether hopelessness is associated with health-related quality of life (HRQOL). This study aims to describe the prevalence of hopelessness and its correlated factors of hopelessness in Japan. We also investigated the relationship between hopelessness and HRQOL. METHODS We used the data from the Japanese General Social Survey 2010, which consists of a nationally representative samples aged from 20 to 89 years old who lived in Japan. The data was collected using a face-to-face interview and a self-administered questionnaire. Hopelessness was measured by hopelessness scale; those with scores of 6 or above and those with scores 3–5 were considered to have high hopelessness and moderate hopelessness, respectively. HRQOL were measured by the 12-item Short-Form Health Survey. Logistic regression models were used to examine the relationship between hopelessness and its correlated factors. Analysis of covariance was used to examine the relationship between hopelessness and HRQOL. RESULTS The survey targeted 4,500 people, with 2,496 ultimately responding to the questionnaire. A total of 2,449 subjects were included in the analysis. A mean age of the respondents was 53.4 years, 46.3 % were men, and 54.7 % were free from chronic diseases. The prevalence of high and moderate hopelessness were 9.9 %, 57.6 %, respectively. Being a current smoker, having two or more chronic diseases, and low house income (less than 4.5 million yen) were associated with hopelessness. In addition, subjects with high and moderate hopelessness had lower QOL scores than subjects with no or low hopelessness. All eight subscales were a significant correlation with degree of hopelessness (test for trend, p \ 0.001). CONCLUSIONS The results show hopelessness was common, and associated with socio-economic status and co-morbid conditions in Japan. Given that both the presence and the degree of hopelessness are inversely related to HRQOL, hopelessness may be one of the social determinants of health among Japanese. Public health policy should pay attention to hopelessness at the same time as further researches are needed to investigate appropriate interventions to address hopelessness.
(101.5) Quality of life of homeless patients suffering from severe mental health disorder in France Ce´cile Fortanier, PhD, Assistance Publique Hoˆpitaux de Marseille— EA 3279, France; Vincent Girard, Assistance Publique Hoˆpitaux de Marseille—EA 3279; Aure´lie Tinland, Assistance Publique Hoˆpitaux de Marseille; Benjamin Videau, Assistance Publique Hoˆpitaux de Marseille—EA 3279; Fanny Olive, Assistance Publique Hoˆpitaux de Marseille; Guillaume Suderie, ORSMIP; Amandine Albisson, ORSMIP; Achille Djiena, F2RSM; Christian Laval, Assistance Publique Hoˆpitaux de Marseille—EA 3279; Pascal Auquier, PhD, Laboratoire de Sante Publique, France AIMS Homelessness remains a persistent public health concern. Those living with mental illness and addiction are overrepresented among this population. They are the most vulnerable; they have substantial and complex health needs. In France, the ‘‘Un chezsoi d’abord’’ research program has been conducted since 2011 August in order to reduce homelessness for the mentally ill persons through a ‘‘Housing First’’ (HF) approach. It tends to answer mainly to what extent can this costly HF approach rationalize expenditures and
123
Qual Life Res (2012) 21:1–132 conduce to best outcomes for patients in terms of health-related quality of life (HRQL) in comparison to services usually proposed. End of 24-month follow-up for 800 patients is planned on 2014 December. Aims of this presentation will be to describe preliminary results on HRQL at the inclusion time on the first included patients.; METHODS In this multi-site randomized controlled trial, 800 homeless persons presenting bipolar affective disorder or schizophrenia will be assigned to receive either experimental HF Intervention or Housing/Services as Usual in 4 cities. HF Intervention offers permanent independent housing associated to a consumerdriven case-management (Assertive Community Treatment). Assessments are carried out at baseline and each 6 months during a 24-month period. Data collection is about socio-demographic, use of health services, social functioning and community integration, housing stability, cost-utility ratios as well as health indicators including HRQL which will be assessed through both generic and specific instruments: the SF36 and the 18-item S-QOL dedicated to people with schizophrenia.; RESULTS To date, 167 patients are included and inclusion data are completed for all of them. 117 are suffering from schizophrenia and 50 from bipolar disorder; there are 140 male and 27 female. Median age is 39 years (19–70). Extensive results will be presented on patients who will be included until 2012 August (approximately n = 200) for a data analysis planned on September. HRQL level for this very specific population will be described according to detailed characteristics of patients and could be compared to norms for the used scales. CONCLUSIONS Results will be informative on the HRQL level of homeless persons in presence of a severe mental disorder in France before benefiting any specific intervention.
102: Patient Reported Outcomes in Cancer Trials (102.1) Value added of patient-reported data to clinician reported toxicity data: preliminary analysis of 3DCRT vs IMRT on RTOG 0126 prostate cancer trial high dose arm Deborah W. Bruner, RN, PhD, FAAN, Emory University, United States; Daniel Hunt, PhD, American College of Radiology; Jeff M. Michalski, M.D., Washington University in St. Louis; Walter Bosch, Ph.D, Washington University in St. Louis; Yan Yan, M.D., M.S., American College of Radiology; James M. Galvin, D.Sc., Thomas Jefferson University-Jefferson Medical College; Jean-Paul Bahary, M.D., FRCP, Centre Hospitalier de L’Universite de Montreal, Canada; Gerard C. Morton, M.D., University of Toronto-Odette Cancer Centre; Matthew B. Parliament, M.D., University of Alberta; Howard Sandler, M.D., Cedars-Sinai Medical Center AIMS Assess value added of patient-reported outcome (PRO) data to clinician reported toxicity in IMRT versus 3DCRT for prostate cancer (PC). METHODS 763 patients were randomized to the 79.2 Gy arm of a trial comparing high to standard dose RT for localized PC. Patient treatment was either IMRT (prostate plus proximal seminal vesicles (PSVs)) or 3DCRT that allowed field reduction to spare PSVs. Toxicity was assessed with CTCv2.0. PRO bowel/bladder function were measured with Functional Alterations due to Changes in Elimination (FACE) and erectile function (EF) with the International Index of Erectile Function (IIEF). Changes over time from 0 (baseline), to 3, 6, 12, and 24 months were evaluated. Tests included: Wilcoxon for modality comparisons, Chi-square to compare pretreatment characteristics between patients with/without PRO/toxicity data, and multivariate logistic regression analysis (MVA) for predictors of PRO/toxicity outcomes. RESULTS Toxicity analysis indicated a benefit in favor of IMRT for late grade 2+ GI (22 % 3DCRT vs 15 % IMRT, p = 0.04) and for combined acute GI/GU grade 2+ toxicities (15 % 3DCRT vs 910 % IMRT, p = 0.042). 499–507 patients completed baseline FACE and IIEF, respectively.
Qual Life Res (2012) 21:1–132 Completion rates were 66 % FACE and 61 % IIEF at 24 mos. There were no statistically significant differences between 3DCRT and IMRT for total FACE score, urinary/bowel subscales, or total IIEF at any time point. EF and satisfaction domains showed significance in change scores between RT modalities at 24 months (both p = 0.04), with less decrease in both domains on IMRT. On MVA including RT method, age, race, time to assessment and penile bulb dose, only age (\70 years) and time of assessment (baseline) were significantly better. For age \70 years, at 24 months RT method was statistically significant on MVA when assessing associations with the EF subscale (odds ratio 4.3; 95 % CI 1.3, 14.1; p = 0.02). CONCLUSIONS Clinician reporting showed better grade 2+ GI and combined GI/GU outcomes for IMRT vs 3DCRT but PROs showed no corresponding difference. A difference that PROs demonstrated in favor of IMRT not found on clinician reporting was EF; a benefit primarily in men age \70 years. This data adds to the complexity of comparative effectiveness discussions.
(102.2) A systematic review in health-related quality of life reporting in small-cell lung cancer clinical trials: an examination into the added value of health-related quality of life findings Efstathios Zikos, MSc, MA, EORTC HQ, Belgium; Corneel Coens, MSc, EORTC HQ, Belgium; Irina Ghislain, MSc, EORTC, Quality of Life Department; Divine Ewane Ediebah, MSc., EORTC HQ, Belgium; Elisabeth Sloan, EORTC, HQ; Chantal Quinten, MSc, EORTC, Belgium; Michael Koller, PhD, Center for Clinical Studies, Germany; Jan P. Van Meerbeeck, PhD, Ghent University Hospital, Respiratory Medicine-Thoracic Oncology; Henning H. Flechtner, PhD, Otto-von-Guericke University (KKJP), Germany; Roger Stupp, MD, 7Department of Oncology of the Riveria/Chablais, Switzerland AIMS Small cell lung cancer (SCLC) represents approximately 15 % of all lung cancer and commonly presents at an advanced stage. Thus treatment is often palliative, and maintenance quality of life a major objective. Increasingly, Health-Related Quality of Life (HRQOL) is evaluated in randomized controlled trials (RCTs). This study examined the HRQOL methodology reporting SCLC RCTs and evaluates the added clinical value of HRQOL data. METHODS A systematic Medline search was conducted. Eligible interventional (e.g. chemotherapy, radiotherapy, surgery) RCTs reported in English between April 1991–February 2011 were included if they included HRQOL assessments. Only RCTs with a sample size = 100 were included. Two independent reviewers evaluated all selected RCTs. RESULTS Twenty-nine RCTs, involving 10,493 patients, were eligible for inclusion in our review. HRQOL was defined as a primary outcome in only 3 % (1 trial with 243 patients) of all RCTs. Baseline HRQOL assessment was mandatory in 14 % of RCTs. Overall survival (OS) differences between arms was found in 45 % of the RCTs and HRQOL differences were found in 66 %. Among the RCTs reporting a significant OS difference, 31 % found no difference in HRQOL; in 31 % better OS was associated with better HRQOL; in 15 % better OS was seen with worse HRQOL; and in 23 % worse OS was associated with worse HRQOL. The treatment combinations that led to better OS and improvements in HRQOL were: (a) irinotecan and carboplatin versus (vs.) oral etoposide plus carboplatin, (b) paclitaxel, carboplatin, epotoside and phosphate vs. carboplatin, etoposide phosphate and vincristine, (c) ifosfamine, carboplatin, etoposide and vincristine vs. standard chemotherapy, and (d) etoposide, cisplatin, cyclophosphamide and 40 -epidoxorubicin vs. etoposide plus cisplatin. HRQOL differences between arms often concerned fatigue, sleep problems, pain, appetite loss, general well-being, diarrhea, hair loss, and physical and emotional functioning. CONCLUSIONS HRQOL assessment in SCLC RCTs provides added value to the studies even when no OS difference is seen. Importantly HRQOL data appears
5 to provide valuable information for these treatments where better HRQOL was associated with overall survival benefit.
(102.3) Patient-reported satisfaction with reconstructed breasts in the long-term survivorship period: comparison of autologous and non-autologous breast reconstruction Amie M. Scott, MPH, Memorial Sloan-Kettering Cancer Center; Adelyn L. Ho, MD MPH, University of British Columbia, Canada; Anne Klassen, D Phil, McMaster University, Canada; Stefan Cano, BSc PhD, Peninsula College of Medicine & Dentristry, United Kingdom; Nancy VanLaeken, The University of British Columbia; Peter Cordeiro, Memorial Sloan Kettering Cancer Center; Andrea Pusic, MD MHS FRCSC, Memorial Sloan Kettering Cancer Ctr AIMS Breast cancer patients undergoing mastectomy may choose to have reconstruction performed using either their own tissue or an implant. As many patients are candidates for both, valid and reliable patient-centered outcomes data are crucial to shared medical decisionmaking. The objective of this study is to determine whether patientreported satisfaction with their reconstructed breasts is dependent on type of reconstructive surgery and length of time from reconstruction. METHODS Participants were recruited from Memorial Sloan-Kettering Cancer, NY and the University of British Columbia, Canada. Patients completed the BREAST-Q, a new patient-reported outcome measure for breast surgery patients. The dependent variable was the BREAST-Q Satisfaction with Breast score, a 16-item scale scored from 0 to 100. Procedure type, time since surgery, and patient characteristics were independent variables. Univariate analysis and clinical judgment were used to identify variables included in the model, and multivariate linear regression models were constructed to control for confounders. RESULTS The study sample consisted of 510 women (response rate 62 %). The sample was on average aged 54.3 ± 9.3 (range 21–81), surveyed 3.71 years ± 1.55 (range 1–9) after surgery, 66 % were reconstructed using an implant. Type of surgery and laterality were found to be variables that predicted higher patient satisfaction with their breasts after controlling for radiation therapy, follow-up time, timing of surgery, age, body mass index, and major complications (surgery type p \ 0.001; laterality p \ 0.001, R-square = 0.17). CONCLUSIONS As there is a growing population of breast cancer survivors, understanding how a woman’s satisfaction with her reconstructed breasts changes over time is essential. This study suggests that patient satisfaction with breast reconstruction depends on the type of reconstruction a woman undergoes. This patient-centered outcome data can be used to enhance shared medical decision-making by providing patients with information about realistic expectations for satisfaction with breasts related to type of surgery chosen.
(102.4) Health-related quality of life following breast reconstruction after mastectomy: a large multi-centre cohort study of patient-reported outcomes for different surgical techniques John Browne, PhD, University College Cork, Ireland; Ranjeet Jeevan, Royal College of Surgeons of England; David Cromwell, London School of Hygiene and Tropical Medicine, United Kingdom; Jerome Pereira, James Paget University Hospitals NHS Foundation Trust; Chris Caddy, Sheffield Teaching Hospitals NHS Foundation Trust; Carmel Sheppard, Portsmouth Hospitals NHS Trust; Kimberley Greenaway, NHS Information Centre for Health and Social Care; Jan Van Der Meulen, London School of Hygiene and Tropical Medicine AIMS There are many surgical techniques for breast reconstruction following mastectomy and considerable uncertainty over the
123
6 comparative effectiveness of these options. The UK National Mastectomy and Breast Reconstruction Audit is a prospective cohort study which provides unique levels of statistical power and generalisability for an analysis of the relationship between operative technique and patient-reported outcomes. METHODS The study recruited women undergoing surgery between January 2008 and March 2009 at 147 NHS acute trusts and 79 independent sector hospitals. 2,956 women were sent the Breast-Q outcome measure 18 months after their surgery and responses were received from 2,384 (81.0 per cent). Outcome scores on four scales (aesthetic appearance, emotional well-being, sexual wellbeing and physical well-being) were adjusted for women’s age, deprivation index, performance status, smoking status, and whether they had undergone radiotherapy and/or chemotherapy postoperatively, when comparing the results of different surgical techniques. Rasch scaling was used to translate numerical effect sizes to verbal benchmarks. RESULTS Women undergoing pedicle and free flap reconstruction surgery reported higher outcome scores on the aesthetic, emotional well-being and sexual well-being scales than women undergoing less complex, implant-only reconstruction. This effect was observed in both the immediate and delayed reconstructions settings. Patients undergoing delayed reconstruction procedures reported better outcomes on the aesthetic, emotional well-being and sexual well-being scales than patients undergoing immediate reconstruction when matched according to operative technique. CONCLUSIONS Reconstruction procedures using flap transfer techniques were associated with better patient-reported outcomes. The use of Rasch scaling with BreastQ modules allowed the clinical relevance of the findings to be demonstrated and this technique should be used to augment the information provided to women about what they can expect after different types of reconstructive surgery. The differences observed between immediate and delayed reconstruction should be interpreted with caution. Women undergoing a delayed breast reconstruction have lived without a breast/ breast mound for a period of time, and consequently may experience a response shift which increases their appreciation of the aesthetic improvement associated with reconstruction surgery.
103: Utility Measures (103.1) Revaluing the SF-6D using ordinal methods for eliciting preferences Richard Norman, MSc, University of Technology, Sydney, Australia; John E. Brazier, PhD, Univ of Sheffield, United Kingdom; Rosalie Viney, PhD, Centre for Health Economics Research & Evaluation, Australia; Leonie Burgess, University of Technology, Sydney, Australia; Paula Cronin, University of Technology, Sydney, Australia; Madeleine T. King, PhD, Psycho-oncology Cooperative Research Group (PoCoG), Australia; Julie Ratcliffe, University of Flinders; Deborah Street, University of Technology, Sydney, Australia AIMS SF-6D is a preference-based index derived from the SF-36 and has become one of the most widely used generic preference-based measures of health. Preference weights have been obtained from members of the general population in the UK and 6 other countries using standard gamble, but this task is complex and the values can be contaminated by attitudes to risk. In this paper we examine the use of an ordinal task based on the discrete choice experiment to derive a new index from the SF-6D. METHODS A DCE to obtain values for SF-6D health state was implemented in an Australia-representative online panel (n = 1,017). Respondents were asked to compare alternatives comprised of the six dimensions of the SF-6D states and one of seven possible levels of survival (1–20 years). A number of specifications were estimated involving a time main effects term for the full health state at different durations and interaction terms between the ill health state levels and duration using a random-effects
123
Qual Life Res (2012) 21:1–132 probit model. Models were compared using AIC and BIC. RESULTS The results reflect the broadly monotonic design of the SF-6D: coefficients are negative and increase in size with lower levels of each dimension (e.g. mental health values the decrements were 0.06, 0.07, 0.18, 0.26). Resultant coefficients were then transformed onto the QALY scale where full health is one and dead is zero using a ratio of Marginal Utilities approach. The predicted mean health state values range was from -0.35 for the poorest state up to 1.0 for the best state with 12 % of the 18,000 states less than zero. This contrasts with the original UK standard gamble value range of 0.3–1.0, with none less than zero. CONCLUSIONS DCE offers a promising alternative way to value health states compared to conventional cardinal techniques like standard gamble or time trade-off. More work is required to extend this to other countries and to examine the performance of the new DCE based index to the original DCE one in terms of sensitivity and responsiveness.
(103.2) Mapping the Childhood Health Assessment Questionnaire to the Health Utilities Index Mark 2 and Mark 3 in children with juvenile idiopathic arthritis Kit Simpson, Medical University of South Carolina; James W. Shaw, Abbott Laboratories AIMS Cost-utility analysis may be used to evaluate treatments for juvenile idiopathic arthritis (JIA) since the condition has a significant impact on patients’ quality of life. However, health state utilities are infrequently collected in JIA clinical trials, and the published utilities for children are often inadequate for use in Markov models of JIA treatments. A study was conducted to develop and evaluate algorithms for predicting Health Utilities Index (HUI) Mark 2 (HUI2) and Mark 3 (HUI3) multi-attribute utilities from responses to the Child Health Assessment Questionnaire (CHAQ). METHODS CHAQ items were selected and matched by response category to HUI items measuring similar attributes. Since no CHAQ items corresponded to the HUI vision and pain attributes, suitable items administered in a clinical trial of adalimumab for the treatment of JIA (DE038) were selected for inclusion in the mapping algorithms. Items measuring hearing, speech, cognition, and emotion were not administered in the clinical trial. The utilities for these attributes were fixed at 1.0. The mapping algorithms were applied to data collected in DE038 to predict HUI2 and HUI3 utilities for participants’ own current health. Subsequently, mean utilities were compared among five health states developed for use in a Table 1
Mean predicted utilities and other health outcomes by JIA health state
Measure
Remission with no limitations
Remission with limitations
Active mild
Active Active Range moderate severe
Predicted HUI2 utility
0.98
0.96
0.94
0.79
0.56
0.18–1.00
Predicted HUI3 utility
0.97
0.95
0.94
0.76
0.51
0.04–1.00
CHAQ score
0.03
0.08
0.03
0.42
1.37
Pain score
4.3
7.6
12.4
34.3
58.5
0–100
Global DA score
4.1
8.0
12.6
35.3
59.2
0–100
No. LOM joints
0.03
4.0
4.9
9.3
13.9
0–66
No. active joints
0
0
8
20
32.2
0–168
Weighted joint score
0.01
0.02
6.1
11.7
15.5
0–47
No. patient-years
249
224
499
256
0–2.9
47
Means were estimated over the number of person-years spent in each health state in the clinical trial CHAQ Childhood Health Assessment Questionnaire, DA disease activity, global LOM limitation on passive motion Source: Prince FHM et al. Annals of Rheumatic Diseases. 2010; 69:138–142
Qual Life Res (2012) 21:1–132 Markov model. RESULTS Mean predicted HUI2 and HU3 utilities varied among the five health states with lower values being observed for poorer states of health (Table 1). The mean predicted HUI2 and HUI3 utilities for patients in remission were 0.98 and 0.97, respectively, while mean predicted HUI2 and HUI3 utilities for patients having active severe disease were 0.56 and 0.51, respectively. The range of predicted HUI3 utilities exceeded the range of predicted HUI2 utilities; however, the two distributions were qualitatively similar. CONCLUSIONS The mapping algorithms developed in this study predict utilities that discriminate among Markov model health states and appear to have construct validity when compared to other subjective and objective health measures. The HUI3 mapping algorithm yields predictions that are similar in magnitude to HUI3 utilities observed in a study of the quality of life of children with refractory JIA (Prince et al. 2010). The two algorithms may be used to generate preferences to inform future cost-utility analyses of JIA treatments. Reference Prince FHM et al. Annals of Rheumatic Diseases. 2010; 69: 138–142.
(103.3) Mapping responses on the functional assessment of cancer therapy-general (FACT-G) to the HUI2 and HUI3 Paulos Teckle, PhD, Canadian Centre for Applied Research in Cancer, Canada; Peacock Stuart, BC Cancer Agency; Kim Van Der Hoek, BC Cancer Agency AIMS To help facilitate economic evaluations of interventions for treating cancer, we estimated utility indices for the frequently used cancer-specific instrument of quality of life, by mapping it onto each of the HUI-2 and HUI-3 general health preference-based indices. METHODS A sample of 367 cancer patients from the Vancouver Cancer Centre completed the FACT-G, HUI-2 and HUI-3 healthrelated quality of life questionnaires. Models of the relationships between the FACT-G and each of the preference based indices were estimated using regression analyses. We examined three alternative modeling approaches: ordinary-least-squares (OLS); generalized-linear-modeling (GLM) using a Gaussian distribution and log link; and censored-least-absolute deviations (CLAD). The performance of the models was assessed in terms of how well the responses to the cancerspecific instrument predicted utilities from each of the preferencebased instruments using the mean absolute error (MAE) and root mean square error (RMSE). We adjusted for the socio-demographic and clinical characteristics of patients. RESULTS Results from the final models of the three approaches did not differ significantly. Physical, functional and emotional subscales of the FACT-G were significant predictors of the HUI-2 and HUI-3 utility scores. The root mean square error for the HUI-3 was lower, suggesting better predictions than for the HUI-2. The OLS/GLM predicted HUI-2 and HUI-3 scores matched the observed values more closely than the CLAD. CONCLUSIONS There is potential to estimate both HUI-2/ HUI-3 utilities using responses from the FACT-G cancer-specific measure of quality-of-life, even though it was not designed as a utility instrument. Our results suggest that it is possible to estimate QualityAdjusted-Life-Years from studies where only cancer-specific instruments have been administered.
(103.4) EQ-5D and SF-6D performance in different patient groups Pedro L. Ferreira, PhD, University of Coimbra, Portugal; Lara N. Ferreira, PhD, University of the Algarve-ESGHT, Portugal; Luis N. Pereira, University of the Algarve-ESGHT; CIEO
7 AIMS Over the last years the use of preference-based instruments has been increasing. The choice of the instrument may lead to different utility values and hence different results in cost-utility analysis. This has heightened the importance of studying the performance of the instruments across patient groups. Previous studies have focused their attention on a specific diseases and therefore with specific characteristics. As far as we are aware, few have analyzed patient groups simultaneously as a whole group suffering from a chronic disease. This research intends to explore the performance of the SF-6D and the EQ-5D in patients suffering from asthma, cataracts, chronic obstructive pulmonary disease (COPD), and rheumatoid arthritis (RA). METHODS A sample of 643 patients (asthma-115, cataracts352, COPD-72, RA-104) completed both the SF-36 v2 and the EQ-5D. The SF-6D and the EQ-5D indexes were computed using the UK value sets. The utilities were analyzed globally, by disease group and by sociodemographic group to explore the discriminative ability of the instruments. Ceiling and floor effects were studied. The association and level of agreement between the indexes and by similar dimension were accessed using correlation coefficients. Pairedsamples t-tests were used to identify differences between the scores. Differences of 0.03 were considered as minimally important differences. Regression models were used to explore the nature of relationship between the SF-6D and the EQ-5D. Furthermore, the discriminative properties of the instruments were evaluated and compared using effect sizes and ROC curves. RESULTS In the overall sample both indexes had the same mean (0.72). However the mean EQ-5D was 0.05 higher than the mean SF-6D in the individuals suffering from asthma and COPD. There was a strong correlation between both indexes (0.68) and there were similar correlation coefficients by disease. Agreement between the instruments was as expected higher in similar dimensions. Both instruments were able to discriminate between sociodemographic groups. The EQ-5D presented a large ceiling effect (23 %), even considering that all individuals suffered from a specific chronic condition. The regression models showed a non uniform relationship between the indexes. This result was confirmed by hypothesis tests that compare the regression coefficients estimates for each disease with the estimates of the overall model. The area under the ROC curves showed that the SF-6D is more efficient in detecting differences between health state groups in almost all cases. CONCLUSIONS This research confirmed the hypotheses that the SF-6D generates higher utility values in less healthy individuals whilst the EQ-5D gives higher utility values in healthier individuals. According to our results the SF-6D and the EQ-5D seem to perform differently in each of the diseases studied. The regression models showed that the coefficients estimates for each disease group differed from the overall model. These results do not allow seeking for an overall adjustment between both measures independently from the health state of the individual. Results show that the EQ-5D and the SF-6D generate different utility values, and the reason for these differences should be further investigated.
104: Quality of Life Evaluation in Cancer and Chronic Illness (104.1) Defining chronic cancer: patient experiences and self-management needs Simon A. Pini, The University of Leeds, United Kingdom; Clare Harley, PhD, University of Leeds, United Kingdom; Kiera Bartlett, University of Sheffield, United Kingdom; Galina Velikova, MD, University of Leeds, United Kingdom AIMS In comparison to survivorship and advanced cancer, research focusing on patients living with cancer as a chronic condition is scarce. Chronic cancer is therefore poorly defined and strategies for supporting patients during this disease phase are lacking. The current research project aims to define chronic cancer, explores patient
123
8 experiences, and reviews patients’ support needs against those described in the 2007 Department of Health Generic Choice Model for Long-term Conditions (DoH-GCM). METHODS Patients were considered ‘chronic’ in this research if; they had a diagnosis of active advanced or metastatic cancer that could not be cured; they had active anti-cancer treatments available that could lead to symptom control, slow disease progression, or prolong life; and were not considered to be at the end-stage of cancer. Semi-structured interviews with chronic patients were audio-recorded, transcribed, and the resulting data explored for emergent themes. A priori themes from the DoH-GCM were applied and these were: clinical support; self-care and self-management; supporting independence; psychological support; and social and economic factors. RESULTS 56 patients were recruited from five clinical groups at a cancer centre in Yorkshire, England: breast (n = 11); renal (n = 11); colorectal/gastrointestinal (n = 12); gynaecological (n = 12); prostate (n = 10). Most patients aspired to living normal lives. Challenges included frequent and lengthy hospital appointments, long-term symptom control, and uncertainty. Only renal and prostate patients reported having routine access to specialist nursing. Uptake of support services was varied and there was generally poor understanding of support pathways for non-medical problems and issues occurring when patients were not receiving active treatment. There was variation in coping strategies and ability of patients to attain a positive outlook on life. CONCLUSIONS For patients to do well in this cancer phase requires good self-management of symptoms plus taking an active role in accessing appropriate services as needed. Care planning at the point of transition to the chronic phase of cancer should focus on evaluating patients’ needs, clarifying support pathways, increasing the profile and involvement of community services and organizations, and supporting patients and families develop effective selfmanagement skills.
(104.2) Patients with chronic diseases quality of life, moral expectations and consensual platform Estelle Boetti, MA, Public Health Nord University Hospital, France; Auquier Pascal, MD PhD, Laboratoire de Sante Publique, France; Le Coz Pierre, Medical Philosophy, La Timone University Hospital of Marseille, France AIMS The number of researches on QOL is increasing. However, the QOL concept is characterized by lack of agreement and showed conceptual problems. The objective of our study is to understand how philosophy can help to renew our understanding of the QOL concept. METHODS Analytical process started with a literature review in Philosophy (Antiquity to contemporary) and Public Health (in Medline-1990–2010). Conceptualization and measurement theories of well-being and satisfaction have been also studied in order to reveal the underlying theories. Through these concepts we have isolated patient’s moral expectations. These moral expectations are formalized as values corresponding to the four ethical principles developed by Beauchamp and Childress, in The principles of biomedical ethics. These principles constitute a recognized consensual platform, based on the four international principles: autonomy, beneficence, non-maleficence and justice. RESULTS The Public Health approaches of QOL don’t focus enough on the moral expectations. To assess the applicability of the principles in the field of health, we confront 31—QOL’s instruments to ethical principles, in order to determine how the moral expectations of children are explored. The comparison shows that all the instruments don’t tackle the four fundamental moral expectations. Nevertheless, a change in perspective has taken place during these four decades: questionnaires made since the late 1990s put more emphasis on the four ethical principles. CONCLUSIONS In examining the ethical principles
123
Qual Life Res (2012) 21:1–132 underlying construction or contents of the QOL instruments, we illustrate that a philosophical approach provides health professionals with a different appraisal of QOL concept. Integrate international ethical principles in assessing QOL opens up new perspectives in patient’s care.
(104.3) Measure-ABLE: ensuring that health outcome and performance measures are accessible and valid for people with disabilities and chronic medical conditions Susan Magasi, Northwestern University, United States; Mark Harniss, University of Washington AIMS People with disabilities and chronic medical conditions represent a large and growing segment of the population. These groups tend to be heavy users of healthcare and social services. Accurate measurement of service needs and outcomes is vital to ensure that services are effective, meaningfully targeted and appropriately allocated. Technological advances are leading to the development of increasingly sophisticated computer-based assessments. For some people with physical, cognitive and sensory impairments, the human–computer interaction poses accessibility challenges and may limit the interpretability of test scores. This presentation will (1) provide a framework for integrating accessibility considerations into the instrument development process and (2) make end-users aware of how accessibility issues can impact the interpretability and validity of test scores.; METHODS Over the course of 3 studies, we conducted expert reviews of all NIH Toolbox measures and of the PROMIS user interface using the standards of Section 508 of the Rehabilitation Act. We also conducted usability and accessibility testing of the NIH Toolbox (n = 30 people with traumatic brain injuries, spinal cord injuries and stroke) and the PROMIS user interface (n = 14 people with vision, reading and mobility impairments).; RESULTS Using The Standards for Educational and Psychology Testing’s Strategies for Test Modification as an organizing framework, we highlight accessibility barriers to measurement identified in our research. Specifically, presentation formats that rely on a single sensory modality, such as vision, limit people’s abilities to access instructions, test items and stimuli. Response formats and scoring algorithms that integrate motor response time; and lengthy fast paced assessment batteries may selectively disadvantage users with sensory, motor, and cognitive impairments. Inappropriate use of reasonable accommodations, such as administrator reading of instrument content or entering participant answers, may threaten the validity of some measures.; CONCLUSIONS Developers and end-users must be aware of how human– computer interactions influence participant performance and the interpretability of findings. To ensure that outcome and performancebased measures balance accessibility and validity considerations, we advocate a 3-phase accessibility review process that includes (1) consultation between content, information technology and accessibility experts, (2) expert review, and (3) usability testing with people with disabilities and chronic medical conditions.
(104.4) Quality of life (QoL) and symptoms in patients with myelodysplastic syndromes (MDS). Investigating the role of transfusion dependency in an International study Fabio Efficace, PhD, Italian Group for Adult Hematologic Diseases, GIMEMA, Italy; Gianluca Gaidano, Amedeo Avogadro University of Eastern Piedmont, Novara, Italy; Reinhard Stauder, Innsbruck University Hospital, Innsbruck, Austria.; Giovanni Caocci, University of Cagliari, Cagliari, Italy; Maria Teresa Voso, University of Rome ‘‘Cattolica S. Cuore’’, Rome, Italy., Italy; David Bowen, St James’s
Qual Life Res (2012) 21:1–132 University Hospital. Leeds, UK.; Dominik Selleslag, AZ Saint-Jan AV, Department of Hematology. Brugge, Belgium.; Anna Angela Di Tucci, Hospital ‘‘Businco’’, Department of Hematology. Cagliari, Italy; Monica Lunghi, Amedeo Avogadro University of Eastern Piedmont, Italy; Massimo Breccia, Sapienza University, Rome, Italy AIMS The main objective was to investigate whether transfusiondependency, at the time of diagnosis, was associated with different QoL profiles in patients with high-risk MDS. QoL in such patients is critical for a number of reasons including: associated comorbidity, symptom burden, and very limited life expectancy. METHODS Data were gathered through an ongoing international prospective observational study involving 15 countries and 50 centers that recruits newly diagnosed patients with MDS. QoL was assessed with the EORTC QLQ-C30 questionnaire. Associations between transfusion dependence and QoL, socio-demographic characteristics (i.e., family status, age, gender, education) and clinical data (i.e., performance status, disease severity) were investigated using Fisher’s exact test and Wilcoxon-Mann–Whitney test as appropriate. Transfusion-dependency was defined according to international guidelines (i.e. having received had at least one Red Blood Cell transfusion every 8 weeks over a period of 4 months). Statistical comparisons were adjusted for multiple testing. RESULTS Overall, 258 patients (33 % female and 67 % male) participated. Mean age of patients was 69 years (range 31–89) and 33 % were transfusion dependent. No major differences were found in socio-demographic characteristics between transfusion dependent versus transfusion independent patients. Transfusion dependent patients reported an overall worse QoL profile. In particular, the most impaired functional scales were Role and Physical functioning. Differences in these two scales were clinically significant (means scores differences of 14.5 and 13.3 respectively), while only physical functioning remained statistically significant after multiple testing adjustment (p = 0.0002). Transfusion-dependent patients also reported clinically meaningful higher levels of fatigue (mean score difference of 10.7). CONCLUSIONS Transfusion-dependent MDS patients reported an overall worse QoL and symptom profile. Major impairments were noted in physical aspects. This data might help clinicians in better identifying specific QoL concerns associated with transfusion-dependency.
105: Patient and Health Outcomes in Chronic Conditions (105.1) Health-related quality of life does not differ between short-, long- and very long-term cancer survivors in the Swedish general population Anna Wikman, Karolinska Institutet, Sweden; Therese Djarv, MD PhD, Karolinska Institutet, Sweden; Mohammed A. Johar, B.Sc.(Hons), M.Sc., Karolinska Institutet, Sweden; Pernilla Lagergren, PhD RN, Karolinska Institute, Dept. of Molecular Med, Sweden AIMS Previous studies have found that symptoms of fatigue, pain and sleep problems persist in very long-term ([10 years since diagnosis) cancer survivors, whereas longer time since cancer diagnosis is associated with decreased impairments in cognitive difficulties and social concerns. In population-based studies on the effects of cancer on health-related quality of life (HRQL) the influence of time since cancer diagnosis is rarely accounted for. Therefore, the aim of this study was to assess the relationship between time since cancer diagnosis and impairments in HRQL in the Swedish general population. METHODS A cross-sectional population-based survey of 4,910 Swedish adults aged 40–79 years was conducted between April and June 2008. Three hundred and nineteen cases of cancer were identified in this sample, whereas 4,591 participants reported no
9 cancer. Analysis of co-variance (95 % Confidence Intervals), adjusting for age, gender, education, marital status, and non-cancer co-morbidity, was performed to compare HRQL ratings, as measured by the EORTC QLQ-C30, between short- (\5 years since diagnosis), long- (5–10 years since diagnosis) and very long-term ([10 years since diagnosis) cancer survivors, and with that of the no cancer comparison group. RESULTS No clinically or statistically significant differences were observed between the survivor groups across the HRQL scales assessed, with the exception of symptoms of fatigue which were greater among the short- and long-term survivors compared with the very long-term survivors. For the majority of the HRQL scales, ratings were clinically and statistically significantly poorer in the group of cancer survivors \5 years since diagnosis compared with the no cancer group. Further, impairments were also observed in the long-term survivors on global quality of life, physical function and fatigue, compared with the no cancer group. HRQL of very long-term survivors did not differ from the group reporting no cancer. CONCLUSIONS Beyond 5 years of cancer diagnosis, HRQL was mostly comparable to the general population without a cancer history. Although HRQL was similar between short-, long- and very long-term survivors, findings suggest the assessment of symptoms of fatigue may be warranted even 10 years post cancer diagnosis.
(105.2) Health-related quality of life in persons treated for active and latent tuberculosis, versus persons screened but not treated for tuberculosis from diagnosis through 6 months of follow-up Melissa Bauer, McGill University, Montreal Chest Institute, Canada; Christina Greenaway, Sir Mortimer B. Davis—Jewish General Hospital; Marek Lalli, Montreal Chest Institute, Canada; Allison Leavens, Montreal Chest Institute, Canada; Ashley Waynne, Montreal Chest Institute, Canada; Bilkis Vissandjee, Faculty of Nursing, Universite´ de Montre´al, Canada; Andrea Benedetti, Montreal Chest Institute, Canada; Richard Menzies, Montreal Chest Institute, Canada; Kevin Schwartzman, Montreal Chest Institute, Canada AIMS Active tuberculosis (TB) can impose substantial morbidity, while treatment for latent TB infection (LTBI) has frequent side effects. Our aim is to compare health utilities and quality of life reported by patients treated for active TB or LTBI with others screened, but not treated for TB. METHODS Subjects were recruited at two hospitals in Montreal, Canada (2008–2011), and completed SF-36 and Standard Gamble questionnaires at initial and follow-up visits: 1, 2, 4, 6, 9, 12 months. SF-6D utility scores are derived from SF-36 responses. Linear mixed models were used to compare changes in health utilities reported by those treated for active TB disease and LTBI infection to those of TB-free, healthy controls within 2 weeks of diagnosis through 2 months of longitudinal follow-up. Results through the 6-month follow-up visit will be included in final presentation. RESULTS The mean age of the three subject groups varied from 35 to 39 years, women represented about half of each group, and nearly all of the subjects were foreign-born. Average Standard Gamble scores of the subjects treated for active TB disease improved significantly through the first month of treatment; this change was +0.11 (95 % CI: +0.03, +0.19). No statistically significant changes in mean Standard Gamble scores were detected for the subjects treated for LTBI. Although mean SF-6D values showed improvement in HRQoL from baseline through 2 months of treatment among those treated for active TB, these estimates were not statistically significant. Subjects treated for LTBI reported similar SF-6D scores to the healthy controls. CONCLUSIONS Our results show a negative impact of active TB disease on HRQoL at diagnosis, with subjects’ HRQoL improving during the first month of active TB treatment. These findings will be incorporated into cost-utility assessments of TB control interventions.
123
10
Qual Life Res (2012) 21:1–132
Table 1 Demographic characteristics at initial interview (N = 301) At initial interview
Active TB Latent TB Controls (N = 61) (N = 121) (N = 119)
Median age, years (IQR)
39 (13.8)
Females, N (%)
27 (44)
Foreign-born, N (%) 57 (93) Median years in 2.5 (0.4, 8.6) Canada (IQR)
35 (11.3) 66 (56)
37 (11.0) 64 (54)
115 (95) 3.0 (0.6, 9.7)
105 (88) 2.3 (0.9, 8.7)
Table 2 Regression analysis for changes in mean SF-6D and Standard Gamble scores of subjects treated for TB compared to controls from baseline through 2 months of treatment Health utility metric
Baseline to 1 month of treatment
1–2 months of treatment
Active TB
+0.11 (+0.03, +0.19)
-0.01 (-0.09, +0.08)
LTBI
-0.02 (-0.08, +0.05)
0.00 (-0.07, +0.08)
-0.03 (-0.07, +0.01)
-0.02 (-0.06, +0.02)
Active TB
+0.06 (0.00, +0.11)
+0.05 (-0.01, +0.11)
LTBI
-0.01 (-0.06, +0.04)
0.00 (-0.05, +0.05)
Controls
-0.01 (-0.04, +0.02)
0.00 (-0.03, +0.04)
Standard gamble
Controls SF-6D
a
b
a
Model comparing active TB subjects to controls was adjusted for age at baseline, sex, and number of individuals residing with the subject at baseline. Model comparing LTBI subjects to controls was adjusted for age at baseline, sex, and number of hours of work or school missed due to illness b Model comparing active TB subjects to controls was adjusted for age at baseline and sex. Model comparing LTBI subjects to controls was adjusted for age at baseline, sex, and number of hours of work or school missed due to illness (105.3) Risk factors for weight loss and influence of nutritional symptoms among patients surviving 5 years after oesophageal cancer surgery Lena Martin, PhD, Karolinska Institutet, Sweden; Pernilla Lagergren, PhD RN, Karolinska Institute, Dept. of Molecular Med, Sweden AIMS To identify risk factors for postoperative weight loss = 15 % and to clarify the influence of postoperative nutritional symptoms on such weight loss in long-term survivors after oesophageal cancer surgery. METHODS A population-based study in Sweden 2001–2005 included surgically treated oesophageal cancer patients (97 %). Patient and tumour details were collected from medical records based to a predefined study protocol. Patients were followed up with questionnaires regularly until 5 years postoperatively. Current weight, weight at operation and average weight as adult were assessed by a self-administered questionnaire. Nutritional symptoms were selected from the EORTC QLQ-C30 and QLQ-OES18 including fatigue, nausea and vomiting, dyspnoea, appetite loss, diarrhoea, and dysphagia, eating difficulties, reflux and pain. Body mass index (BMI) at operation (\25 or =25), gender and preoperative weight loss (\10 % or =10 %) were tested as risk factors. The outcome was defined as weight loss of =15 % between preoperative weight and weight 5 years postoperatively. Logistic regression estimated relative risks of weight loss of =15 %, expressed as odds ratios (ORs) with 95 %
123
confidence intervals (CIs). Clinically relevant and statistically significant differences in nutritional symptoms between weight loss groups (\15 % and = 15 %) were analysed using linear regression models presented as mean score differences (MSD) with 95 % CI. RESULTS A total of 117 patients were included. Patients with preoperative BMI = 25 were at a threefold increased risk (OR 3.2, 95 % CI 1.4–7.3) for postoperative weight loss of =15 % 5 years after operation. Fatigue (MSD 13.7, 95 % CI 2.6–24.5), nausea and vomiting (MSD 13.1, 95 % CI 95 % CI 5.4–20.5), appetite loss (MSD 16.1, 95 % CI 5.2–28.7) eating difficulties (MSD 10.7, 95 % CI 1.3–20.0), and pain (MSD 11.0, 95 % CI 1.8–18.9) was clinically relevant and statistically significantly worse experienced symptoms among patients with weight loss of =15 %. CONCLUSIONS Overweight at time for operation is a risk factor for postoperative weight loss of =15 % in a long term perspective. Patients with weight loss of =15 % experience more severe nutritional symptoms 5 years after operation compared with patients without such weight loss.
(105.4) Health status and quality of life in long-term survivors of childhood leukaemia Julie Berbis, Department of Public Health, Medicine University, France; Ge´rard Michel, Pediatric Haematology Departments, University Hospital of Marseille, France; Pierre Bordigoni, Pediatric Haematology Department, University Hospital of Nancy, France; Pascal Chastagner, Pediatric Haematology Department, University Hospital of Nancy, France; Nicolas Sirvent, Pediatric Haematology Department, University Hospital of Montpellier, France; Franc¸ois Demeocq, Pediatric Haematology Department, University Hospital ClermontFerrand, France; Justyna Kanold, Pediatric Haematology Department, University Hospital ClermontFerrand, France; Dominique Plantaz, Pediatric Haematology Department, University Hospital of Grenoble, France; Karine Baumstarck, France; Pascal Auquier, Public Health Department, University Hospital of Marseille, France, France AIMS Advances in cancer treatment have changed the prognosis of children with acute leukaemia (AL). Currently, a significant proportion of survivors have undergone haematopoietic stem cell transplantation (HSCT), which remains a high-risk treatment. The French LEA program was initiated in 2003 to study the health status and QoL of long-term survivors of childhood acute leukaemia (AL). After presenting a comparison of late side effects and QoL after treatment with or without HSCT, the aim of this study was to compare survivor’s QoL with the QoL reported by a French general population of reference. METHODS From 2004 to 2009, survivors of childhood AL diagnosed after 1980 from five French onco-paediatric centres were included. Medical visits were conducted to detect the occurrence of late effects. Based on patient age at the time of evaluation, different questionnaires were used to assess the survivors’ QoL SF-36 for adults and VSP-A (proxy-version) for children and adolescents. To evaluate the association between HSCT and each type of late effect or QoL dimension, appropriate multivariate regressions were performed. QoL mean scores were compared with those obtained for age-and sex-matched French controls. RESULTS Among 943 patients included, 256 underwent HSCT (27.1 %). Of all the survivors, 674 (71.5 %) had at least one late effect. The risk of developing late effect for transplanted survivors was 4.0 [2.5–6.7] times increased. For children survivors, parent-reported scoring of QoL showed no significant differences between the treatment groups. Adult HSCT survivors reported lower physical dimension QoL scores than chemotherapy survivors. Compared with French norms, the all adult survivor’s group reported a significantly lower mental composite score, when physical composite score showed no significant difference. CONCLUSIONS Transplanted survivors have a high risk of developing late effects, resulting in decreased physical well-being in
Qual Life Res (2012) 21:1–132 adulthood. However, long after treatment completion, childhood leukaemia survivors report that effects on psychological well-being are more important than they are on physical QoL dimensions. This could require to be better taken into account by the actors of the health care system.
(105.5) Maximizing the effects of asthma interventions: predictors of perceived asthma control Owis J. Eilayyan, McGill University, Canada; Sara Ahmed, PhD, McGill University, Canada; Amede Gogovor, McGill University, Canada AIMS To estimate the extent to which symptom status, physical, mental, and social function predict the perceived asthma control (PAC) over a 6 month period of time. METHODS This study is a secondary analysis using data from a longitudinal study that examined health outcomes of asthma. 170 participants, recruited from a specialty clinic at a Montreal tertiary hospital, were evaluated over 2 time points. Multivariate logistic regression (MLR) was used to estimate the correlation between predictors. A path analysis model was used to estimate the predictors of PAC, which was modeled based on the Wilson & Cleary and ICF models. This path model hypothesizes that FEV, symptom and mental health affect the level of role of emotional, physical and social function. All of these variables can be affected by the personal factors and also affect the PAC directly and indirectly. RESULTS MLR indicates that FEV (B = -1.1) is a moderator of the impact of physical function (B = 1.2) on PAC. A new path model was modeled with more specified paths. It hypothesizes that symptom, mental health, role of emotional, physical and social function, and health care utilization (HCU) affect the PAC directly. We also hypothesize that PAC can be affected indirectly by (1) FEV through symptom and physical function, (2) Symptom through Mental health, physical and social function and HCU, (3) Mental health through role of emotional, physical and social function, (4) HCU through mental health, and (5) role of emotional through social function. The path model, however, indicates no significant moderator effect on PAC. It also indicates that social function and HCU predict PAC significantly over 6 months. Symptom status, mental health, and role of emotional affect the social function and HCU significantly. Symptom status, mental, physical, social function, and HCU explained 15 % of PAC over 6 months. The fit statistics indicates that we have a good fit model (Chi-square = 39.27, P value = 0.12, RMSEA = 0.043, and CFI = 0.97). CONCLUSIONS This path model suggests different predictors of PAC, to assist the care team in identifying individuals at risk of a future exacerbation and for designing tailored asthma interventions. MLR and path analysis provided different conclusions regarding predictors of PAC.
106: Utility Measures (106.1) Appropriate mapping functions in HRQL: a series of models mapping from FACT-G to EQ-5D Tracey A. Young, PhD MSc BSc CStat, University of Sheffield, United Kingdom; Donna Rowen, PhD, University of Sheffield, United Kingdom; Clara Mukuria, PhD, University of Sheffield, United Kingdom; John E. Brazier, PhD, Univ of Sheffield, United Kingdom; Louise Longworth, Jr., Health Economics Research Group, Brunel University AIMS Mapping, also known as cross walking, can be used to establish relationships between health related quality of life (HRQL) measures and produce mapping functions so that it is possible to obtain predictive HRQL scores for one instrument based on responses from
11 another. The aim of this study was to develop mapping functions to the generic EQ-5D HRQL measure from the cancer-specific FACT-G instrument. METHODS Mapping models were fitted to predict EQ-5D utility scores from the FACT-G measure using ordinary least squares (OLS), Tobit, two part models, and splining, and to EQ-5D item level responses using multi-level modelling. A series of models were fitted using FACT-G overall scores, domain scores or responses to specific items, interaction terms, squared terms, clinical and demographic information. Model goodness of fit was assessed using adjusted R2 and mean square errors (MSE). Predictive ability was assessed by comparing the observed mean to predicted values, number of cases predicted in full health (no problems on all dimensions), and percentage of scores within 0.1 and 0.05 of the actual EQ-5D value. RESULTS Mean observed EQ-5D value was 0.722 (SD = 0.224), ranging from -0.135 to 1 with 17 % of participants reporting full health. OLS models accurately predicted mean EQ-5D scores but only the models including specific FACT-G items were able to predict observations in full health (0.2 %). Adjusted R2 ranged from 0.33 to 0.52 and MSE ranged from 0.024 to 0.033. The best fitting model included significant item level responses. Two-part models used logistic regression to predict full health and OLS to estimate scores for those not in full health and were not as efficient at predicting mean scores (0.02 overprediction), but were better at predicting full health (2 % predictions). CONCLUSIONS OLS models were accurate at predicting mean ED-5D scores but poor at predicting very good health (fully active or restricted in strenuous activities only). The goodness of fit and predictive ability of the twopart models were best. Although marginally over-predicting the mean value, they were more accurate at predicting observations in full health. Further research into the usage of two-part models for mapping to EQ-5D utility scores is recommended.
(106.2) Comparison of the EQ-5D-5L and the EQ-5D-3L in a Spanish urological cancer sample Oriol Cunillera Puertolas, MSc, Fundacio´ IMIM, Spain; Carlos G. Forero, PhD, IMIM, Spain; Stefanie Schmidt, IMIM (Hospital del Mar Medical Research Institute), Spain; Juan M. Cabase´s, Universidad Pu´blica de Navarra; Juan M. Ramos Gon˜i, Servicio Canario de Salud, Spain; In˜igo Gorostiza, Basurto University Hospital—Osakidetza; Luis Rajmil, PhD, Age`ncia d’Informacio´ Avaluacio´ i Qualitat en Salut (AIAQS), Spain; Jordi Alonso, MD PhD, IMIM-Hospital del Mar Medical Research Institute, Spain; Montserrat Ferrer, MD PhD MPU, Institut Municipal d’Investigacio Medica (IMIM), Spain AIMS The new version of the widely used preference-based instrument Euroqol, the EQ-5D-5L, is called to overcome the flaws of its predecessor, the EQ-5D-3L. The aim of this study is to compare the performance of the new Spanish 5-level version in front of the classic EQ-5D-3L, and to study the calibration of their items using Item Response Theory, in a sample of Spanish patients with urological cancer. METHODS Data comes from an inception cohort of patients with either urinary bladder cancer or prostate cancer (n = 840) enrolled from 8 Spanish hospitals. The two versions of the EQ-5D were randomly ordered at the beginning or the end of the self-completed survey. Performance was studied evaluating redistribution properties in each dimension of EQ-5D between the two versions, informativity using Shannon indexes, and ceiling effects for both versions of the instrument. The calibration of the levels was studied applying the Item Response Theory approach, drawing category response curves for the 5 dimensions for the EQ-5D-3L and the EQ-5D-5L. RESULTS Global ceiling effect (percentage of individuals reporting no problems in any dimension) descended from 48.4 % for the EQ-5D-3L to 39.9 % for the 5-level
123
12 version. Among those patients reporting no problems at each EQ-5D-3L dimension, less than 16 % redistributed to non-extreme EQ-5D-5L response options; the ‘‘No problem’’ levels were the less redistributed ones, with the other EQ-5D-3L levels redistributing more than 54 % for all dimensions. The EQ-5D-5L showed significantly higher informative quantity (Shannon entropy H) and information efficiency (J) for all dimensions, with an average difference of 0.32 for H and 0.19 for J. Self Care dimension showed the lowest values and 3–5L improvement, from 0.58 (95 % CI 0.47–0.67) to 0.71(95 % CI 0.58–0.82) for H index, and from 0.19 (95 % CI 0.16–0.22) to 0.3 (95 % CI 0.25–0.35) for J index. IRT Category response curves showed an important overlap between the two mildest levels for all dimensions except Anxiety/Depression. CONCLUSIONS The new Spanish EQ-5D-5L increased the informativity compared to the 3-level version. The redistribution was somehow limited in the mild levels, and also IRT analysis suggested mild categories for most dimensions added less discriminant ability than more severe levels.
(106.3) Evidence from four countries on the validity of the discrete choice model in quantifying health-states for the new EQ-5D-5L Paul F.M. Krabbe, PhD, University of Groningen, Dept. of Epidemiology, Netherlands; Nancy Devlin, PhD, Office for Health Economics, United Kingdom; Elly Stolk, Erasmus University Rotterdam, Netherlands; Koonal Shah, Office of Health Economics; Mark Oppe, Erasmus University Rotterdam, Netherlands; Ben Van Hout, University of Sheffield; Simon Pickard, PhD, University of Illinois at Chicago, College of Pharmacy, United States; Feng Xie, MSc, National University of Singapore, Singapore AIMS To improve descriptive richness and discriminatory power of the EQ-5D, the EuroQol Group has recently developed a new EQ-5D by adding 2 intermediate levels to the existing 3 levels (EQ-5D-5L). Because of the increased number of health states (3125) associated to the EQ-5D-5L and the complexity of the TTO technique, discrete choice (DC) modeling was explored to derive values. The aim of this multi country study is to investigate the feasibility and comparability of DC for the EQ-5D-5L. METHODS A large study based on the same protocol was executed in Canada, The Netherlands, UK, and US. In each country at least 400 respondents were invited (representative samples from the general population). All respondents had to make a forced choice between 10 different pairs of EQ-5D-5L states. For this multi country study an interactive, data-collecting web-system was developed: EQ-VT software. Therefore, all the instructions and the format of the DC tasks were identical in all four countries. In total 200 pairs of states (order of the pairs and order within each pair completely randomized by the EQ-VT software) were judged in each country. DC analysis was performed utilizing the asclogit routine (Stata). RESULTS The average time needed for performing one DC task was between 32.5 (Netherlands) and 45.2 (UK) s. All 20 parameters (4 for each of the 5 domains) were statistically significant, except level 2 for Usual Activities in the Netherlands (p = 0.55). Only 3 modest illogical ordering of regression coefficients were observed (Netherlands: level 3 Pain, UK: Level 3 Usual Activities and Pain). Predictions for the complete set of 3125 EQ-5D-5L states were quite similar for the four countries. Correlations between the countries were high: from 0.87 (Netherlands vs. US) through 0.97 (Canada vs. UK). However, severe health states in the UK were valued more worse than in the other countries. CONCLUSIONS We conclude that the estimated parameters of the DC model are overly consistent and logical. Moreover, collecting DC data as done in this study (web-based, sample of the general population, no interviewer assistance) is feasible.{return}{return}
123
Qual Life Res (2012) 21:1–132 (106.4) Australian algorithms for the EQ-5D-3L and EQ-5D-5L Richard Norman, MSc, University of Technology, Sydney, Australia; Rosalie Viney, PhD, Centre for Health Economics Research & Evaluation, Australia; John E. Brazier, PhD, Univ of Sheffield, United Kingdom; Paula Cronin, University of Technology, Sydney, Australia; Madeleine T. King, PhD, Psycho-oncology Cooperative Research Group (PoCoG), Australia; Julie Ratcliffe, University of Flinders; Deborah Street, University of Technology, Sydney, Australia AIMS Until recently, economic evaluation of healthcare in Australia has relied on overseas estimates of utility weights for the EQ-5D, the most common utility instrument. Time Trade-Off weights have recently been developed (Viney et al., 2011), but the methods for doing so remain concerning. Alternative approaches, such as the discrete choice experiment, have been proposed (Bansback et al., 2012), but are yet to be widely adopted. The aim of this study was to estimate Australian algorithms for the standard 3-level EQ-5D and the recently developed 5-level version, for use in economic evaluation. METHODS Discrete choice experiments were run in a populationrepresentative online panel of 1,000 respondents per experiment. The experiments were constructed using a shift-generator approach, in which the effects of interest are explicitly designed for. Randomeffect probit models were used to analyse the data, which was then converted on to the necessary 0–1 scale for use in economic evaluation. Respondent heterogeneity was explored using a range of more sophisticated data modelling techniques, including mixed logits and generalised multinomial logits. RESULTS Mobility, Pain/Discomfort, and Anxiety/Depression were the dimensions most important to respondent for both surveys. The results reflected the monotonic structure of the two instruments. The range of scores was however larger than that observed in most Time Trade-Off studies in the area (the pits states in the 3L and 5L were valued at -0.516 and -0.676 respectively). More sophisticated modelling of heterogeneity improved the model fit, but did not impact on the mean response. CONCLUSIONS Discrete choice experiments offer an attractive alternative to existing preference elicitation techniques. However, significant methodological issues remain in their use and interpretation. These studies provide important methodological advances in the area, and will allow Australian policy makers to reflect local preferences for health in their decisions.
107: Outcome Evaluation in Cancer (107.1) Breast reconstruction and radiation: an analysis of quality of life and satisfaction among patients undergoing implant reconstruction Claudia R. Albornoz, Memorial Sloan-Kettering Cancer Center, United States; Evan Matros, MD, Plastic Surgery Department, Memorial Sloan Kettering Cancer Center, United States; Andrea Pusic, MD MHS FRCSC, Memorial Sloan Kettering Cancer Ctr; Anne Klassen, D Phil, McMaster University, Canada; Stefan Cano, BSc PhD, Peninsula College of Medicine & Dentristry, United Kingdom; Colleen McCarthy, MD MS, Memorial Sloan Kettering Cancer Center; Joseph J. Disa, Memorial Sloan-Kettering Cancer Center; Babak J. Mehrara, Memorial Sloan Kettering Cancer Center; Peter Cordeiro, Memorial Sloan Kettering Cancer Center AIMS Patient satisfaction and quality of life (QoL) are key outcomes following breast reconstruction. Among patients undergoing implant breast reconstruction, radiotherapy has been associated with high rates of scar tissue formation (capsular contracture) and reconstructive failure. However, long-term satisfaction and QoL have not been previously been measured using a condition-specific, validated patient-reported (PRO) instrument. The aim of this study is to
Qual Life Res (2012) 21:1–132 evaluate PROs among implant breast reconstruction patients to understand the impact of radiotherapy on satisfaction and QoL. METHODS A cross-sectional survey of 270 patients who underwent implant-based breast reconstruction with and without radiation was performed. Outcomes evaluated were satisfaction with breast, satisfaction with outcome, psychosocial well-being, sexual well-being and physical well-being (chest) using the BREAST-Q (Reconstruction Module). Univariate and multivariable analysis was performed to evaluate the impact of radiotherapy on patient satisfaction and QoL. RESULTS Mean follow up was 44.6 months for irradiated patients (n = 89) and 42.2 months for non-irradiated (n = 181). Age, body mass index (BMI), laterality and implant type was similar between groups. Satisfaction with breast was significantly higher for nonirradiated patients (58.7 vs. 53.6, p = 0.03). Satisfaction with outcome, psychosocial well-being, physical well-being and sexual well-being were not different between the groups. Multivariate analysis identified BMI, bilateral reconstructions and silicone implants as predictors of ‘satisfaction with breast’. Satisfaction with breast was not significantly different between patients who received radiotherapy before the mastectomy (lumpectomy + radiation) compared those with radiotherapy post implant reconstruction (55.4 vs. 52.4, p = NS). CONCLUSIONS Satisfaction with breast is higher for patient with implant reconstruction without radiation. There is no difference regarding the time of radiation. Additional predictors of satisfaction with breasts are lower BMI, bilateral reconstructions and silicone implants.
(107.2) Quality of life and satisfaction among prostate cancer patients followed through a disease-specific survivorship clinic Scott Gilbert, University of Florida, Department of Urology; Rodney Dunn, MS, University of Michigan Dept of Urology; Daniela Wittmann, University of Michigan; Jeffrey S. Montgomery, University of Michigan; John M. Hollingsworth, University of Michigan, Department of Urology, United States; Brent Hollenbeck, MD, MS, University of Michigan; John T. Wei, MD, University of Michigan, United States; James Montie, University of Michigan AIMS Integrating patient-reported outcomes, such as health-related quality of life (HRQOL), into survivorship clinics is critically important to: (1) provide empirical information on impairment and health consequences related to cancer and cancer treatment, (2) increase the clinical utility of HRQOL through more timely reporting to providers and patients and (3) tailor subsequent care for individual cancer survivors. To date, however, little evidence exists differentiating a tailored approach to survivorship care that includes HRQOL measures from routine care that does not systematically assess HRQOL. Given this gap, we sought to compare HRQOL and Satisfaction outcomes between a standard model of routine follow-up care and dedicated survivorship care. METHODS We conducted a prospective quasi-experimental before-after study consisting of 235 men with prostate cancer followed in a routine urology clinic following radical prostatectomy (before group) and 100 men with prostate cancer followed through a dedicated survivorship clinic following surgery (after group). The survivorship clinic (SC) consisted of two distinguishing features compared to the routine care (RC) provided through the urology clinic: (1) patientreported HRQOL outcomes were integrated into the clinic visit at the point-of-care, and (2) protocol-based management and referral for sexual and incontinence problems were tailored to each patient based on reported HRQOL impairments. For both groups, we assessed baseline and post-operative patient-reported quality of life at 6 and 12 months using the Expanded Prostate Cancer Index (EPIC) questionnaire. EPIC consists of 26 items divided among 5 domains: urinary incontinence, urinary irritation, sexual, bowel and hormonal. Satisfaction with
13 various elements of care (outcome, provider skill and manner, information and access) was evaluated using the Service Satisfaction Scale-Cancer (SSS-Ca). We evaluated differences in HRQOL and Satisfaction between patients followed with routine care and those managed through the survivorship clinic with the nonparametric Wilcoxon rank-sum test. RESULTS Although baseline EPIC scores were comparable between groups, HRQOL scores were generally higher among patients followed in the survivorship clinic at 6 months and 1 year compared to those followed with routine care. In particular, sexual function scores were significantly higher among patients managed in the survivorship clinic despite lower baseline scores (52.2 vs. 33.6 at 1 year, p \ 0.01). Satisfaction scores in all four domains were consistently higher in the survivorship clinic group compared to the routine care group (all p \ 0.05). CONCLUSIONS HRQOL and Satisfaction scores were higher among patients managed through a novel prostate cancer-specific survivorship clinic. These data suggest that in addition to high levels of patient satisfaction, disease-specific survivorship clinics may yield clinical and functional benefits compared to less tailored approaches to patient care following cancer therapy.
(107.3) Spiritual well-being in lung cancer survivors Marlene H. Frost, PhD, Mayo Clinic; Paul J. Novotny, Mayo Clinic-Rochester, United States; Mary E. Johnson, Mayo Clinic; Matthew M. Clark, Mayo Clinic; Jeff Sloan, PhD, Mayo Clinic, United States; Ping Yang, Mayo Clinic AIMS Spiritual well-being (SWB) among cancer survivors has not been well-delineated. Additionally, little is known about how SWB is affected over the trajectory of the disease process. The aims of this study were to examine the SWB of individuals with a diagnosis of lung cancer, to assess the stability of SWB over time and to identify factors associated with SWB. METHODS A prospective cohort of lung cancer patients who received care at Mayo Clinic was followed over a 10 year period. Study entry was at the time of diagnosis or referral to Mayo Clinic and participation involved annual survey using the Functional Assessment in Chronic Illness Therapy–Spiritual Well-Being, Short Form 8, and Quality of Life Linear Analog Scale Assessment. Associations were explored using Fischer’s exact, Chi-square, and Kruskal– Wallis tests and Spearman correlations. Linear regression was used to explore multivariate relationships. RESULTS There were 1,578 participants over a 10 year period of time. Group SWB scores were relatively high and stable over time (x¯ = 77.1–79.3, SD = 14.47– 18.46, possible scale 0–100). However, individual scores varied widely across almost the entire scale (2.1–100) and revealed a chaotic trajectory for SWB. Males, current smokers and those with higher pack-years experienced lower SWB compared to females, non-smokers and those with lower pack-years (p \ 0.0001, 0.0455, 0.0004 respectively). SWB was strongly associated with overall QOL. CONCLUSIONS SWB is an individualistic experience that can change dramatically over time for cancer survivors. Ongoing assessments are important to identify when intervention may be needed to improve SWB.
(107.4) Identifying and characterizing trajectories of QOL in persons with advanced cancer: important contributors to decreasing QOL in people with cancer Ana Maria Rodriguez, MSc BSc PT, McGill University, Canada; Nancy E. Mayo, PhD, McGill University, Royal Victoria Hospital, Canada AIMS When dealing with cancer, we aim at maintaining the Quality of Life (QOL) of patients. Yet minimal work examines predictors of QOL constructs over time. The aim of this study was to explore the
123
14 temporal sequence leading to optimal QOL over time of key components of the Wilson Cleary Model. METHODS 212 persons with a variety of advanced cancer from the McGill University Health Center (MUHC) were evaluated using nine patient-reported outcomes and seven direct measures over a period of 18 months. Rasch analysis was used to model symptoms (Appetite, Pain, Fatigue, Emotional status, Memory), function (physical, emotional, social), general health perceptions (GHP), and overall QOL latent constructs. Additionally, biological variables such as recalled weight loss, C-Reactive protein (CRP) serum concentration, Skeletal Muscle Mass, and Body Mass index were measured. The latent QOL construct was then modeled over time using ‘‘group-based modeling’’. Probability of group membership was finally predicted using linear regression using the biological, symptoms, function, and GHP constructs of the Wilson-Cleary model at study entry, which coincided with the time of cancer diagnosis. RESULTS The Rasch QOL model over time resulted in 5 distinct trajectories: a low and slightly linear increasing trajectory representing 26 % of the sample, two flat medium and high performing trajectories representing 26 and 17 % of the sample respectively, a quadratic increasing trajectory representing 25 % of the sample, and a linear abruptly decreasing trajectory representing 5 % of the sample. The variables or latent constructs from the time of diagnosis that statistically significantly predicted membership in a QOL trajectory were age, sex, cancer type, recalled weight loss, CRP, social support, emotional status, and fatigue. CONCLUSIONS Using a combination of Rasch, group-based trajectory modeling, and linear regression, we were able to discriminate between relevant QOL subgroups of patients. Most importantly, we were able to predict QOL trajectory from the time of diagnosis with advanced cancer. This enabled us to make preliminary conclusions about the most important contributors to QOL over time, and emphasizes the importance in assessing social support, fatigue level, and emotional status of patients from the time of diagnosis to maintain their QOL throughout their living experience with cancer.
108: Psychosocial and Health Interventions (108.1) Efficacy of cognitive behavioral therapy and physical exercise in alleviating treatment-induced menopausal symptoms in women with breast cancer: final results of a randomized controlled trial Neil K. Aaronson, PhD, The Netherlands Cancer Institute, Netherlands; Saskia F.A. Duijts, The Netherlands Cancer Institute/VU University Medical Center; Marc Van Beurden, The Netherlands Cancer Institute; Martijn M. Stuiver, The Netherlands Cancer Institute, Netherlands; Jacobien Kieffer, Ph.D, The Netherlands Cancer Institute, Department of Psychosocial Research, Netherlands; Myra S. Hunter, King’s College London; Hester S.A. Oldenburg, The Netherlands Cancer Institute AIMS To evaluate the effect of cognitive behavioral therapy (CBT), physical exercise (PE) and of these two interventions combined (CBT/PE) on menopausal symptoms (primary outcome), body image, sexual functioning, psychological well-being and health-related quality of life (HRQoL) (secondary outcomes) in breast cancer patients experiencing treatment-induced menopause. METHODS Breast cancer patients who reported experiencing treatment-induced menopausal symptoms (N = 422) were randomly assigned to CBT (N = 109), PE (N = 104), CBT/PE (N = 106) or to a waiting list control group (N = 103). Self-report questionnaires were completed at baseline, 12 weeks, and 6 months. Multilevel procedures were used to compare the intervention groups with the control group over time.
123
Qual Life Res (2012) 21:1–132 RESULTS Compared to the control group, the intervention groups had a significant decrease in levels of endocrine symptoms (FACTES; p \ 0.001; effect size (ES) 0.31–0.52), and urinary symptoms (BFLUTS; p = 0.002; ES 0.29–0.33), and an improvement in physical functioning (SF-36 PF; p = 0.002; ES 0.37–0.46). The groups that included CBT also showed a significant decrease in perceived burden of hot flushes and night sweats (HF/NS Problem rating; p \ 0.001; ES 0.39–0.56), and increase in sexual activity (SAQ Habit; p = 0.027; ES 0.65). Most of these effects were observed at both 12 week and 6 month follow-up. CONCLUSIONS CBT and PE can have salutary effects on endocrine symptoms and, to a lesser degree, on sexuality and physical functioning of breast cancer patients experiencing treatment-induced menopause. Future work is needed to improve the design and the planning of these interventions to improve program adherence.
(108.2) Empowermental Partnership in Coping System (EPICS) Index: measuring treatment impacts and quality of life in chronic illness Joshua C. Gandi, RN RPN HND BSc MSc, University of Jos, Nigeria; Paul S. Wai, University of Jos AIMS Empowermental Partnership in Coping System (EPiCS) model is a treatment approach within which practitioners use a ‘‘technically eclectic and theoretically consistent mix’’ that blends to achieve coping, healing and recovery among persons with chronic health challenge(s). However, despite the proliferation of instruments, there still exist some gaps in the practicality of measures that could adequately reflect holistic perspective of the model and effectively meet its unique informational needs. The main purpose of this study was to design and construct EPiCS index which meets psychometric requirements and the unique circumstances or informational needs to effectively measure treatment impacts and quality of life in chronic illness. METHODS Four categories of participants, namely peer-review participants (N = 10), expert review participants (N = 5), research assistants/facilitators (N = 10), and the target population participants (N = 108) were used. The 10 research assistants/facilitators were sampled out with the aid of Professional Quality of Life (ProQOL R-IV) scale while the 108 participants (M = 58, F = 50) were selected from four clinical settings by stratified random sampling technique. Using some laid-down criteria, the instrument construction followed three phases which include designing phase, development phase and micro pilot-testing phase. The data were analyzed by means of a principal component analysis, with varimax rotation. RESULTS All the original items met selection criteria, as reflected in the correlation matrix, showing how each item correlated against itself and against other items. The ‘initial’ indicates perfect correlation of the items (1.0) against themselves while the ‘extraction’ gives real value for each item. Pre-treatment reliability statistics showed favorable Cronbach’s alpha (A = .446 and B = .459) with a significantly high overall Cronbach’s alpha (.729) for all items (N = 40), while post-treatment reliability statistics had high Cronbach’s alpha (A = .465 and B = .596) with the overall Cronbach’s alpha (.833) more significantly very high. CONCLUSIONS The various indications of factorability, including the solution for the residuals, were good and the empirical finding holds the promise of defining EPiCS index as a composite instrument. However, there’s need for further study to validate this emerging instrument, towards ensuring robust psychometric requirements, for overall health-related quality of life assessment and research.
Qual Life Res (2012) 21:1–132 (108.3) Quality of life outcomes of a randomized counseling trial for cervical cancer survivors Lari Wenzel, PhD, University of California, Irvine; Kathryn Osann, Ph.D., University of California, Irvine, United States; Susie Hsieh, University of California, Irvine; Nissa Chantana, University of California, Irvine; Sandra Sappington, University of California, Irvine; Diana Marquez, University of California, Irvine; Edward Nelson, University of California, Irvine, United States AIMS Cervical cancer survivors experience quality of life disruptions associated with disease and treatment. Many of these disruptions persist into survivorship, and are potentially amenable to change. The purpose of this study is to determine if patients randomized to a psychosocial telephone counseling (PTC) intervention differed from those receiving usual care, comparing psychosocial and quality of life study endpoints. METHODS A randomized biobehavioral clinical trial was conducted for cervical cancer survivors enrolled = 9 and \30 months after diagnosis (n = 204), comparing PTC to usual care. PTC included five weekly sessions (i.e., interview, stress management, health and well-being, social and sexual concerns) and a 1 month booster. Patient-reported outcome measures and biological specimens (saliva and blood) were collected at baseline, 4 and 9 months post enrollment. Changes over time were tested by repeated measures analysis of variance. RESULTS The mean participant age was 44.7 (range 22–73), with 40 % Hispanic and 52 % nonHispanic White, and 41 % with a high school education or less. Approximately 52 % received surgery plus radiation and chemotherapy. Study drop-out rates were 18 % at 4 months, and 28 % at 9 months. Participants who dropped from the study prior to the 4 month evaluation reported significantly greater depression and anxiety at baseline compared to those who were retained (p \ 0.001 and p = 0.017 respectively). After adjusting for baseline scores, results indicated that participants in the PTC study arm had significantly improved depression (p = 0.015), coping (p = 0.022), and gynecologic problems (p = 0.043) compared to those randomized to usual care at the 4 month measurement interval. These differences were not sustained at 9 months post enrollment. CONCLUSIONS This large randomized trial confirmed benefits associated with psychosocial telephone counseling, designed to improve quality of life, stress and distress associated with disease and treatment in a multi-ethnic, underserved population. The significant differences observed at 4 months post enrollment, close in time to the period of active treatment, were not sustained at 9 months. This suggests that a more prolonged treatment approach is needed to realize extended benefits.
(108.4) Effectiveness of a patient empowerment programme for Chinese patients with diabetes mellitus Cindy L.K. Lam, MD, The University of Hong Kong, Peoples Republic of China; T.P. Elegance Lam, PhD, The University of Hong Kong, Hong Kong; William C. Wong, the University of Hong Kong AIMS To determine the effectiveness of a Patient Empowerment Programme (PEP) in Chinese patients with diabetes mellitus (DM) in terms of patient-reported enablement and clinical improvement of HbA1c. The PEP was designed to empower patients on selfmanagement and to enhance coping with illness. METHODS A longitudinal study on 2,407 Chinese adult patients with DM who participated in a 4-session PEP, recruited from public primary care clinics in Hong Kong. A random sample of 100 and 1,248 DM patients who had not participated in the PEP was recruited as comparison groups on the patient-reported outcome (PRO) and clinical outcome, respectively. PRO was measured by the Chinese version of
15 the Patient Enablement Instrument (PEI) at the end of the programme. The secondary (clinical) outcome was the change in HbA1c from baseline to 9–15 months after the programme. The proportion of subjects who reported being enabled (PEI score [0), the mean PEI score and the change in HbA1c were determined and compared between participants and non-participants. Multivariate linear regression was used to determine factors that were associated with PEI score. Pearson correlation was used to test the relationship between PEI and change in HbA1c. RESULTS 975 subjects completed the PEI at the end of the PEP, 96.3 % of them reported more enabled with a mean PEI score of 6.88. The outcomes were significantly better in PEP participants than non-participants (60.0 % reported more enabled with a mean PEI score of 1.81). Participants had a greater improvement in HbA1c than non-participants but the difference was not statistically significant. No significant predictor of PEI score was found and there was no significant correlation between PEI score and change in HbA1c. CONCLUSIONS The Patient Empowerment Programme (PEP) was effective in enabling DM patients to cope with their illness. The Patient Enablement Instrument (PEI) was sensitive in detecting the effect of the intervention and in differentiating between groups.
109: Symptom Monitoring in Clinical Research and Practice (109.1) The Myotonic Dystrophy Type-1 Health Index (MDHI): a reliable outcome measure for clinical trials and patient monitoring Chad Heatwole; Rita Bode, Northwestern University; Jeanne Dekdebrun, The University of Rochester; Nuran Dilek, The University of Rochester; Nicholas Johnson, The University of Rochester; Elezabeth Luebbe, The University of Rochester; William Martens, The University of Rochester; Charles Thornton, The University of Rochester; Richard Moxley, III, The University of Rochester AIMS Myotonic dystrophy type-1 (DM1) is a dominantly inherited multisystem disorder. The Myotonic Dystrophy Health Index (MDHI) is a custom designed, disease-specific multifaceted health survey designed to measure the most important physical, social, emotional, and symptom specific themes in this population. The items in this instrument were selected utilizing qualitative interviews and a crosssectional study of 278 adults with DM1. The MDHI is scored from 0 to 100 with a higher score representing greater disease burden. The primary aim of this research is to evaluate the test–retest reliability of the MDHI in adults with DM1. METHODS We serially administered the MDHI to all DM1 patients and each completed the MDHI twice within 4 weeks. We calculated total MDHI scores and 17 subscale scores for each participant. We analyzed the test–retest reliability for the entire MDHI, each of the individual questions, and each of the subscores in the MDHI. RESULTS Twenty-two DM1 patients 19–69 years of age completed the MDHI on two occasions with the average time between the two test periods being 10.2 days with a range of 5–31 days. Thirteen of the 22 participants were female. Two of the original 115 questions were dropped because of intraclass correlation coefficients (ICC) lower than 0.60. The resulting final MDHI had an ICC of 0.95. The MDHI’s seventeen subscores were found to have ICCs ranging from 0.97 (the hearing subscale) to 0.69 (the myotonia subscale). Fourteen of the seventeen subscales had an ICC of .80 or greater. CONCLUSIONS The MDHI is a valid and reliable instrument to study changes in patient reported health in the DM1 population. The total score of the MDHI is stable when administered over short periods of time. The reliability of the MDHI is sufficient for its use as an outcome measure in future clinical trials.
123
16 (109.2) Asking the right questions to get the right answers: using cognitive interviews to review the acceptability, comprehension and clinical meaningfulness of patient self-report adverse event items in oncology patients Patricia Holch, PhD, University Of Leeds, United Kingdom; Lorraine Warrington, University of Leeds; Lucy Ziegler, University of Leeds, United Kingdom; Ceri Hector, University of Leeds, United Kingdom; Ada Keding, University of Leeds, United Kingdom; Leon Bamforth, University of Leeds, United Kingdom; Clare Harley, PhD, University of Leeds, United Kingdom; Sandra A. Mitchell, PhD CRNP AOCN, National Cancer Institute; Ethan Basch, MD MSc, Memorial Sloan-Kettering Cancer Center; Galina Velikova, MD, University of Leeds, United Kingdom AIMS eRAPID (Electronic patient self-Reporting of Adverse-events: Patient Information and aDvice) is a programme to develop, evaluate and implement an online system for reporting and managing symptomatic adverse events (AE) during cancer treatment. Pivotal to the success of the system is that self-report items be consistently understandable by patients. Thus, patient reported (PR) Common Terminology Criteria for Adverse events (CTCAE) items for 17 selected common toxicities were examined using cognitive interviews (CI) with patients. METHODS 60 patients (34 female, 26 males) undergoing biological and chemotherapy and experiencing a range of toxicities were purposively sampled by age, gender and tumour group (median age 61.5, range 35–84, 12 breast, 12 gynaecological, 13 colorectal, 12 lung and 11 renal). Patients responded to items reflecting 77 symptomatic CTCAE toxicities drawn from two sources: (1) the National Cancer Institute’s Patient-Reported Outcomes Version of the CTCAE (PRO-CTCAE); and (2) items developed by Leeds Research Group (LRG). PRO-CTCAE items assess the presence, frequency, severity and interference dimensions of CTCAE toxicities; LRG items were developed by rephrasing CTCAE terms into a patient self-report format. Completion was on a touch-screen computer. A subsequent audio recorded CI explored the comprehension, interpretation and adequacy of items and response options via verbal probing techniques. RESULTS Content analysis and coding were performed in 3 iterative rounds (n = 20 patients/round). Overall 16 changes were made to NCI items: adding 9 items e.g. severity and interference of skin-rash, 5 descriptors i.e. nausea (feeling sick or queasy); and 2 item ordering changes. 37 changes were made toLRG items: 60 % related to meaning, 7 % order and 33 % response option. The NCI PRO-CTCAE items, using standardized questions were easy to understand. The LRG items required modifications to enable understanding whilst reflecting the CTCAE severity grading. CONCLUSIONS The changes made from patient feedback via the CI have resulted in a bank of consistently understandable PR toxicity items for remote self-report in the eRAPID project. The NCI-PRO-CTCAE items developed in USA were clear and understandable for UK patients. Working towards an internationally valid system for patient-reporting of AEs in clinical trials and in patient care is essential for improving patient safety and experiences during cancer treatment.
(109.3) Validity and utility of the M. D. Anderson Symptom Inventory (MDASI) among prostate cancer patients: data from E2Z02: symptom outcomes and practice patterns (SOAPP) Tito R. Mendoza, PhD, Univ. of Texas M.D. Anderson Cancer Ctr.; Fengmin Zhao, Dana-Farber Cancer Institute, Boston, MA; Linda J. Patrick-Miller, University of Chicago Center for Clinical Cancer Genetics, Chicago, IL; Charles S. Cleeland, PhD, UT-MD Anderson Cancer Center; Lynne I. Wagner, Northwestern University, Chicago, IL; Michael Fisch, University of Texas M. D. Anderson Cancer Center, Houston, TX
123
Qual Life Res (2012) 21:1–132 AIMS Optimal treatment outcomes for patients with prostate cancer are dependent upon rigorous symptom assessment. The MDASI, a patient-reported outcome measure of symptoms and their impact validated in a sample of patients of diverse cancer types has the potential to inform symptom experience and guide treatment specific to prostate cancer patients. METHODS SOAPP, an Eastern Cooperative Oncology Group (ECOG) study, was conducted at academic and community medical oncology clinics and included patients with prostate cancer of all stages and phases of care. At baseline and 4 weeks, patients completed a measure of symptom severity and functional interference (MDASI). Symptom experiences and psychometric properties of the MDASI in prostate cancer patients (n = 320) were analyzed. RESULTS The 5 most prevalent moderate/ severe symptoms reported at baseline were fatigue (34 %), disturbed sleep (25 %), drowsiness (18 %), pain (17 %), and dry mouth (15 %); moderate/severe nausea (5 %) and vomiting (2 %) were the least prevalent. At baseline ECOG performance status (PS) was = 2 in 9 % of patients. Pain, sadness, and difficulty remembering things occurred in 17 %, 14 %, and 13 % respectively, and among patients whose PS worsened, severity of these symptoms had the highest change scores at follow-up. Internal consistency and test–retest reliability of the MDASI were good, with Cronbach coefficient alphas = 0.84 and intra-class correlations = 0.78 for all subscales. Significantly higher scores and moderately large effects for the MDASI severity scale were reported by patients with poorer PS (2.2 vs 1.0, ES = .83) and tumor progression (2.4 vs 1.2, ES = .73) p \ .05. Results were similar for the interference scale, demonstrating MDASI known-group validity. Patients whose quality of life declined showed significantly greater increase in severity (1.1 vs .14, p \ .01) and interference (1.9 vs .28, p \ .004) from baseline to follow-up than patients whose quality of life was unchanged, demonstrating MDASI sensitivity to change. CONCLUSIONS The MDASI is a valid, reliable, and sensitive symptom assessment method for descriptive and clinical studies, and patient care in medical oncology outpatients with prostate cancer.
(109.4) Symptom frequency, severity and interference represent non-overlapping attributes of symptomatic treatment toxicity in patient-reported outcomes version of the common terminology criteria for adverse events (PRO-CTCAE) Thomas Atkinson, Memorial Sloan-Kettering Cancer Center, United States; Amylou C. Dueck, PhD, Mayo Clinic, United States; Yuelin Li, PhD, Memorial Sloan-Kettering Cancer Ctr., United States; Sandra A. Mitchell, PhD CRNP AOCN, National Cancer Institute; Lauren Rogak, Memorial Sloan-Kettering Cancer Center; Laura Sit, George Mason University; Antonia V. Bennett, PhD, Memorial Sloan-Kettering Cancer Center; Tito R. Mendoza, PhD, Univ. of Texas M.D. Anderson Cancer Ctr.; Steven B. Clauser, PhD, National Cancer Institute; Ethan Basch, MD MSc, Memorial Sloan-Kettering Cancer Center AIMS The National Cancer Institute’s Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) is comprised of a set of self-report items reflecting 78 symptomatic treatment toxicities. Nine of these symptoms are evaluated by three questions that capture patients’ experiences for the attributes of frequency, severity, and interference. To understand the degree to which each attribute contributes unique information about the symptom experience, we explored the strength of the relationship among the three attributes. METHODS The data were derived from an NCI-sponsored study examining the validity of PRO-CTCAE in patients (n = 869; 56 % female; median age 59; 32 % minority; 34 % high school or less) receiving treatment for mixed tumor types (27 % breast, 20 % lung, 16 % head/neck) at 4 NCI-designated cancer centers
Qual Life Res (2012) 21:1–132 and 5 NCI Community Cancer Centers Program sites (HHSN261201 000043C). Partial correlations were examined for the nine symptom terms from PRO-CTCAE that elicit all three symptom attributes. For each symptom, we calculated the pairwise correlation between two items independent of the effect of the third [e.g. r(f,i|s) for pain indicates the partial correlation between pain frequency and interference after controlling for severity]. RESULTS For pain, r(s,f|i) = 0.47; r(s,i|f) = 0.53; and r(f,i|s) = 0.28. The lower correlation between frequency and interference after partialing out severity indicates that pain severity accounted for the largest proportion of shared variance among the attributes. A generally similar pattern was observed for the other symptoms. However, for headache and myalgia, the degree of collinearity among attributes was approximately equal. Across symptoms, partial correlations ranged from: r(s,f|i) = 0.40–0.56; r(s,i|f) = 0.38–0.53; r(f,i|s) = 0.20–0.39. All partial correlation coefficients were statistically significant (p \ 0.001). CONCLUSIONS For a majority of the examined symptoms, the severity attribute makes the strongest unique contribution to the shared variance among the three attributes. Results also suggest that frequency and interference attributes add unique information, beyond that provided by severity. Future PRO-CTCAE consortium efforts are needed to confirm and extend these observations, and to determine how the symptom attributes should be interpreted to inform toxicity grading. Friday, 26 October 2012 201: Health Outcome Evaluation and Development (201.1) Patient reported outcome measures: including people with low literacy skills and/or learning disabilities Deepa Jahagirdar, Healthcare Improvement Scotland, United Kingdom; Thilo Kroll, University of Dundee; Karen Ritchie, Healthcare Improvement Scotland; Sally Wyke, University of Glasgow, United Kingdom AIMS Patient reported outcome measures (PROMs) are increasingly used for quality improvement and to assess patients’ progress. However their complexity means that over 1 in 5 people in the UK with low literacy skills and 1.5 million with learning disabilities may be excluded from using them. To address this gap, using COPD as an example, this paper describes the development of an evidence-based guide to support health professionals to use PROMs more inclusively. METHODS Semi-structured interviews with people with low literacy skills and/or learning disabilities and focus groups with health professionals and people with COPD were conducted. Using PROMs recommended for COPD as prompts—the generic EQ-5D and condition-specific St George Respiratory Questionnaire—discussions covered the format and the delivery of PROMs. The data were analysed thematically and the evidence was then used to build a multiformat practical guide for health professionals. RESULTS Three themes were identified: accessibility, ease of use and contextual factors. Accessibility issues included larger font sizes and more white space. ‘Ease of Use’ issues included the logistics and administration of PROMs. Participants preferred settings where they had enough time and privacy to complete PROMs, and help from people they trust, rather than waiting room completion. Contextual Factors included wider issues associated with completing PROMs such as understanding the purpose of PROMs to reduce intimidation (for patients) and focusing on using the PROM to monitor patients’ progress (for professionals). These themes formed the core of a practical guide developed in three formats to maximize its reach: an online interactive flash presentation, an interactive PDF report and an Easy Read hard copy version. In the Flash and PDF reports, the user can navigate through an overview of PROMs, the problem of exclusion and practical solutions to implement. CONCLUSIONS Adjusting the format of PROMs, giving patients a choice of where to complete PROMs and
17 ensuring PROMs’ purpose is clearly conveyed can help prevent the exclusion of people with low literacy skills and/or learning disabilities. Professionals can easily consult the user guide containing this information to administer PROMs to a larger proportion of their clients, supporting PROMs’ potential in clinical practice.
(201.2) A new measurement system to evaluate environmental barriers and facilitators to community participation Jin-Shei Lai, PhD, Northwestern University; Allen W. Heinemann, Northwestern University, United States; Noelle Carlozzi, University of Michigan; Sofia Garcia, PhD, Northwestern University; David Gray, Washington University; Elizabeth Hahn, MA, Northwestern University; Joy Hammel, University of Illinois; Susan Magasi, Northwestern University, United States; David Tulsky, PhD, University of Michigan; Patrick Semik, Rehabilitation Institute of Chicago AIMS Environmental barriers and facilitators are key components of social participation which significantly impact health-related quality of life for individual with disabilities. Valid scales measuring environmental factors are needed to identify barriers and facilitators to enable effective intervention. This presentation reports the development of such a new measurement system. METHODS We developed and tested 273 items covering 6 domains: ‘‘social environment’’, ‘‘economic quality of life’’, ‘‘built and natural environment’’, ‘‘assistive technology’’, ‘‘access to information and technology’’, and ‘‘systems, services and policies’’. These items/domains were generated via extensive literature review and results from focus groups; assigned to appropriate domains by experts and study investigators; and refined based on results of cognitive interviews with traumatic brain injury (TBI; n = 15), spinal cord injury (SCI; n = 15) or stroke (n = 15). Items were pilot tested in patients with TBI (n = 100), SCI (n = 105), and stroke (n = 100). Unidimensionality was evaluated for each domain using confirmatory factor analysis (CFA; criteria: comparative fit index [0.9; RMSEA \0.08, residual correlation \0.15; R-square [0.3). Rasch analysis was used to identify erratic response patterns of items (criterion: mean-square \1.4). The study team compared results of both CFA and Rasch analysis, with reference to item content, for item inclusion/ exclusion or revision. RESULTS Average age of study participants was 48 years, 64 % were male, 95 % were living at home, and 49 % were ambulatory using a mobility device. 117 items that met criteria of both CFA and Rasch analysis were retained for the domains of social environment (item n = 82) and economic quality of life (item n = 35). Unidimensionality of other four domains could not be confirmed due to excessive local dependency (i.e., high residual correlations between items). Thus, brief forms were developed with local dependent items removed or revised. Conditional items (e.g., types of assistive device used) were also added to better measure the environmental barrier/ facilitators. CONCLUSIONS This measurement system was developed to build upon the International Classification of Function conceptual model. The aforementioned results were used to modify this new measurement system, which is currently being tested in additional 600 patients. We expect to complete data collection by the end of 2012.
(201.3) Systematic evaluation of disease-specific health-related quality of life instruments for patients with shoulder pathology Stefanie Schmidt, IMIM (Hospital del Mar Medical Research Institute), Spain AIMS Shoulder pathology is a prevalent condition, especially in an aging society. Evaluating health-related quality of life (HRQL) in these patients is crucial, but depends largely on the right selection, use and interpretation of the most appropriate instrument. The aim of this
123
18 study was to conduct a systematic expert evaluation about the measurement properties of specific HRQL instruments developed for patients with shoulder pathology. METHODS We conducted a systematic literature review (PUBMED) to identify available shoulder disease-specific HRQL instruments and their metric information. Titles and abstracts of the retrieved articles were reviewed by two experts independently to select the information that met the inclusion criteria: patient-reported, shoulder-specific, at least a 3-point scale response option, neither being population nor specific for a certain shoulder pathology. For every instrument identified, two experts reviewed the corresponding full text articles and applied the EMPRO tool (‘‘Evaluating the Measurement of Patient-Reported Outcomes’’). EMPRO has been designed to systematically evaluate the conceptual measurement properties of HRQL instruments. EMPRO scores are transformed to a scale from 0 (worst) to 100 (best). RESULTS In the systematic literature review, 52 disease-specific HRQL instruments were identified out of 2,325 articles. Eleven instruments met the inclusion criteria and were considered for the systematic evaluation. The number of articles with metric information varied from 2 to 25. Overall EMPRO scores for the original version of the questionnaires ranged from 82.7 to 24.7. ‘‘Reliability’’ and ‘‘interpretation’’ were the attributes where most data was lacking. Best reliability scores were reached for ASES-s (100), FLEX-SF, SPADI and SST ([70). The highest validity score was obtained for FLEX-SF (86.7), followed by ASES-s, SDQ-NL and SST ([70). SDQ-NL and SST reached the highest possible score (100) in the responsiveness attribute, followed by ASES-s, OSS, SDQ-UK, SPADI and SRQ ([75). CONCLUSIONS The majority of the 11 evaluated instruments presented EMPRO scores C50 in the most important attributes. ASES-s, SDQNL and SST presented the highest scores in the majority of the 8 attributes evaluated. Nevertheless, the selection of the right instrument depends mainly on the study purpose. The presented detailed EMPRO results can facilitate this selection.
(201.4) Evaluation of a newly developed questionnaire to measure motivation for self-care among people living with type 2 diabetes (T2DM): The Motivation for Exercise and Diet Questionnaire (MED-Q) Shana B. Traina, Janssen Global Services; Susan D. Mathias, MPH, Health Outcomes Solutions; Hilary H. Colwell, Health Outcomes Solutions; Ross Crosby, PhD, Neuropsychiatric Research Institute AIMS Effective diabetes management requires healthy eating and appropriate physical activity. Motivation is necessary for execution of these behaviors among people living with T2DM. No existing measures evaluating motivation for self-care behaviors in this population could be identified. Therefore the 17-item MED-Q, based on the Theory of Planned Behavior, was developed. Content of the MED-Q was developed based on literature reviews, input from people living with T2DM, and healthcare professionals and measures motivation to exercise, diet, and maintain/lose weight. This study evaluated some psychometric properties of the MED-Q. METHODS An online survey of US adults self-reported as living with and taking medication for T2DM was implemented. The survey, which included the MED-Q, Multidimensional Diabetes Questionnaire (MDQ), and Summary of Diabetes SelfCare Activities (SDSCA), was completed at baseline and 3–7 days later. Internal consistency reliability (using Cronbach’s alpha), test–retest reliability [using intra-class correlation coefficients (ICC) for those reporting no health changes], convergent validity (via Pearson correlations), known-groups validity (based on BMI calculated using selfreported height and weight), and minimum important difference (MID) [using distribution-based approaches] were evaluated. RESULTS Of the 1,015 respondents, (n = 529 at 3–7 day retest), 60 % were male, mean age 63 years, 84 % white/non-Hispanic, 64 % married, and 59 % had a
123
Qual Life Res (2012) 21:1–132 BMI [ 30. Factor analyses confirmed a 4-factor solution; Exercise, Diet, Weight, and Locus of Control (LOC). Cronbach’s alphas and ICCs were above the acceptable threshold of [0.70 for Exercise, Diet, and Weight and were [0.60 for LOC. Correlations between MED-Q items and related MDQ and SDSCA items were high (0.60–0.84, p \ 0.01)). There were significant differences between BMI categories on all MEDQ scales (p \ 0.001), except LOC. Combining distribution-based results yielded an 8–11 point MID estimate for Diet, Weight, and Exercise, and 6–9 points for LOC. CONCLUSIONS The MED-Q has acceptable psychometric properties; however, if used as an outcome measure, its responsiveness should be evaluated. Measuring motivation for healthy behaviors among those with T2DM using the MED-Q may be useful when personalizing treatment approaches that require adherence to selfcare activities.
(201.5) Understanding meaning from measurement in movement disorder clinical trials: Rasch analysis of the Parkinson’s Disease Dyskinesia Scale (PDYS-26) Jennifer Petrillo, PhD, Novartis, United States AIMS The Parkinson’s disease Dyskinesia Scale (PDYS-26) is a PRO instrument of daily functioning for people with PD levodopa-induced dyskinesia (PD-LID). Its key role in clinical trials should be supported by evidence that it is both clinically meaningful and scientifically sound. The clinical content of the PDYS-26 is well considered, but its psychometric performance in clinical trials has received less attention. Rasch Measurement Theory (RMT) methods are being increasingly used as they provide detailed diagnostic information about the measurement properties of PRO instruments. In this study, we used RMT methods to examine the PDYS-26. METHODS PDYS-26 clinical trial data from 1,112 patient measurements were evaluated at six time points: Screening, Baseline 1, Baseline 2, Day 21, Day 42, and Day 84. RMT based psychometric performance was evaluated across all six languages and time points. The scope included: item threshold ordering; item fit; targeting of the scale to the sample; item dependency; reliability; and stability. RMT analyses were carried out using RUMM2030. RESULTS All items had ordered thresholds. However, there was noticeable item misfit, and high Chisquare values, especially for PDYS1 (Sitting), PDYS3 (Talking face-toface to family/friends), PDYS9 (Writing), and PDYS19 (Keeping still when stressed/excited). There was reasonable targeting to this sample, but reduced item coverage at the higher end of the daily functioning continuum. Eight pairs of items reflected residual correlations [0.30, indicating local dependency. Reliability was high (PSI = 0.96). Responses between different countries (language versions) showed differential item functioning in seven items. CONCLUSIONS These mixed findings suggest that there are areas for improvement in the PDYS-26 in PD-LID. Evidence from the fit and threshold ordering analyses suggests that there is likely more than one clinical concept underpinning the scale. This has an impact on measurement precision, in this range, and thus the ability of the scale to detect differences between and changes within higher functioning patients. Item dependency, and stability between different countries (language versions), also need to be addressed. Our study further supports the benefits for using RMT methods to help to solve measurement problems by identifying limitations in scale performance and indicating how they can be overcome.
202: Trials and Intervention Evaluation (202.1) Joint modeling of longitudinal health-related quality of life data and overall survival Divine Ewane Ediebah, MSc., EORTC HQ, Belgium; Corneel Coens, MSc, EORTC HQ, Belgium; Efstathios Zikos, MSc, MA, EORTC HQ, Belgium; Chantal Quinten, MSc, EORTC, Belgium;
Qual Life Res (2012) 21:1–132 Jolie Ringash, MD, The Princess Margaret Hospital, Canada; Carolyn C. Gotay, PhD, UBC School Pop & Pub Health, Canada; Eva Elfriede Greimel, PhD, Medical University Graz, Austria; David Osoba, FRCPC, QOL Consulting, Canada; Madeleine T. King, PhD, Psycho-oncology Cooperative Research Group (PoCoG), Australia; Henning H. Flechtner, PhD, Otto-von-Guericke University (KKJP), Germany; Martin Taphoorn, PhD, Medical Centre Haaglanden, Neurology, The Hague, The Netherlands; J. Schmucker von Koch, PhD, University of Regensburg, Medical Ethics, Regensburg, Germany; Joachim Weis, PhD, University of Freiburg, Psychooncology, Freiburg, Germany; Charles S. Cleeland, PhD, The University of Texas MD Anderson Cancer Center, Houston, TX, United States; Bryce B. Reeve, PhD, University of North Carolina at Chapel Hill, Unites States; E.F. Smit, PhD, Free University Hospital, Dept. of Pulmonology, Amsterdam, the Netherlands; Andrew Bottomley, PhD, EORTC, Quality of Life, Brussels, Belgium on behalf of the Lung Cancer Cooperative Group, Quality of Life Department and Patient Reported Outcome and Behavioral Evidence (PROBE) AIMS In cancer clinical trials, outcomes may co-vary. Specifically, treatment-related impairment of health-related quality of life (HRQoL) may hasten death, and thus mask survival benefits of treatment. Our aim is to investigate the combined effect of treatment on patient’s HRQoL scores and overall survival (OS). METHODS We analyzed data from an EORTC randomized clinical trial (RCT) of 480 advanced non-smallcell lung cancer (NSCLC) patients comparing gemcitabine + cisplatin (arm B), paclitaxel + gemcitabine (arm C), versus standard arm paclitaxel + cisplatin (arm A). HRQoL was assessed via the global health status/QOL (GHS) scale of the EORTC QLQ-C30, at baseline and after each chemotherapy cycle. Joint modeling was used to assess the combined effect of treatment on HRQoL and survival. We compared results from different analysis strategies by applying Cox model with treatment and time-dependent GHS (effect of GHS on treatment) [M1] and the joint model (M2) to examine the treatment effect. RESULTS The hazard ratio (HR) for treatment effect was estimated as 0.99 (95 % CI 0.78–1.28) for arm B and as 1.28 (0.99–1.65) for arm C for M1. For M2 the HR was estimated as 1.27 (0.83–1.94) for arm B and as 1.55 (1.01–2.39) for arm C. The HR for the effect of GHS on treatment was estimated as 0.88 (0.83–0.91) for M1 and as 0.78 (0.71–0.85) for M2. The risk of death is higher in M2 (27 % for arm B and 55 % for arm C compared to the standard arm A) than in M1 (1 % for arm B and 28 % for arm C compared to the standard arm A). Also there was an increased risk of death for the effect of GHS on treatment in M2 (22 %) than in M2 (12 %). CONCLUSIONS Our findings suggest that part of the survival benefit of a new treatment can be masked by the negative effect that treatment has on HRQoL. In our study, up to 27 % of the theoretical treatment efficacy was lost through impaired HRQoL affecting survival. Our analysis also supports the association between HRQoL and OS. More attention to effective supportive care during treatment could potentially correct negative HRQoL effects and unmask survival benefits.
(202.2) Is ‘In-Trial’ quality of life guidance lacking? A systematic review employing qualitative content analysis Derek G. Kyte, MSc MACP MMACP, University of Birmingham, United Kingdom; Heather Draper, University of Birmingham; Jonathan Ives, University of Birmingham; Clive Liles, University of Birmingham, United Kingdom; Adrian Gheorghe, University of Birmingham; Melanie Calvert, PhD, The University of Birmingham, United Kingdom AIMS Anecdotal evidence from clinical academics/data managers involved in clinical trials, suggests a lack of guidance for researchers involved in Health-Related Quality of Life (HRQL) data collection
19 and processing (i.e., ‘In-Trial’ activity). Particular concerns have been raised with regard to: patient information provision, which professionals should have access to HRQL data, how a researcher should deal with sensitive information contained within a quality of life questionnaire and the associated implications for individual patient care within a trial. The purpose of this review was, therefore, to determine the extent and nature of published guidelines addressing these issues. METHODS Systematic review of 1,362 articles identified 18 eligible papers containing ‘In-Trial’ guidelines. Two independent authors undertook a qualitative content analysis of the selected papers. Guidelines presented in each of the articles were coded according to an a priori coding frame, developed using a concept-driven strategy and refined during the pilot phase using data-driven methods. The definitive coding frame demonstrated reliability (pooled Kappa 0.86–0.97), and validity (\2 % residual category coding). RESULTS The majority of guidelines present were concerned with ‘Pre-Trial’ activities (72 %), for example outcome measure selection and study design issues, or ‘Post-Trial’ activities (16 %) such as data analysis, reporting and interpretation. ‘In-Trial’ guidelines represented less than 10 % of all guidance across the papers reviewed. The content in this dimension primarily focused on compliance and quality control, proxy assessment and reporting of data collection. No guidelines referred to the management of sensitive HRQL data or its access and there was minimal consideration of patient information provision/ understanding or whether HRQL data should inform patient intervention within a trial. CONCLUSIONS The findings highlight that guidelines aimed at HRQL data collection and processing in clinical trials are lacking. It is of concern that no guidance exists for researchers involved with the handling of sensitive HRQL data, as it has been reported that this activity may be associated with considerable anxiety and ethical tension. Further research is needed to produce guidelines aimed at supporting researchers so they can manage patient data appropriately and deal effectively with these tensions.
(202.3) Core outcome sets and the COMET (core outcome measures in effectiveness trials) initiative; improving the efficiency and value of the research process Jane Blazeby, MD, University of Bristol, United Kingdom; Doug G. Altman, University of Oxford, United Kingdom; Mike Clarke, Queen’s University Belfast; Elizabeth A. Gargon, University of Liverpool, United Kingdom; Paula R. Williamson, University of Liverpool, United Kingdom AIMS Selection of specific outcomes or domains is crucial to clinical trials designed to compare directly the effects of different interventions in ways that minimise bias. Evidence shows that trials in specific conditions often report different outcomes, or address the same outcome in different ways. Inconsistency in reported outcomes causes well known problems for those who attempt to synthesise evidence, and meta-analyses often exclude key studies because relevant outcomes are not reported. Furthermore, the measured outcomes may not always be important to patients or health service users. Much would be gained if an agreed core outcome set (COS) of a minimum number of appropriate and important outcomes was measured and reported in all clinical trials/audits in a specific condition. This paper will describe the COMET Initiative (Core Outcome Measures in Effectiveness Trials) which aims to foster and facilitate methodological research in the area of standardising outcomes, to provide guidance on developing COS, methods to include user involvement, integrating patient reported outcomes (PROs) into core outcome sets, and preparing reporting standards for such projects. METHODS The COMET Initiative started in 2010 supported by UK MRC methodology Hub funding. (http://www.comet-initiative.org). It brings together researchers interested in the development and application of
123
20 COSs and it will provide methodological guidance in this area. Currently work is ongoing to identify, collate and maintain all relevant resources in an online searchable database (http://www.come tinitiative.org/studies/search) which will allow sharing of ideas and COSs and prevent duplicity of work. RESULTS Searches have identified 130 completed projects that are developing and reporting COSs in various areas of health. Examples exist where the involvement of patients in COS development has identified an outcome that might not have been considered by practitioners alone. Examples of the multiplicity of PROs in specific disease sites exists meaning that there is need for work to define what is ‘core’ amongst these many outcomes. Examples of COSs with limited or no patient involvement have been identified. COMET is also working with COS development in some 25 clinical areas, with a further 32 in discussion. In the months since the launch of the COMET website and database (August 2011), 1,061 searches have been undertaken, with 3,289 individuals from 87 countries visiting the site. CONCLUSIONS COS will increase the efficiency and value of the research process. Design of new trials will be simplified, risk of measuring inappropriate outcomes reduced, and selective reporting of outcomes less likely. By improving the evidence base, COMET will make it easier for people to make well-informed decisions about healthcare.
(202.4) Patient satisfaction with participation, adverse events and hospitalizations in phase II/III NCCTG clinical trials: was it worth it? (N0392) Cynthia Chauhan, MSW, Mayo Clinic Breast SPORE; Jeff Sloan, PhD, Mayo Clinic, United States; Pamela Atherton, Mayo Clinic; Amylou C. Dueck, PhD, Mayo Clinic, United States AIMS Patient (pt) satisfaction with their clinical trial experience may give insight to quality of life (QOL) factors affecting accrual, retention and outcome. This trial examined pt opinion of their experience. METHODS Pts enrolled on designated North Central Cancer Treatment Group (NCCTG) phase II or III treatment trials completed the Was It Worth It (WIWI) satisfaction assessment after cycle 1 and treatment completion. This secondary analysis focused on the relationship between pt satisfaction and adverse event (AE) incidence. Chi-Square tests compared pt opinion across subgroups. RESULTS As of 02/01/2012, 271 pts were enrolled on 25 protocols and treated at 79 sites. 97 % of pts were in phase II studies and 89 % had stage IV disease. Pt satisfaction was no different when comparing pts with/ without AEs at 1 month and the end of treatment (p = 0.99 and p = 0.43). Pts with an AE were more likely to report a worsening of QOL during the study (19 vs. 9 %, p = 0.026). Pts with hospitalizations were more likely to report worsened QOL (35 vs 20 %, p = 0.01) and a worse than expected study experience (24 vs 11 %, p = 0.01). Pts hospitalized during cycle 1 were more likely to report a detrimental effect on overall QOL (OR = 3.32, p = 0.005). CONCLUSIONS Contrary to popular beliefs, experiencing adverse events did not influence pt satisfaction with trial participation. Hospitalizations reduced pt satisfaction and overall QOL. Assessing pt satisfaction will inform future study design that can potentially improve pt accrual and retention.
203: Analysis and Interpretation of Outcomes (203.1) Matched health-related quality of life reference mean scores as a proxy for ‘‘true baseline’’ when assessing recovery Mohammed A. Johar, B.Sc.(Hons), M.Sc., Karolinska Institutet, Sweden; Therese Djarv, MD PhD, Karolinska Institutet, Sweden; Pernilla Lagergren, PhD RN, Karolinska Institute, Dept. of Molecular Med, Sweden
123
Qual Life Res (2012) 21:1–132 AIMS When assessing recovery of health-related quality of life (HRQL) values from a background population can be used as a proxy for ‘‘true baseline’’. Co-morbidities negatively affect HRQL and may therefore be important to be included as a factor together with age and gender when calculating matched HRQL reference mean scores (MRS). The aim was to demonstrate the use of MRS when evaluating the effects of co-morbidities on HRQL recovery over time in longterm survivors of oesophageal cancer. METHODS Data from a Swedish nationwide cohort of oesophageal cancer patients treated surgically during 2001–2005 was used. Patients were followed up with HRQL measures longitudinally over 5 years using the validated EORTC QLQ-C30 and QLQ-OES18. A cross-sectional populationbased study in 2008 randomly selected people (71 % response rate = 4,910 participants) from the background population to respond to the same questionnaires as above. Each patient was matched with the reference population on age (±2.5 years), gender, and co-morbidities with a mean of 97 people per patient. The mean scores of these matched people were used as MRS for the respective patient. Linear mixed models analyzed the effects of co-morbidities on longitudinal recovery of HRQL. The adjusted mean score differences (MSD) and 95 % confidence intervals (CI) were calculated. RESULTS Among the 141 patients who survived 5 years, 56 % had co-morbidities. Patients with co-morbidities had worse MRS in all aspects of HRQL compared to patients without co-morbidity. Recovery in HRQL over the followup period was generally better in patients without co-morbidities than patients with co-morbidities. Statistically significantly differences between groups were seen for global quality of life (MSD 8.5, CI 3.1–13.8), physical function (MSD 6.2, CI 1.3–11.0) and fatigue (MSD -8.6, CI -14.7 to -2.5). The time interaction within the group with co-morbidity was statistically significant for all HRQL aspects. CONCLUSIONS Patients with co-morbidity have lower MRS scores in general showing that co-morbidity should be adjusted for when using MRS as a proxy for ‘‘true baseline’’. When using MRS, recovery in global quality of life, physical function, fatigue was statistically significantly better in patients without co-morbidities than patients with co-morbidity in long-term survivors of oesophageal cancer.
(203.2) Analysis of the Spanish PROMIS pain behavior item bank for language related DIF Leo Morales, MD, Group Hlth Research Institute; Sylvia H. Paz, PhD, UCLA; Karen L. Spritzer, UCLA; Ronald D. Hays, PhD, UCLA David Geffen AIMS To investigate the newly translated PROMIS pain behavior (PB) item bank for language DIF. The PB bank includes 39 items measuring self-reported external manifestations of pain including observable displays (sighing, crying), pain severity behaviors (resting, guarding, facial expressions, and asking for help), and verbal reports of pain. All items have six response categories. METHODS The study included 527 Spanish-speaking and 678 English-speaking adults all in the United States. The Spanish sample mean age was 36 (SD = 10.5); 63 % were female; and 32 % had a high school diploma or less, 33 % had some college, and 34 % had college degree or more. Nine percent spoke only Spanish, 53 % spoke Spanish better than English, 37 % spoke Spanish and English equally well, and \1 % spoke English better than Spanish. By comparison, the English sample mean age was 51 years (SD = 18.9 years); 54 % were female; and 19 % had a high school diploma or less, 35 % had some college and 46 % had a college degree or more. The PROMIS PI bank consisted of 41 items each with 5 response categories. The PROMIS protocol was used to translate the items to Spanish and involves two forward translations, 1 reconciled version, 1 back-translation, comparison of original with the back-translation, and reviews by 3 bilingual experts. Cognitive
Qual Life Res (2012) 21:1–132 interviews with native Spanish speakers followed to evaluate the comprehension of the items. Analyses included assessments of dimensionality, DIF and DIF impact simulation studies. RESULTS Coefficient Alpha for the item bank was 0.98; the mean item-rest correlation was 0.71 (range from 0.58 to 0.77); and a one-factor confirmatory model had a CFI of 0.79 and TLI of 0.98—all suggesting one underlying dimension. LORDIF in R was used to assess DIF. Applying a criterion of a pseudo-R2 = 0.02, 6 items were found to show DIF (24, 31, 40, 41, 47, 50) (see Appendix). CAT simulation studies using Firestar were conducted to assess DIF impact. Theta estimates for the Spanish sample generated using DIF and nonDIF English and equated-Spanish item parameters were moderately highly correlated (r = 0.93). Excluding DIF items from the item bank resulted in somewhat more correlated thetas (r = 0.96). CONCLUSIONS Six items in the Spanish PB item bank showed significant DIF. The impact of the DIF in CAT simulation studies was small but not negligible. We recommend either using equated Spanish item parameters to score items with DIF or excluding items with DIF from the item bank altogether. English parameters may be used for nonDIF items.
(203.3) Power and sample size determination for the group comparison of PROs in IRT: robustness of a method based on the Cramer-Rao bound Myriam Blanchin, University of Nantes, France; Jean-Benoit Hardouin, PhD, University of Nantes—EA4275, France; Alice Guilleux, University of Nantes—EA 4275; Ve´ronique Se´bille, University of Nantes, France AIMS Patient-reported outcomes (PROs) are important as endpoint in clinical trials and epidemiological studies. Models from the Item Response Theory (IRT) are increasingly used to analyze these particular outcomes that introduce a latent variable in the model as these outcomes cannot be directly observed. A method based on the Cramer-Rao bound and the mixed Rasch model has been proposed for sample size and power determination for the comparison of PROs in cross-sectional studies comparing two groups of patients when an IRT model, the Rasch model, is intended to be used for analysis. Compared to simulations, the so-called Cramer-Rao method has already shown good results regarding power determination when the model fits the data well. The objective of this work was to study the robustness of the Cramer-Rao method to some deviations of the mixed Rasch model. METHODS The local independence hypothesis states that the item responses are conditionally independent from the latent variable. In the mixed Rasch model, the latent variable is considered as a random variable. In the CramerRao method, the latent variable is assumed to follow a normal distribution. Datasets were simulated with dependent items or a nonnormal distribution of the latent variable. Powers for the test of group effect obtained from these datasets by simulations were compared to the results from the Cramer-Rao method that assumed that the items are locally independent and that the latent variable distribution is normal. RESULTS First results indicate that the shape of the latent variable distribution has no impact on the Cramer-Rao method. Previous studies have shown that the violation of the local independence did not affect the power of the Wald test for another IRT model, the 2-PLM. We can expect the same results for the Rasch model and that the power for the Cramer-Rao method will still be close to the power for datasets containing dependent items. CONCLUSIONS The Cramer-Rao method to determine power for group comparison of PRO seems to be robust to both studied deviations of the mixed Rasch model: the local independence of the items and the normality of the latent variable.
21 (203.4) Beyond age and gender adjustment when using Quality of Life reference data: a propensity score matching approach Francesco Cottone, PhD, GIMEMA (Italian Group for Adult Hematologic Diseases), Italy; Gary Collins, PhD, Univ of Oxford, Ctr for Statistics in Med, United Kingdom; Fabio Efficace, PhD, Italian Group for Adult Hematologic Diseases, GIMEMA, Italy AIMS Direct comparisons of Health Related Quality of Life (HRQoL) outcomes between non-randomized groups might be biased, as outcomes are confounded by imbalance in pre-treatment patient characteristics. Such bias can be reduced by adjusting on observed covariates. This is the setting of HRQoL comparisons with reference data, where age and gender (AG) adjustment is commonly used for this purpose. However, other observed covariates can be used to lessen this bias and yield more precise estimates. The objective of this study is to show that more accurate HRQoL comparisons with reference data can be obtained, accounting for few additional covariates beyond gender and age by propensity score matching (PSM) methods. METHODS Simulations were used to compare five different adjustment methods in seven diverse scenarios. Three methods were AG-based, i.e. weighted means, linear regression and exact matching. Remaining methods were unadjusted PSM and PSM with further regression adjustment. For each scenario, 1,000 case groups and one reference dataset were generated according to the corresponding impact of five covariates (including age and gender) on the probability of assignment to the case group. Results were compared in terms of mean scores, bias and percent bias reduction. The same methods were applied to normative data of the SF-36 questionnaire for the general health (GH) scale. RESULTS PSM-based methods resulted in better performance across all simulated scenarios for estimating a prespecified 6-point difference in HRQoL (on a [0–100] scale). Ranges of estimates were [-7.1; -6.0] for PSM and [-12.1; -7.4] for AG. The corresponding bias reduction ranges were [89.4 %; 99.3 %] and [33.2 %; 84.3 %] respectively. On a real data set, cases reported a worse mean general health than healthy controls, with estimated differences of -6.4 and -6.6 points for PSM-based methods and -10.5, -10.6 and -11.2 for AG-based techniques (above the 8-points clinical significance of the SF-36 GH scale). CONCLUSIONS PSM methods have the potential to significantly reduce bias and we recommend that wherever possible they should be considered in the design and analysis of non-randomized HRQoL based studies.
204: Mental Health and Patient Outcomes (204.1) Cognitive reserve: an important emerging construct for quality-of-life research Carolyn E. Schwartz, ScD, DeltaQuest Foundation, Inc., United States; Brian R. Quaranto, DeltaQuest Foundation; Ralph HB Benedict, Dept. Neurology, Jacobs Neurol. Instit., Univ. of Buffalo; Erin Snook, Department of Kinesiology, University of Massachusetts Amherst; Timothy Vollmer, Dept. Neurology, Univ. Colorado Denver, Rocky Mountain MS Ctr; Bruce D. Rapkin, PhD, Yeshiva University AIMS The concept of cognitive reserve is increasingly relevant in research on neurological disorders and aging populations. Comprised of both pre-morbid childhood enrichment activities (passive reserve) and current leisure and educational stimulation (active reserve), it has prognostic significance in predicting better disease trajectories in people with Alzheimer’s disease, multiple sclerosis (MS), and the elderly. We introduce this construct to quality-of-life (QOL) researchers by describe our empirical work on its measurement, patterns of relationships between passive and active reserve and health and well-being outcomes, and the underlying cognitive
123
22 appraisal processes. METHODS Cross-sectional data (n = 1,142) were drawn from the North American Research Committee on MS (NARCOMS) Registry, from whom additional patient-reported outcome (PRO) data were collected. Cognitive reserve was measured using the Stern and Sole-Padulles measures. Generic (SF -12v2, Perceived Deficits Questionnaire, Ryff Psychological Well-Being, Diener Satisfaction with Life Scale) and disease-specific (PatientDetermined Disease Steps, Performance Scales) PROs were included. Regression models examined relationships between cognitive reserve, demographics, and PROs. The QOL Appraisal Profile assessed underlying appraisal processes. Multivariate models compared groups within appraisal parameter sets, and t-tests compared mean item responses within appraisal parameters. RESULTS The cognitive reserve measures assessed distinct but related constructs, passive reserve reflecting pre-morbid enrichment and active reserve reflecting current cultural stimulation activities. Active and passive reserve are associated with better generic and disease-specific outcomes, independent of demographic factors. Active reserve is relevant for perceived disability, cognitive deficits, physical and mental health, and eudemonic well-being. In contrast, passive reserve has an independent effect only for eudemonic well-being. Appraisal analyses revealed that high-reserve individuals emphasize the positive, focus on controllable life aspects, and less fantasy-based. Hierarchical multivariate models comparing the two aspects of reserve revealed more differences comparing high-low active than passive reserve. Thus, although premorbid differences influence appraisal, there are more notable differences related to current engagement in intellectually and other stimulating activities that maintain brain health. CONCLUSIONS Adaptation and compensation in the face of changing pathology may be better understood by considering the concept of cognitive reserve. Appraisal processes underlying selfreport elucidate paths to resilience and adaptability in the face of disability.
(204.2) The challenge of translating the picture naming subtest of the repeatable battery for the assessment of neuropsychological status (RBANS) Elisabeth Rendu, MAPI Institute, France; Brian Harel, CogState; Athena Nomikos, Cogstate; Angela Caveney, University of Michigan; Catherine Acquadro, MD, MAPI Research Trust, France; Caroline Anfray, MAPI Institute, France AIMS The Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) was initially developed as a neuropsychological screening tool for dementia. It is now commonly used to screen for abnormal cognitive decline across a range of clinical populations. The Picture Naming subtest, a confrontation naming task requiring subjects to name 10 line drawings, is used to assess basic verbal functions and language skills. This abstract presents the challenges faced during the translation of 40 pictured words (e.g., bed, hammer, scissors, candle) into 13 different languages (i.e., Bulgarian, Chinese for Taiwan, Czech, English and French for Canada, German, Hindi, Hungarian, Kannada, Korean, Marathi, and Spanish for Mexico and the USA). METHODS The traditional translation process had to be adapted. In each country, a thorough forward translation was performed with translators and a speech therapist, all of whom were native speakers of the target languages. Criteria for selecting substitute items were developed for when the direct translations were culturally inappropriate for a country (i.e., substitutes should match the frequency of use of the source terms, belong to the same semantic field, be familiar to all people and be drawable). Finally, a clinician’s review was performed in each country. RESULTS Twenty-nine of the 40 source words were found appropriate for all countries, leaving 11 to be substituted. The languages that required the most substitutions
123
Qual Life Res (2012) 21:1–132 were Bulgarian, Chinese, Hindi, Hungarian, Kannada, Korean, and Marathi, whereas very few substitutions were needed for languages culturally close to US English such as Canadian English and Canadian French, German, and US Spanish. The most culturally problematic picture words were ‘‘Barn’’ and ‘‘Toboggan/Sled’’. ‘‘Barn’’ was replaced by the word ‘‘House’’ and its corresponding picture into eight out of the 13 languages. In Canadian French, the picture for ‘Ball’ was modified because the direct translation would have been the same for the source terms ‘Ball’ and ‘Balloon’. CONCLUSIONS This process demonstrated that identifying words equivalent to the source ones used to assess basic verbal functions and language skills is challenging and requires the collaboration of various specialists in each target language.
(204.3) Trajectories and predictors of the course of mental health in persons with spinal cord injury Christel M. Van Leeuwen, PhD, Center of Excellence for Rahabilitati Rehabilitation Center De Hoogstraat; Trynke Hoekstra, VU University Medical Center Amsterdam; Casper F. Van Koppenhagen, Rehabilitation Center De Hoogstraat; Sonja De Groot, Reade, centre for rehabilitation and rheumatology; Marcel Post, PhD, Institute for Rehabilitation Research, Netherlands AIMS To study the course and predictors of mental health in the period between the start of active spinal cord injury (SCI) rehabilitation and 5 years after discharge. The hypothesis was that different mental health trajectories can be identified. METHODS Multi-centre prospective cohort study in eight Dutch rehabilitation centres with specialized SCI units with measurements at the start of active rehabilitation, after 3 months, at discharge, 1, 2, and 5 years after discharge. Persons (n = 206) with recently acquired SCI and age between 18 and 65 years participated in the study. The main outcome measurement was the five-item Mental Health Index (MHI-5) with a total score between 0 (lowest mental health) and 100 (highest mental health). Latent Class Growth Mixture Modeling and multinomial regression analyses were used for the analysis. RESULTS Latent Class Growth Mixture Modeling revealed 5 trajectories: (a) high scores (above 80) at all time-points (52 %), (b) low scores (60 or below) at all time-points (4 %), (c) early recovery from 40 to scores above 70 (13 %), (d) intermediate scores from 60 to scores above 70 (29 %), and (e) severe deterioration of scores above 70 to scores below 30 (2 %). Pain, gender, and educational level were predictors to distinguish between the five trajectories. CONCLUSIONS Five different mental health trajectories were identified between the start of active rehabilitation and 5 years after discharge. About a third of the persons with SCI still perceived moderate to severe mental health problems 5 years after discharge. Pain, gender, and educational level only predicted a small part of the variance in mental health trajectories.
(204.4) The effects of stroke on depression: a functional outcome of patients with ischemic stroke Lv Meijun, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Hai Ying, Department of Neurology, Affiliated Hospital of Liaoning University of TCM, Peoples Republic of China; Yu Li, Department of Neurology, Affiliated Hospital of Liaoning University of TCM, Peoples Republic of China; Yu Xiu, department of neurology, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yuan Dongchao, department of neurology, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Qi Wencheng, department of neurology, Liaoning University of Traditional Chinese Medicine, Peoples
Qual Life Res (2012) 21:1–132 Republic of China; Lu Yujuan; Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS Studies have revealed that patients diagnosed with stroke developed post stroke depression (PSD) which affects the quality of life of those patients. This study is aimed to investigate the effects of restoration stroke on depression. METHODS Patients diagnosed with ischemic stroke were recruited (National Institutes of Health Stroke Scale, NIHSS = 5, and ‘‘level of consciousness’’ of NIHSS = 1). The NIHSS and Barthel Index (BI) were used to assess the neurological function and quality of life (QOL). Three mood items (depressed mood, psychic anxiety and somatic anxiety) were extracted from the Hamilton Depression Rating Scale (HAMD) to study the patients’ depression. The Spearman correlation was used to investigate the relationship between depression and BI. Additionally, patients were divided into neurological function deficient and non-deficient groups by the item score (0 and non-0) in each of the 11 NIHSS items, to investigate the effect of neurological function deficient on depression. We also measured the differences for depression between acute and restoration stroke by t test. RESULTS Of the 1,982 patients (age 62.59 ± 10.27 years; 1,101 male, 55.55 %), 1,781 were in the restoration stage (3 weeks to 6 months, 89.86 %). The HAMD scores were varied agreeing with the severity of the BI (p = 0.05), the correlation coefficient was 0.108 (p = 0.00) for the BI and the HAMD. Between the neurological function deficient and non-deficient groups, significant differences (p = 0.05) were observed in all items except for ‘‘motor arm’’ and ‘‘motor leg’’. NIHSS items were all related with the HAMD (r = 0.068–0.139, p = 0.05) except for ‘‘level of consciousness’’, ‘‘motor arm’’ and ‘‘motor leg’’ and the item ‘‘best language’’ had the highest correlation coefficient (r = 0.139). The restoration stage had higher depression scores than the acute (p = 0.00), which means restoration stroke patients may be more depressive. CONCLUSIONS Nearly all aspects of stroke appeared to influence depression. Depression affects the QOL of stroke patients. Therefore, we should pay attention to PSD, especially in patients with language function deficient, and in the restoration stage to improve their quality of life.
205: Quality of Life Measure Development (205.1) Testing the generalisability of the EORTC QLQ-C30 core questionnaire across primary cancer sites using multi-group confirmatory factor analysis Daniel S. J. Costa, PhD, Psycho-oncology Co-operative Research Group (PoCoG), Australia; Neil K. Aaronson, PhD, The Netherlands Cancer Institute, Netherlands; Peter Fayers, PhD, Aberdeen Medical School, United Kingdom; Julie F. Pallant, PhD, School of Rural Health, Univ of Melbourne, Australia; Galina Velikova, MD, University of Leeds, United Kingdom; Madeleine T. King, PhD, Psycho-oncology Cooperative Research Group (PoCoG), Australia; The MAUCa Consortium, Psycho-oncology Research Group AIMS The EORTC Quality of Life Questionnaire is a widely-used cancer-specific quality of life instrument consisting of a core set of 30 items (QLQ-C30) supplemented by cancer site-specific modules. The purpose of this paper was to examine the generalisability of the measurement properties of the QLQ-C30 across patients with cancers originating in different primary sites. METHODS Multi-group confirmatory factor analysis was employed to test whether a measurement model of the QLQ-C30 was invariant across cancer sites. The measurement model comprised the multi-item subscales of the established
23 structure of the QLQ-C30 (physical functioning, role functioning, emotional functioning, social functioning, cognitive functioning, pain, fatigue, nausea/vomiting and overall health and quality of life). Invariance between each pair of 10 cancer site groups was assessed on a total sample of 2,616 patients by comparing the comparative fit index (CFI) for a model where the factor loadings were free to vary to a model where the factor loadings were constrained to be equal for the corresponding items in each group. RESULTS The fit of the measurement model for the entire sample was adequate. Five of the 45 pairwise comparisons between primary cancer site groups exhibited differences (delta CFI [ 0.01), all of them involving the testicular cancer group. The response distributions for the testicular cancer patients indicated these patients had few difficulties, reflecting the fact that they were disease-free survivors. CONCLUSIONS The multiitem domain structure of the QLQ-C30 appears to be generalisable across a variety of primary cancer sites, including all the most common cancer sites. These results confirm that the QLQ-C30 can be used with confidence across sites, and suggest that comparisons of domain scores across different primary cancer sites are robust.
(205.2) Metric properties of the Bladder Cancer Index proposed for patients with bladder cancer Angels Pont, Hlth Svcs Research Unit, Spain; Yolanda Pardo, Institut de Recerca Hospital del Mar (IMIM); Stefanie Schmidt, IMIM (Hospital del Mar Medical Research Institute), Spain; Ricard Riel, Institut Catala` de la Salut; Montserrat Ferrer, MD PhD MPU, Institut Municipal d’Investigacio Medica (IMIM), Spain AIMS The Bladder Cancer Index (BCI) is a disease-specific instrument that was developed to assess the Health Related Quality of Life (HRQoL) in a wide range of bladder cancer states and treatments (Gilbert et al. Cancer 2007; May 1;109(9):1756–62). The aim of this study is to evaluate the metric properties of the Spanish version of this questionnaire. METHODS Prospective study of a cohort of patients with a new diagnosis of bladder cancer from seven Spanish Hospitals. Participants fulfilled the BCI, SF-36, and EQ-5D previous treatment and after 6 and 12 months. BCI is a 36-item questionnaire that is composed by Urinary, Bowel and Sexual Domains, including also Function and Bother subscales for each domain. Cronbach’s alpha was calculated to assess internal consistency. Validity was assessed using a multi-trait-multi-method matrix of correlations between BCI, SF-36 and EQ-5D. Discriminate validity was tested by comparing means between known groups (stage 0a, 0is, I vs II-IV). RESULTS 216 patients with a mean of 69.4 years (SD = 10.5) were recruited, 81.6 % presented a non-muscle invasive stage (0a, 0is, I). Missing items were low except for Sexual Domain: 24.1, 31.3 and 28.5 % of patients with missing scores on Sexual Domain, Sexual Function and Sexual Bother. There’s a moderate ceiling effect in Urinary and Intestinal domains, and in Sexual Bother subscale (41.9, 34.2 and 46.2 % respectively). While only Sexual domain presents high floor effects (20.2 %). Internal consistency was high for all scores, ranging from 0.70 to 0.95. Inter-scale correlations between scales of the BCI support the convergent validity of this instrument: moderate correlations among function and bother subscales of urinary and bowel scales (0.507 and 0.510, respectively); and low correlations between the scales of different domains (ranged from 0.20 to 0.43). The correlations of the BCI scales with other generic HRQoL instrument (SF-36 and EQ5D) were low to moderate (0.04–0.47). Significant differences were found in mean BCI Urinary and Bowel Domains and its subscales according to the cancer stage, showing lower HRQoL impact for nonmuscle invasive stages. CONCLUSIONS Our findings indicate that the Spanish version of BCI is equivalent to the original instrument and similarly reliable and valid. Although further research is needed to test responsiveness, results suggest that the BCI may already be used in Spain.
123
24 (205.3) The WHOQOL-OLD module: psychometric properties of the Turkish national data pool Erhan Eser, MD, Celal Bayar University, School of Medicine, Dept. Public Health, Turkey; Sultan Eser, MOH, Izmir Cancer Registry, Turkey AIMS WHOQOL-OLD Turkish version is the only valid generic instrument for the assessment of HRQOL in Turkish older adults. This study aimed to present the psychometric properties of the Turkish National WHOQOL-OLD data pool by Turkish WHOQOL centre. METHODS WHOQOL-OLD is the older adults module of the WHOQOL generic instrument. It has 6 domains. Possible scores range is between 0 and 100 and higher scores indicate better QOL. Confirmatory approach was employed during the reliability and validity analyses. RESULTS National WHOQOL-OLD data pool consists of 4,555 subjects aged 65 and over. Of the data 37.7 % came from representative samples, 55.5 % consisted institutional and 6.8 % mixed residency data. Of the older adults, 52.5 % were female; mean age was 73.4 (sd 15.6); 44.3 % of them had no school education and 28.8 % had primary education. The mean WHOQOLOLD domain scores were as follows: Sensory-Abilities (45.7 ± 14.8); Autonomy (40.3 ± 23.1); Past, present and future activities (45.1 ± 15.2); Social participation (51.6 ± 20.8); Death and dying (48.9 ± 22.0); Intimacy (60.4 ± 21.0) and the Overall score (48.8 ± 12.5). Internal consistency (IC) of the overall data was good in 4 of 6 domains (Autonomy, a = 0.76; Social participation, a = 0.81; Death and dying, a = 0.78 and Intimacy, a = 0.93). Unsatisfactory IC results were obtained for Sensory-Abilities (a = 0.23) and Past, present and future activities domains (a = 0.14). Overall IC was acceptable (a = 0.84). If item deleted alpha analyses revealed that the items no 3,4,5 and 10 were found to have weak contributions to the IC of their domains. Scale success for the majority of domains were good (75.0–100.0 %). As for the Construct validity of WHOQOL-OLD: Known groups validity results were satisfactory (scores: younger age [ older age; male [ female; higher education [ lower education). Confirmatory factor analysis results revealed a moderate to poor scale fit (CFI = 0.90 and RMSEA = 0.126). CONCLUSIONS The preliminary results indicated the use of Sensory abilities and Past, present and future activities domains with caution based on these results. A larger representative national sample is needed to come evaluate the problematic items and also for normative evaluations of the WHOQOL-OLD Nevertheless even at this point Turkish WHOQOLOLD can be useful to assess HRQOL of the elderly in clinical practice and for research purposes.
(205.4) Cross-cultural development of an EORTC measure to assess quality of life (QoL) in patients with chronic myeloid leukemia (CML) Fabio Efficace, PhD, Italian Group for Adult Hematologic Diseases, GIMEMA, Italy; Massimo Breccia, Sapienza University, Rome, Italy; Susanne Saussele, Universita¨t Heidelberg, Mannheim, Germany; Ute Kossak-Roth, Universita¨t Heidelberg, Mannheim, Germany; Annarita Cardoni, Italy; Giovanni Caocci, University of Cagliari, Cagliari, Italy; Weichu Chie, National Taiwan University, Taipei, Taiwan; Adel Naeem, University of Baghdad, Baghdad, Iraq; Ourania Nicolatou-Galitis, University of Athens, Greece; Kim Cocks, University of York, York, UK; Marco Vignetti, Health Outcomes Research Unit, GIMEMA Data Center, Rome, Italy; Michele Baccarani, University of Bologna, Bologna, Italy; Franco Mandelli, Health Outcomes Research Unit, GIMEMA Data Center, Rome, Italy; Mirjam Sprangers, PhD, University of Amsterdam, Amsterdam, The Netherlands
123
Qual Life Res (2012) 21:1–132 AIMS The treatment of chronic myeloid leukemia has changed dramatically with the advent of targeted therapies (TT). In this work we present current results of the cross cultural development of an EORTC QoL questionnaire for patients with CML. METHODS The development is following the EORTC Quality of Life Group module development guidelines as well as the Food and Drug Administration (FDA) recommendations. In phase I, an extensive literature search on potentially relevant HRQOL issues for CML patients was conducted and the list was administered, on an international basis, to patients and CML experts. In phase II, a shorter list of relevant issues, identified through statistical analyses on patient and expert phase I evaluations, was constructed into items compatible with the EORTC QLQ-C30. In the current phase III, patient evaluations are performed on an international basis. To further strength content validity of the questionnaire, CML Patient advocacy groups were also involved. RESULTS More than 600 CML related articles were scrutinized and an initial list of potentially relevant QoL 74 issues was drafted. To date 575 CML patients participated in this project. Three hundred eight-four CML patients were interviewed in 9 European and nonEuropean countries and 191 patients provided their input through international patient advocacy websites. The data stemming from this latter group were used for supportive analyses only. In phase I, 59 CML experts were also interviewed on possible relevance of issues. At the time of interviews, patients were in first line treatment with Imatinib (56 %), in first line treatment with second generation of tyrosine kinase inhibitors (TKIs) (11 %) and in second line treatment with second generation of TKIs (31 %). Forty-six percent were in treatment for more than 5 years ensuring that our sample included a reasonable percentage of long-term cancer survivors. The initial list of 74 issues was reduced to 30 items currently tested in phase III. CONCLUSIONS We expect the final EORTC QoL-CML questionnaire to be used as a key outcome measure, in conjunction with the EORTC QLQ-C30, to provide the patients’ perspective on the burden of treatment in future clinical trials.
206: Patient Reported Outcomes in Clinical Practice (206.1) Threshold for determining clinically relevant urinary incontinence using the Incontinence Symptom Index (ISI) in a population-based cohort Anne M. Suskind, MD, University of Michigan, United States; Rodney Dunn, MS, University of Michigan Dept of Urology; Daniel M. Morgan, University of Michigan; John O. L. DeLancey, University of Michigan; John T. Wei, MD, University of Michigan, United States AIMS Urinary incontinence is a highly prevalent condition that affects up to 50 % of women and 14 % of men in the U.S. Many, however, do not receive treatment, as primary care providers lack appropriate quality of life measures to screen patients for clinically significant urinary incontinence. In this study, we used a population-based sample of women to determine the optimal thresholds to adapt a quality of life measure as a clinical screening tool. METHODS The Establishing the Prevalence of Incontinence (EPI) Study is a case control study of community dwelling women ages 35–64. Participants with and without incontinence were invited to have urodynamic testing and a clinical evaluation to establish the presence of and type of urinary incontinence. The Incontinence Symptom Index (ISI) is an instrument that was developed for urinary incontinence treatment and is brief (10 items), clinically relevant, and validated. The ISI has sub-domains for both stress and urge urinary incontinence (each with scores that range from 0 to 12) and a total severity domain (with scores that range from 0 to 32). Of those in the EPI study, 214 completed the ISI. Their ISI scores were calibrated to the physician’s final diagnosis using diagnostic test criteria (sensitivity and specificity) to determine
Qual Life Res (2012) 21:1–132
25 professionals; and should pay careful attention to the content and format of the feedback intervention. The optimal use of PROMs will remain elusive until issues associated with effectiveness are revealed.
(206.3) A systematic review and narrative review of qualitative studies about the use of PROs in clinical practice Fig. 1
optimal threshold scores where women were likely to have clinically relevant stress and urge incontinence. RESULTS ROC curves for the stress and urge sub-domains and the total severity domain are shown in Fig. 1. An ISI threshold of = 3 (sensitivity 77 %, specificity 73 %) for the stress urinary incontinence sub-domain, of =5 (sensitivity 86 %, specificity 76 %) for the urge urinary incontinence sub-domain, and of =7 (sensitivity 84 %, specificity 75 %) for the total severity score were determined to be optimal indicators for clinically relevant stress and urge incontinence. CONCLUSIONS Threshold scores were developed to determine clinically relevant cases of stress and urge urinary incontinence, with high sensitivity and specificity. These scores will be useful to assist primary care providers in guiding management and referral for patients with urinary incontinence.
(206.2) Does feedback of patient-reported outcomes to healthcare professionals result in an improvement in patient outcomes? A systematic review and meta-analysis Maria B. Boyce, MA, University College Cork, Ireland; John Browne, PhD, University College Cork, Ireland AIMS To investigate the impact of providing feedback on patientreported outcome measures (PROMs) to healthcare professionals on patients’ outcomes.; METHODS Systematic review including controlled studies investigating the effectiveness of PROMs feedback and the impact at the individual patient-level and the group-level (providers/hospitals). Study quality was assessed using a modified Jadad scale. Changes in PROMs scores were examined by identifying differences between groups that received feedback and didn’t receive feedback. Where possible, results were pooled using standardized mean difference (SMD) for meta-analysis.; RESULTS Twenty studies met the selection criteria and three reported sufficient data for pooling. Standardized mean differences were calculated using PROMs scores. The meta-analysis showed that feedback had no effect on outcomes, but this was not statistically significant (SMD: -0.08 [95 % CI, -0.33 to 0.17], P = 0.54). In a broader narrative synthesis, 11 studies (55 %) found one positive result favouring PROMs feedback. PROMs feedback was more successful when targeted towards specific patient groups as 82 % of studies with a positive result focused on a specialized group of patients. In the majority of studies (95 %) feedback was provided at the individual patient-level, the remainder provided feedback at a group-level to clinicians as a performance indicator.; CONCLUSIONS The use of PROMs as a quality improvement tool is rapidly expanding; however research into the effectiveness of PROMs feedback is lagging far behind. There is uncertainty regarding the use and effectiveness PROMs in clinical practice. PROMs have different functions leading to lack of clarity on their intended role. Traditionally, PROMs feedback has been provided at the patient-level. However, the PROMs Programme (England) introduced PROMs feedback at the group-level as a performance indicator. Future studies should investigate the impact of PROMs: at the group-level to clinicians or hospitals; on a greater variety of specific healthcare issues where variation in outcomes is present; on different healthcare
Daniela C. Gonc¸alves, University of Oxford, United Kingdom; Elizabeth J. Gibbons, MSc, University of Oxford, United Kingdom; Raymond Fitzpatrick, PhD, Univ of Oxford, Dept of Public Hlth, United Kingdom; Jose M. Valderas, MD, PhD, MPH, University of Oxford, United Kingdom AIMS To identify patients and healthcare professionals perspectives regarding the use of patient reported outcome (PROs) in clinical practice, in order to inform the development of systems for the routine collection and use of this information. METHODS A systematic search strategy was developed, based on an integrated model for health outcomes, previously published reviews, and direct contact with experts. A snowball technique was also employed. Only empirical studies that employed qualitative or mixed methods were considered. RESULTS Twelve publications were identified, referring to ten studies published between 1990 and 2010 in specialised care (n = 8), and in primary care and community settings (n = 3). They used a variety of methods for data collection (e.g., focus groups, semistructured interviews). Eight studies reported on perspectives of health professionals, and five on patients’ perspectives. Only five studies were based on actual experience of routine use of PROs. Both providers and patients considered that the use of PROs in clinical practice would be feasible and that it would enhance clinical practice, by improving the understanding of the patient’s health status, communication, and detection of problems. Healthcare professionals emphasised that the use of PROs could be considered as an indicator of patient-centeredness and good quality of care. Potential risks included impact on the consultation and the ensuing rapport between professionals and patients. In relation to eventual implementation, both patients and professionals agreed that there was a need for tailoring measurement to the specific needs of each patient. The main barriers to the use of PROs identified by providers were associated costs, staff and patients burden, and absence of evidence regarding their clinical utility. Patients highlighted the redundancy of some items, and logistic impediments associated with questionnaire completion. CONCLUSIONS Patients and health professionals anticipate both risks and benefits in the routine use of PROs in clinical practice. Perceived implementations barriers have changed little over the last decades and fall under two broad categories: logistic/practical and perceived limited clinical utility. Better use of IT and computer adaptive tests might overcome the former, whereas the latter could be informed through additional evidence obtained from experiments (clinical trials) studying the impact of using PROs in clinical practice.
(206.4) The Norwegian study in RENEWING HEALTH: development, implementation and evaluation of an intervention with an electronic diabetes diary—the few touch application (FTA)—with and without health counseling. A randomized controlled trial Lis Ribu, PhD, Oslo and Akershus University College, Norway; Eirik Arsand, Norwegian Centre for Integrated Care and Telemedicine, Norway; Astrid Torbjornsen; Heidi Holmen; Astrid Grottland AIMS The overall objectives of this study is to evaluate whether the introduction of technology supported self-management with an electronic diabetes diary—the Few Touch application (FTA)— through self-monitoring and health counseling, produce benefits in
123
26 terms of HbA1c, health-related quality of life (HRQOL), self-care behaviors and lifestyle changes such as improved dietary habits and increased physical activity, compared with usual care. This is a low cost-, and low intensity treatment intervention, and the FTA is designed to give patients a central role in the management of their own disease. METHODS The MRC Framework for complex interventions provided a structured approach to design the intervention. The theoretical phase entailed reviewing the literature to identify current practices and earlier research. In Phase I (modeling), findings were used to design the intervention. The participants receive a mobile phone with the self-management system FTA, and they are requested to self-monitor their blood glucose levels, eating behavior and physical activities at a daily basis, but at a phase and at intervals that they feel are manageable. The health counseling intervention is theoretically based on principles from Motivational Interviewing, Prochaska and DiClemete’s Stages of Change model, as well as cognitive behavioral therapy. In the practical development of the intervention we have used problem-solving models developed by IAPT/Reach out in England. The intervention is currently being evaluated in a randomized controlled trial (pilot) (Phase II). The feasibility of delivering the intervention, and its acceptability are tested on 150 persons with Type 2 diabetes, and the evaluation is based upon a Health technology assessment. HRQOL is an important outcome, and will be a primary outcome in comparative analysis in the EUstudy. RESULTS Study start with inclusion of patients started 01.03.2011, and 133 participants are included in the Norwegian study by now. We have compared our baseline data with a diabetes sample from the general population in Norway, and the findings indicate that our sample is representative according to sociodemographic data. A comparison of HRQOL (SF-36) among the groups, indicated a significantly poorer HRQOL among the participants in the present study. CONCLUSIONS Short-term findings (4 months) will be presented in Budapest.
207: Outcome Measure Development (207.1) The development of an EORTC communication questionnaire: first steps Juan Ignacio Arraras, PhD, Spain; Karin Kuljanic Vlasic, University Hospital Center Rijeka. Croatia; Monika J. Sztankay, Innsbruck Medical University, Austria; Lisa M. Wintner, Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Austria; Anna Costantini, Sant’Andrea Hospital. Sapienza University of Rome. Italy; Wei-Chu Chie, PhD, Inst of Epdeimiology and Preventive Medicine, Nat’l Taiwan Univ, Taiwan; Astrid Helene Liavaag, Sørlandets Sykehus, Norway; Eva Elfriede Greimel, PhD, Medical University Graz, Austria; Anne Bredart, PhD, Institut Curie, France; Maria Arnott, EORTC AIMS To present the steps performed in the development of an EORTC Communication questionnaire specific for cancer patients: a PRO instrument that aims to evaluate patients’ communication with professionals. The questionnaire is based on the Patient Centered Communication Model. Cross cultural aspects have an important role. Communication is a key element of cancer patients’ Quality of Life. METHODS Questionnaire development follows the EORTC Quality of Life Group Module Development Guidelines. Literature review, patients’ and professionals’ interviews have been performed. Literature review. Literature searches carried out in PUBMED and PSYCHINFO data bases 2003–2010. Questionnaire data bases PROQOLID, BIBLIOPRO and the FACIT system, several monographs on Communication, and the EORTC item bank were reviewed. A list of issues (about 400) was created and organized in areas. The possible areas (and issues) to include in the questionnaire were selected based on the Patient Centred Communication Model. Interviews: three steps
123
Qual Life Res (2012) 21:1–132 1. A semi structured interview with 23 patients (Austria Croatia, Italy, Spain): a general question about communication with professionals, and questions about the main areas that were considered after the literature review. 2. A variety of 23 professionals from Austria, Croatia, Norway, Spain, and Taiwan have evaluated the list of issues: its relevance and importance, and possible new issues. 3. A second group of 18 patients (Austria, Croatia, Italy, and Spain) have evaluated the list of issues in the same way. Decision rules to keep issues were agreed. RESULTS A provisional list of issues has been created and organized in provisional scales: (1) Patient’s active role—behaviors; (2) Enhancement of clinician—patient relationship: (2A aspects of the relationship): therapeutic alliance, interaction, time, professional’s abilities creating a relationship; (2B skills): verbal and non verbal language; (3) Management of patient’s emotions: evaluates, empathy, and replies; (4) Attention to patient’s information preferences and needs; (5) Information; (6) Satisfaction; (7) the environment CONCLUSIONS EORTC methodology has helped to select key communication areas. The literature review and the interviews will increase the questionnaire validity. Next steps: the list of issues will be converted into questions. A provisional module will be constructed, translated and pilot tested in an international study.
(207.2) Assessing validity without a conceptual model? The Neck Disability Index as an example Luc Ailliet, VU Medical Center, Department of Epidemiology and Biostatistics; Sidney M. Rubinstein, VU Medical Center, Department of Epidemiology and Biostatistics; Dirk L. Knol, Department of Epidemiology and Biostatistics, VU University medical center; Henrica C.W. de Vet, Sr., PhD, EMGO Institute VU Medical Center, Netherlands; Maurits W. Van Tulder, Department of Health Sciences, Faculty of Earth and Life Sciences, VU University; Caroline B. Terwee, PhD, VU University Medical Center, Department of Epidemiology and Biostatistics, Netherlands AIMS The Neck Disability Index (NDI) is the most commonly used instrument for measuring outcomes in patients with neck pain. The NDI aims to measure ‘disability’ according to the developer. Most researchers use the instrument as a measure of physical functioning, but the NDI seems to measure more than functioning. The aim of this study was to assess content validity of the NDI. METHODS Four requirements for good content validity were evaluated: (1) to assess if all items refer to relevant aspects of the construct the literature was searched on the development of the NDI. The developer of the NDI was contacted to clarify the construct. In addition, 11 neck pain experts were invited to comment on the construct of the NDI; (2) to assess if all items are relevant for the study population 10 neck pain patients commented on the relevance of the items; (3) we considered whether all items were relevant for evaluative purpose; (4) to assess if all items together comprehensively reflect the construct the experts and patients were asked whether important items were missing. RESULTS (1) The concept ‘disability’ was poorly defined by the developer of the NDI as ‘‘the perceived effect of pain and impairment on the patient’s performance and enjoyment of activities of daily living’’. Experts indicated that the NDI measures more than the performance of daily activities (e.g. pain, symptoms, mental functioning). Most items were considered relevant for measuring a broader construct, like ‘‘burden of disease’’, although there was doubt about the relevance of some items. (2) Patients with high scores on the NDI considered all items relevant, but patients with low scores considered most items not applicable to them. (3) We considered all items relevant for evaluation. (4) The experts suggested that a number of items seems missing to measure ‘‘burden of disease’’ such as sports,
Qual Life Res (2012) 21:1–132 mobility, mental issues, dizziness, radiation and computer work. The NDI seems to be outdated. CONCLUSIONS It is unclear what the NDI aims to measure. Most researchers use the instrument as a measure of physical functioning, but the NDI measures more than functioning. If the construct is considered broader, then important items are missing. So for both interpretations of the construct the NDI has poor content validity. Given current quality standards for PRO instruments, the NDI seems not good enough (anymore) for use as a primary outcome measure in research. We suggest that this may apply to many PRO instruments developed in the 90’s of the past century.
(207.3) Development and evaluation of environmental factor candidate items for persons with stroke, traumatic brain injury and spinal cord injury Allen W. Heinemann, Northwestern University, United States; Jin-Shei Lai, PhD, Northwestern University; Noelle Carlozzi, University of Michigan; Joy Hammel, University of Illinois; Sofia Garcia, PhD, Northwestern University; David Gray, Washington University; Elizabeth Hahn, MA, Northwestern University; Susan Magasi, Northwestern University, United States; David Tulsky, PhD, University of Michigan; Allan Kozlowski, Rehabilitation Institute of Chicago AIMS People living with disabilities have long acknowledged the critical influence of physical, cultural, and economic aspects of the environment on their social health and participation. Measuring participation is complicated by multiple environmental factors including physical, social, and attitudinal issues. Additional work is needed to refine the construct of participation, as well as the ways that the environment can impede or enable it. A critical step in this process is operationalizing environmental factors that create barriers or facilitate participation. The objectives of this presentation are to: (1) describe an approach for operationalizing environmental factors; (2) describe the results of a study to develop measures of environmental factors; and (3) to define an initial item bank of environmental facilitators and barriers to participation following stroke, traumatic brain injury and spinal cord injury. METHODS Instrument development included an extensive literature review; focus groups; cognitive testing; item classification, selection and writing. We conducted 20 focus groups with over 200 participants with diverse disabilities, age, race/ethnicity, and socioeconomic status. We used a semi-structured guide and provided captioning that produced a transcript. Fifteen individuals with our target conditions, traumatic brain injury, spinal cord injury and stroke, participated in cognitive interviews. The main outcomes were environmental factor item banks categorized in domains of: assistive technology; built and natural environment; social environment; services, systems, and policies; access to information and technology; and economic quality of life. RESULTS We binned 2,273 items, winnowed this pool to 392 items for cognitive testing, and recommended 265 items for testing with a larger sample. Five domains correspond to the International Classification of Functioning (access to information and technology; assistive technology; built and natural environment; system, services and policies; social environment). We decided it was important to define an economic and financial domain given the importance of economic resources on participation. CONCLUSIONS Five of the six domains correspond closely to the ICF taxonomy; the sixth, economic quality of life, reflects an important construct that reflects the resources that affect participation. Testing with a larger sample is underway to estimate item parameters using item response theory and to evaluate reliability, validity and sensitivity.
27 (207.4) Comparing transition and static measures of patient enablement instruments (PEI) Wong Wendy, PhD, The University of Hong Kong, Hong Kong; Cindy L.K. Lam, MD, The University of Hong Kong, Peoples Republic of China AIMS Background: Patients-reported outcomes relied on the objective events experienced by respondents. The transition or static measures have been the most common methods used in these instruments. It is unknown which methods are the best to capture patient enablement abilities and have never been explored. By using the validated Patient enablement instrument (PEI), the transition measure was tested against static measure on the same sample to determine which was valid for use in the measurement of patient enablement. Aims: To compare the transition and static measures of PEI. METHODS 1,062 patients who visited for a new complaint were recruited and completed the baseline assessment of PEI and 841 of whom completed the 2-week follow-up assessment on the same instrument. The relation between differences in score on baseline (pre-score), 2-week follow-up (post-score), and the global rating of change in enablement over the past 2-week were investigated by the Pearson’s correlation and the linear regression of pre and post score. The change in PEI score (Post minus Pre) was defined as static measure while global rating of change was regarded as transition measure. Validity of transition measure was supported if (1) the correlations (2) between static and the transition measures were at least 0.5; and (2) significant regression coefficients were found in linear regression in which transition measure was dependent variable. RESULTS 70.7, 3.3 and 26 % of patients reported their transition measures to which they had improved, deteriorated and unchanged respectively, Positive correlations between pre-score, post-score, the transition and static measure were identified. Correlation between the transition and static measures (P = 0.121, P \ 0.001) was higher than that between transition and post-score (P = 0.094, P = 0.007), but lower than that between transition and pre-score (P = 0.225, P \ 0.001). In the three regressions, post-score was highly significant but pre-score and static measure did not explain a significant portion of the variance in transition measure (global rating). CONCLUSIONS Low extent of correlations between static and transition measures concluded that global rating of change may not be a good indicator of transition measure. Caution should be paid on the validity of transition measure in the interpretability and establishment of minimum clinically important differences for PEI.
(207.5) Evaluation of a more comprehensive survey item bank for standardizing disease-specific impact comparisons across chronic conditions John E. Ware, Jr., PhD, Univ of MA Medical School; Rick Guyer, JWRG Incorporated; Magdalena Harrington, JWRG Incorporated; Mikel Strom, JWRG Incorporated; Rosemarie Boulanger, JWRG Incorporated AIMS A new QOL Disease Impact Scale (QDIS) was developed to fill the gap between widely-used specific and generic patient-reported outcomes (PRO). Preliminary evaluations of QDIS psychometric properties and empirical validity are reported. METHODS A new standardized bank of 49 disease-specific impact items, differing only in attribution to one condition, was constructed to represent 12 health domains. Data (N = 10,624) were analyzed from Internet-based surveys completed by probability samples of chronically-ill adults representing the general US population, with oversampling of preidentified arthritis, cardiovascular, chronic kidney, diabetes, and
123
28
Qual Life Res (2012) 21:1–132
respiratory conditions. For relative validity (RV) comparisons with QDIS, respondents completed widely-used specific and generic measures at 6 and 9-month intervals after baseline. Analyses included baseline factor analysis and item response theory (IRT) modeling of QDIS and correlational and ANOVA comparisons of RV across severity groups and for predictions of generic health outcomes. RESULTS Analyses confirmed a unidimensional QDIS bank and cross-disease item parameter standardization; tests of model fit of standardized IRT item parameters (adjusted for mean differences) within conditions were not significant (Chi-square = 3,241–7,054, df = 7,595). Percentages of 10-item QDIS scores with reliabilities [0.90 were consistently (across conditions) very high (91–100 %) for moderate-severe groups. Correlations between QDIS and legacy condition-specific measures were high (r = 0.62–0.83 across conditions). Specific measures (QDIS and legacy) were consistently more valid (median RV = 0.82) in discriminating across severity groups in comparison with generic measures (median RV = 0.12). Despite being much shorter, QDIS consistently discriminated well (RV = 0.73–0.97) across mild, moderate and severe groups and as well or better than corresponding legacy tools (RV = 0.52–0.82) and predicted generic physical outcomes substantially (r = 0.41–0.70), sometimes better than legacy specific measures; the reverse pattern was not observed. A similar pattern was observed for QDIS in predicting generic emotional outcomes, with exceptions involving some superior specific mental health measures. CONCLUSIONS QDIS metrics standardized across conditions appear to be sound psychometrically and valid. Combined with norm-based scoring, an integrated ‘‘dashboard’’ making interpretations of both specific and generic health outcomes easier is now feasible. Tests of the usefulness of such combined information in clinical research and practice are warranted.
agreement levels among the clinicians. RESULTS The response rates for Surveys 1 and 2 were 78 and 67 %, respectively. Seventy-nine CTCAE terms remained after survey one. After survey 2, 63 CTCAE terms were selected as being subjective in nature, relevant for measuring in pediatric cancer trials, and amenable to self-report. Most of the CTCAE terms were removed because they relied on laboratory or clinical measures (e.g., high cholesterol, myelodysplastic syndrome) or because they were not applicable for children (e.g., premature menopause). CONCLUSIONS These 63 CTCAE terms will next be translated into child friendly terms (e.g., from ‘‘fatigue’’ to ‘‘tired’’). A questionnaire will be designed and evaluated using qualitative and quantitative methods to ensure validity in capturing the child’s experience. Ultimately, systematic collection of these data will improve care by enhancing the quality and relevance of the available data regarding the toxicities of treatments for childhood cancer.
208: Health Outcomes in Children
AIMS Cleft lip and/or palate (CLP) is the most common congenital craniofacial anomaly. Children with CLP have unique concerns that might not be measured with generic patient-reported outcome (PRO) instruments. We are developing a PRO instrument, the CLEFT-Q, beginning with qualitative interviews of children with CLP. One method to establish content validity and to enable comparison between instruments is to classify what patients describe in the qualitative interviews, and directly compare this to the content of other instruments. The ICF-CY is a classification system based in a broad psychosocial framework that allows for the coding of instruments and qualitative data using a common terminology. The aim of this study was to compare and contrast the content from qualitative interviews informing the CLEFT-Q to content from three generic instruments previously used with this population (Peds-QL 4.0, CHQ, KINDL-R) using the ICF-CY classification. METHODS We performed 53 interviews with children with CLP aged 6–22 years in Canada, the UK, and the USA. Data were coded using a line-by-line approach and the constant comparison method. The codes were then linked to the ICF-CY classification by one rater and verified by a second. Content from the generic instruments was also linked to the ICF-CY by two raters. Any disagreements were resolved with a third rater. The ICF-CY categories from the interviews and the instruments were then compared to allow for a visual depiction of concepts important to patients with CLP compared to concepts measured by the generic instruments. RESULTS The qualitative data were linked to 60 ICF-CY categories. The body functions component was dominant in the qualitative data and the KINDL, whereas activity and participation was dominant in the Peds-QL 4.0 and CHQ. Patients with CLP identified several concepts that were not included in the content of the generic instruments, and several concepts measured in the generic instruments were not identified in the qualitative data. CONCLUSIONS While some health concepts described by patients with CLP overlap with concepts measured in three commonly used generic QOL instruments, we identified some unique health concepts through
(208.1) Identification of subjective adverse event terms to inform the development of a child self-report measure for use in oncology trials Bryce B. Reeve, Ph.D., University of North Carolina at Chapel Hill, United States; Jessica C. Lyons, MS, University of North Carolina at Chapel Hill; Catriona Mowbray, PhD, BSN, RN, CPN, Children’s National Medical Center; Justin Baker, St. Jude Children’s Research Hospital; David Freyer, Children’s Hospital Los Angeles; Stuart Gold, University of North Carolina at Chapel Hill School of Medicine; Mary C. Hooke, PhD, RN, PCNS, CPON, University of Minnesota School of Nursing; Steven Joffee, MD, MPH, Dana-Farber Cancer Institute/Boston Children’s Hospital; Janice Withycombe, Palmetto Health Children’s Hospital; Pamela Hinds, PhD RN FAAN, Children’s National Medical Center AIMS Over 60 % of children diagnosed with a cancer will participate in a clinical trial. Collection of Adverse Event (AE) data in trials is required by law and the current standard is that AEs are reported only by clinicians even when many of the AEs are subjective in nature (e.g., fatigue, anxiety, pain). Our overarching goal is to design a child self-report measure of subjective AEs that will inform the reporting of AEs for the U.S. National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE), the standard lexicon for grading AEs in oncology trials world-wide. The aim of this study is to review the CTCAE to identify which AEs are amenable to self-report by children ages 7 or older. METHODS The CTCAE (v.4) contains 790 AEs described in medical terms. We sought consensus among 184 experienced pediatric oncologists and nurses, from 7 diverse Children’s Oncology Group treatment centers. Two online surveys were used to identify the subset of AEs that are subjective in nature and therefore amenable to self-report by children or by caregiver proxies. Analyses of the survey data were based on reviewing
123
(208.2) Content comparison of the qualitative data informing the CLEFT-Q to generic instruments (Peds-QL 4.0, CHQ, KINDL) using the ICF-CY classification Karen W.Y. Wong, MD MSc FRCSC, Hospital for Sick Children, Canada; Nora Fayed, University Health Network, Canada; Christopher R. Forrest, MD, The Hospital For Sick Children, Canada; Tim Goodacre, Oxford Radcliffe Hospitals, Oxford, UK; Andrea Pusic, MD MHS FRCSC, Memorial Sloan Kettering Cancer Ctr; Arun Gosain, Rainbow Babies and Children’s Hospital, Cleveland, USA; Olaf Kraus de Camargo, McMaster University; Stefan Cano, BSc PhD, Peninsula College of Medicine & Dentristry, United Kingdom; Anne Klassen, D Phil, McMaster University, Canada
Qual Life Res (2012) 21:1–132 qualitative interviewing. Our findings support the need for a condition-specific instrument.
(208.3) Simultaneous cross-cultural development of a condition-specific instrument to assess quality of life (QOL) in short stature youth Monika Bullinger, PhD, Inst for Medical Psychology, Hamburg Univ, Germany; Julia Hannah Quitmann, University Medical Center Hamburg-Eppendorf, Germany; Michael Power, PhD, Edinburgh University, United Kingdom; John E. Chaplin, PhD AFBPS, University of Gothenburg; Emmanuelle Mimoun, University of Toulouse; Michael Herdmann, IMIM University, Insight Consulting and Research; Hartmut Wollmann, Pfizer Ltd.; Andreas Pleil, PhD, Pfizer, Inc. AIMS A simultaneous approach to QoL assessment is expected to facilitate cross-cultural convergence in instrument development and to allow generalization of HRQoL results across countries and languages. The Quality of Life in Short Stature Youth (QOLISSY) project applied this approach to construct a condition-specific instrument for pediatric short stature patients and their parents in 5 European countries. METHODS Focus groups were conducted in each country to identify relevant items within a conceptual model. Subsequent cognitive debriefing and pilot testing yielded basic information on the preliminary questionnaire to be included in a field test to examine its reliability and validity. Cross-cultural aspects were examined in derivation of items from qualitative material (content analysis) as well as in quantitative analysis of psychometric properties within and across countries (differential item functioning, DIF). RESULTS 194 subjects participated in focus groups which collected over 5,000 statements. The few cultural differences identified related to descriptions of specific daily living activities. Statements were reduced and formulated into 124 items to be included for cognitive debriefing and pilot testing in 212 families. Cognitive debriefing results showed few countryrelated differences. Preliminary psychometric testing allowed crossculturally poorly performing items to be discarded. The field test in 336 families, 162 in re-test conducted overall countries combined showed few DIF effects. Reliabilities and validity indicators of the QOLISSY in single countries complied with standards. The construct validity for the total sample was replicable across countries, with a three-domain core QOL structure consisting of 22 items providing the best fit. CONCLUSIONS The cross-cultural simultaneous development of a condition specific tool proved possible. Recruitment challenges resulted in varying sample sizes across countries. Qualitative as well as quantitative methods to analyse cultural variability were helpful in composing a cross-culturally applicable instrument, which is now translated and validated in more cultures/languages.
(208.4) A life course approach to measurement of global health Christopher B. Forrest, MD, PhD, University of Pennsylvania, School of Medicine, United States; Katherine Bevans, PhD, Children’s Hospital of Philadelphia; Ramya Pratiwadi, Children’s Hospital of Philadelphia; JeanHee Moon, Children’s Hospital of Philadelphia; Jin-Shei Lai, PhD, Northwestern University; Carole A. Tucker, PhD, Temple University AIMS Research and health surveillance often require inclusion of both children and adults. Because self-reported health status measures are typically created separately for pediatric and adult populations, it is not possible to evaluate health and the impact of healthcare across the life course. Our aim was to develop a pediatric version of the PROMIS Global Health measure that could be linked with the parallel adult
29 version (Hays et al., 2009) to create a single, life course metric. METHODS A mixed-methods approach was used to design, develop, evaluate, and calibrate the pediatric measure. The concept of global health for children was explored in expert, parent, and child interviews and a literature review; items were tested for comprehensibility in cognitive interviews and a Lexile analysis; and, items were evaluated using factor and item response theory analyses. Data were obtained from 3,635 children and youth 8–17 years-old and 1,608 parents of 5–17 year-old children that participated in a national internal panel survey. RESULTS A pool of 18 developmentally appropriate items was developed for children, 6 of which were obtained from the adult measure. Psychometric analyses led to deletion of 9 items (retention of 4 of the parallel adult items). The 9-item measure was determined to be a unidimensional construct that met IRT assumptions across all ages. Its test–retest reliability ranged from 0.87 to 0.89 across age groups and was comparable between children and parents. We are now evaluating correlations of the new measure with existing (‘‘legacy’’) measures of pediatric global health, and will be linking the pediatric and adult global health measures. CONCLUSIONS This study provides support for the unidimensionality and reliability of the PROMIS Pediatric Global Health 9-item measure. After we have linked the pediatric and adult measures, we will have a common global health measure for assessment of children and across the life course.
Saturday, 27 October 2012 301: Outcome Measure Development (301.1) Numeracy and valuations in quality of life assessment (QOLA) Ivan Barofsky, PhD, The Quality of Life Institute AIMS The defining characteristic of a QOLA is the active valuation (e.g., utility estimates, preferences, and visual analog scales) of some state (e.g., depression, a physical limitation). This principle is as valid when valuing a sensory experience (e.g., a touch), as it is when valuing a life lived. A critical determinant of the quantification of these valuations is the person’s ability to generate and apply numbers. This skill referred to as numeracy—is seldom regularly assessed during a QOLA, yet as many as 22 % of Americans have been shown to lack numeracy skills. This paper will review the published literature (e.g., Reyna et al.; Psychol. Bull, 2009) to illustrate the role that numeracy plays in a variety of assessment formats (e.g., preferences, Likert Scales, visual analog scales) critical to determining the outcome of a qualitative assessment. Available numeracy assessment instruments, as well as empirical outcomes will also be reviewed. METHODS PubMed, PsycINFO and selected texts were reviewed to identify papers which consider numeracy issues directly or indirectly related to QOLA. RESULTS The literature review revealed at least 17 generic and/or disease-specific numeracy assessments. These instruments consisted of 2–68 items with the items consisting of both reading comprehension tasks and/or questions that required manipulating mathematical operations. Studies (e.g., Woloshin et al.; Med Dec Making; 21, 2001) were reported where valuation exercises using the standard gamble or timetrade-off were assessed for persons who differed in levels of numeracy skill (e.g., Schwartz et al., 2001). What was found was that persons with low numeracy showed a negative correlation suggesting that they had difficulty with the quantitative estimation involved in utility estimation, while those with a high numeracy scores showed a positive correlation. Data of this sort may account for some of the observed variability and lack of comparability between various utility estimation methods. CONCLUSIONS Numeracy skills are potentially a major determinant of QOLA outcome and its systematic assessment should be a regular part of any scaled assessments that involves valuations, if not psychosocial assessments, in general.
123
30 (301.2) A systematic review of the use and validation of health-related quality of life instruments in patients with cancer cachexia Sally J. Wheelwright, University of Southampton, United Kingdom; Anne-Sophie Darlington, PhD, University of Southampton School of Medicine, United Kingdom; Jane B. Hopkinson, University of Cardiff; Deborah Fitzsimmons, Swansea University, United Kingdom; Alice White, University of Southampton; Colin Johnson, Mchir, Dept of Surgery, United Kingdom AIMS Cachexia is a multidimensional syndrome characterised by involuntary weight loss, muscle atrophy and physiological changes that leads to progressive functional impairment. More than 80 % of cancer patients develop cachexia before death. The aims of this systematic review were to (1) identify cachexia-specific HRQOL instruments, (2) identify studies which measure HRQOL in cachectic cancer patients and appraise HRQOL reporting, (3) describe the relationship between weight loss and QoL. METHODS Following the Centre for Reviews and Dissemination guidance for undertaking systematic reviews, relevant databases were searched from January 1980 to January 2012 using terms relating to cancer, cachexia and HRQOL for papers including adult cancer patients with cachexia or documented weight loss at baseline. HRQOL development was evaluated using the COSMIN checklist. A modified version of the Efficace minimum data checklist assessed reporting of HRQOL tool use. RESULTS (1) We found one cachexia-specific HRQOL instrument, the Functional Assessment of Anorexia/Cachexia Therapy (FAACT). Using the COSMIN, internal consistency, reliability, measurement error and structural validity were all rated as poor. Criterion validity and responsiveness were rated as fair, whilst hypothesis testing was good and content validity was excellent. The FAACT has not been crossculturally validated. (2) 76 studies assessed HRQOL in cachectic or weight-losing cancer patients. Just 17 % of studies used the FAACT: most used generic cancer QoL instruments, the most common of which was the EORTC QLQ-C30 (46 % of studies). Although the average total score on the modified Efficace checklist was relatively high, 7 out of 10, more than half the studies failed to provide a rationale for instrument selection, information about instrument administration or missing data documentation. 3) 29 studies directly examined the relationship between QoL and weight loss, with 23 finding a broadly negative relationship between the two. CONCLUSIONS This review demonstrates a relationship between weight loss and HRQOL, but since most studies use generic HRQOL measures, they offer little insight to researchers or guidance to clinicians. The development of the only cachexia specific HRQOL instrument is associated with several weaknesses. There is a pressing need for a well-designed HRQOL tool for use with this patient group.
(301.3) International validation of the WHOQOL-SRPB: a study of older adults Suzanne Skevington, PhD, University of Bath, United Kingdom; Juliana Moratelli, United Kingdom; Laura Smith, University of Bath; Ramona Lucas, PhD MD MPH, University of Barcelona, Psychology School, Spain; Marcelo P. Fleck, PhD, Hospital de Clinicas de Porto Alegre, Brazil; Tal Narkiss, Ben-Gurion University of the Negev; Laura Schwartzmann, MD, Interdisciplinary Ctr of Childhood & Poverty, Uruguay AIMS Developed through an international collaboration, the WHOQOL-SRPB (2006) is a multilingual instrument for use in health care that is designed to assess spiritual, religious and personal beliefs alongside other quality of life (QoL) domains. It demonstrates good preliminary psychometric properties, but further international evidence of validity and sensitivity is needed. The aim was to examine these
123
Qual Life Res (2012) 21:1–132 properties cross-culturally. METHODS Older adults (n = 1,321) (60+ years) living in five countries (Brazil, Israel, Spain, UK and Uruguay) were recruited in WHOQOL-OLD centres where WHOQOL-SRPB language versions were also available. Participants completed the WHOQOL-BREF with the SRPB module from the WHOQOL-SRPB, and OLD module from the WHOQOL-OLD; also the SF-12, socio-demographic and health variables. RESULTS The global dataset contained 64 % women and 36 % men. Forty-three % were 60–69 years, and 54 % 70+. Self-reports showed that 23 % were unhealthy and 78 % healthy. WHOQOL-BREF domain means ranged from acceptable physical QoL (58.9) to good environmental QoL (67.1). For SRPB items the best QoL was related to awe and wonder and the poorest to spiritual connection. For the OLD items, lack of sensory functioning reflected the poorest QoL. Excellent discriminant validity was found for the SRPB domain (F = 51.8; p = .0001) and the WHOQOL-BREF domains (physical, psychological, social, environment) (p = .001) when overall QoL ratings from very poor to very good QoL were compared across the continuum. The ‘peaceful’ item from the SF-12 correlated modestly with the SRPB domain (r = -.22; p = .0001), but predictably, correlated more highly with peace (r = -.42) than the other seven SRPB facets. Construct validity of the SRPB module was largely supported through mapped inter-item/domain correlations, and in relation to WHOQOL-BREF dimensions. Internal consistency reliability of the SRPB domain was excellent (alpha = 0.96); for countries alpha ranged from 0.96 to 0.93. The overall alpha for the WHOQOL-BREF was excellent (0.91), supporting individual use. Differences between cultures, languages, world regions, religious beliefs and other variables will be presented. Similarities are examined through multi-level modeling, controlling for country differences. CONCLUSIONS We present the first cross-cultural validity and sensitivity evidence for the WHOQOL-SRPB from an older adult sample which should improve confidence in using this instrument.
(301.4) Qualitative development of the PROMISÒ pediatric subjective well-being (SWB) item banks Janine Devine, University Medical Center Hamburg-Eppendorf, Germany; Katherine Bevans, PhD, Children’s Hospital of Philadelphia; Anne Riley, Johns Hopkins University, United States; JeanHee Moon, Children’s Hospital of Philadelphia; John M. Salsman, PhD, Northwestern University; Christopher B. Forrest, MD, PhD, University of Pennsylvania, School of Medicine, United States; Ulrike Ravens-Sieberer, University Medical Center of Hamburg AIMS The aims of this pediatric Patient Reported Outcome Measurement Information System (PROMIS) study were (1) to conceptualize children’s subjective well-being (SWB), and (2) to produce item pools with excellent content validity and ready for calibration and use in computerized adaptive testing (CAT) applications.; METHODS The content and manifestations of pediatric SWB were defined through semi-structured interviews with experts, children 8–17 years-old, parents, and a systematic literature review. Candidate items were retrieved through the literature, classified, revised, and newly generated to cover the whole spectrum of pediatric SWB. The item pools were evaluated for comprehensibility using cognitive interviews (CI), reading level analysis, and translatability review.; RESULTS Interviews supported children’s SWB as having both hedonic (positive affect, PA) and global components (life satisfaction, LS). Input from experts, children, parents, and the literature indicated the need to measure the eudaimonic dimension (meaning and purpose, MP) in life of children. Item pools for life satisfaction (56 items), positive affect (53 items), and meaning and purpose of life (55 items) were developed, and small differences in comprehensibility of some items were observed between children and adolescents. The final item
Qual Life Res (2012) 21:1–132 banks have excellent content validity and are ready for calibration and use in computerized adaptive testing (CAT).; CONCLUSIONS The conceptualization of children’s SWB is similar to adults; however, the specific manifestations of positive affect, life satisfaction, and meaning and purpose vary across the lifespan. These findings suggest the need for developmentally appropriate measures for children. 302: Personalized Medicine (302.1) Are condition-specific measures necessary to the practice of personalized medicine? San Keller, PhD, American Institutes for Research; Manshu Yang, American Institutes for Research; Christian Evensen, American Institutes for Research; Marsha Treadwell, Children’s Hospital and Research Center at Oakland AIMS The Patient-Reported Outcomes Measurement Information System (PROMIS) aims to provide valid health measurement across chronic conditions. The current study addresses the question of whether PROMIS requires supplementation to be sensitive to specific patient populations. METHODS We conducted formative research with 122 adults living with Sickle Cell Disease (SCD) (in-depth, personal interviews and focus groups) and 15 providers (in-depth, personal interviews) to gather information about how SCD affects adult functioning and wellbeing. Diversity in age, gender and geographic location of adults with SCD was achieved, and, consistent with US epidemiology, most were African American. Data were content analyzed and the content was compared to existing PROMIS measures. Beta-test data on 8 Adult Sickle Cell Quality-of-Life Measurement Information System (ASCQ-Me) short-form (SF) and 6 computer-adaptive (CAT) measures as well as 24 PROMIS measures (10 globals, 11 SFs and 3 CATs) were collected from 490 adults recruited at 7 geographically-diverse US sites. Univariate and multivariate analyses were conducted to determine the sensitivity of measures to SCD severity. RESULTS Qualitative analyses of patient interviews revealed content on functioning and wellbeing not included in PROMIS, which formed the basis of ASCQ-Me questions. As expected, SCD patients were less healthy than the general population across all PROMIS measures. Patients were divided into SCD-severity tertiles based on a medical-history checklist. In univariate analyses, all but one of the ASCQ-Me and two of the PROMIS measures significantly discriminated among the tertiles. Multivariate regression supported the use of the PROMIS physical functioning and pain CATs in patients with SCD. However, ASCQ-Me emotional impact and sleep impact SFs described significant-unique variance in disease severity net of PROMIS. Moreover, the ASCQ-Me pain-episode severity SF accounted for significant-unique variance in SCD disease severity whereas the PROMIS pain impact and behavior SFs did not. CONCLUSIONS PROMIS performed well in this patient population. Nevertheless, qualitative and quantitative data supports the addition of condition-specific content. In SCD, unique PRO content emerged from interviews with patients and providers and some SCD-specific SFs or CATs explained statistically-unique variance in disease severity net of comparable PROMIS measures—unique information on pain, emotional and sleep impact for treatment considerations.
(302.2) Relationship between genetic markers and quality of life (QOL) in stage III colon cancer (CC) patients (pts) prior to adjuvant treatment (N0147) Jeff Sloan, PhD, Mayo Clinic, United States; Qian Shi, Mayo Clinic; Adam Lee, Mayo Clinic; Robert Diasio, Mayo Clinic; Emily Pavey, Mayo Clinic; Dan Sargent, Mayo Clinic; Richard Goldberg, Ohio State; Michelle Mahoney, Mayo Clinic; Steven Alberts, Mayo Clinic
31 AIMS In metastatic colorectal cancer (CRC), previous studies suggested potential relationships between certain drug metabolizing (DM) genes (DPYD, ERCC2, GSTP1 and TYMS) and QOL prior to receipt of chemotherapy (Sloan, Plenary Session ASCO 2004). The current study is the first to examine relationships between similar DM genes and QOL in stage III CC METHODS 1,583 CC pts on a randomized adjuvant stage III, phase III trial received FOLFOX, FOLFIRI, or either ± Cetuximab. Genomic DNA and baseline QOL data via linear analogue self assessment (LASA) scales with 4 QOL variables were obtained. 53 DM SNPs were selected for known functional effect within potential regulatory sites. Two-sample t-tests assessed differences in QOL between pts with variants and wild-type status. Differences of at least 10 points on a 0–100 point QOL score were considered clinically significant. RESULTS 14 relationships between DM genes and QOL were detected at the nominal p \ 0.05 level; six of these were related to DPYD, ERCC2, GSTP1 and TYMS. Relationships were also observed at the 0.05 level with polymorphisms of genes DPYS, ERCC1, MGMT, MTHFR, OPRT and UGT1A1 based on unadjusted and adjusted association tests. GSTP1 I105V and OPRT 5’UTR 28 A[G markers differed on fatigue scores (p = 0.04, 0.005, respectively). Variants on MTHFR R594Q were associated with higher mental acuity scores (p = 0.02). These relationships remained after adjusting for age, gender, race, T stage, PS, disease site, lymph nodes involved, and grade. The robustness of the observed associations varied with the multiple comparison adjustment method. CONCLUSIONS Similar to the earlier study of stage IV CRC, we observed a possible relationship in pts with Stage III CC with regard to DM genes and QOL, although these relationships were more modest than observed in Stage IV pts. These DM genes may be in linkage disequilibrium with other genetic variants that more directly affect behavior and QOL. Further genome wide association studies are needed to clarify these relationships.
(302.3) Screening for psychosocial problems in onco-genetics: development and evaluation of a screening checklist Willem Eijzenga, MSc., NKI/AVL; Neil K. Aaronson, PhD, The Netherlands Cancer Institute, Netherlands; Daniela E.E. Hahn, NKI-AVL; Irma Kluijt, Amsterdam Medical Center, Netherlands; Margreet G.E.M. Ausems, University Medical Center Utrecht; Grace N. Sidharta, NKI-AVL; Eveline M.A. Bleiker, NKI-AVL AIMS Approximately 20 % of individuals who undergo genetic counseling and testing for cancer experience clinically significant levels of psychosocial problems. Such problems, however, remain undetected in approximately 30 % of cases. The purpose of this study is to develop and evaluate a checklist for identifying psychosocial problems experienced by individuals undergoing genetic counseling for cancer. METHODS The four phases of the EORTC Quality of Life Group guidelines for questionnaire development were followed to develop the checklist. To examine the screening properties of the checklist, 127 individuals undergoing cancer genetic counseling completed the checklist and were interviewed by an experienced psychosocial worker (‘‘gold standard’’). The screening properties of the checklist were examined in terms of sensitivity, specificity, and positive predictive value. A randomized, controlled trial is being conducted to evaluate the value of the checklist in daily clinical genetics practice. RESULTS The checklist comprises 26 items organized into six problem-domains. An extra question per domain assesses the perceived need for specialized psychosocial services. The sensitivity of the checklist in detecting psychosocial problems warranting professional intervention ranged from 0.35 to 0.91, the positive predictive value ranged from 0.57 to 0.87, and the specificity ranged from 0.38 to 0.73. Preliminary data from the first 138 participants in the trial indicate that 83 % experienced psychosocial problems related to living with cancer, 51 % related to children, 38 %
123
32 to genetics, 25 % to family and social issues, 19 % to emotions, and 14 % to practical problems. Whether these issues are more frequently detected and addressed in the intervention group as compared to the control group will be evaluated when the study matures. CONCLUSIONS If proven effective, the checklist can aid in facilitating communication, increasing counselors’ awareness, and improving the management of psychosocial problems within cancer genetics.
(302.4) Concept relevance for components of the ENSEMBLE MDS 1.0 Questionnaire Battery Mona L. Martin, RN MPA, Health Research Associates, Inc.; Kelly McCarrier, PhD MPH, Health Research Associates, Inc.; Michael Scanlon, Health Research Associates, Inc; Matthew Wolfe, Health Research Associates, Inc.; Rosirene Paczkowski, Eli Lilly and Company, United States; Don P. Buesching, PhD, Eli Lilly & Company AIMS The ENSEMBLE MDS is a self-administered questionnaire tool designed by Eli Lilly and Company to collect patient-reported phenotypic data in clinical studies to better characterize and define patient subgroups and their likely response to treatment. In order to substantiate the content validity of the ENSEMBLE MDS v1.0 battery, this study used semi-structured qualitative interviews designed to (1) document patients’ understanding of its underlying concepts of general health, social support, stress, anxiety, depression and social status, and (2) identify the meaning of each item included in the battery. METHODS Participants were recruited from internet and newspaper advertisements to participate in qualitative interviews using a semi-structured interview guide. Patient descriptions of the key concepts reported during the interviews were coded using Atlas.ti and patient quotations were evaluated by content matching to determine the relevance of ENSEMBLE MDS v1.0 component measures among the study sample. Any difficulty expressed by patients regarding their ability to interpret each item and formulate an answer was identified. RESULTS Forty-eight participants representing four targeted conditions (Depression, Type 2 diabetes, Psoriasis, Rheumatoid Arthritis) completed qualitative interviews. The study sample was 58 % female, 73 % Caucasian, and 12 % had high school education or less. Participants’ average age was 44. Patient quotes provide qualitative support of concept relevance and good item comprehension for the Self-Reported Health Status scale, Perceived Social Support scale, Psychological Health Questionnaire (PHQ-4), and the Perceived Stress Scale among the target population interviewed. While concept relevance was demonstrated for the remaining component measures, the Subjective Economic Status measure showed substantial variability in the interpretation of the concepts presented, and issues were reported with the response options for the Objective Economic Status component and the EQ-5D that would affect the interpretability of the data. CONCLUSIONS This study provides early general evidence of content validity for the ENSEMBLE MDS v1.0, although difficulties with some scales suggests revisions to the battery may be needed prior to successful use in identifying specific subgroups of treatment responders in clinical studies.
303: Response Shift (303.1) Impact of response shift on time to quality of life scores deterioration in breast cancer patients: is it time to move for QoL RECIST criterion? Zeinab Hamidou, Centre Georges Francois Leclerc, France; Tienhan Sandrine Dabakuyo, PharmD PhD, Centre Georges Francois Leclerc, France; Francis Guillemin, MD, PhD, Ecole de sante publique—
123
Qual Life Res (2012) 21:1–132 Lorraine University, France; Thierry Conroy, Centre Alexis Vautrin, France; Michel Velten, Epidemiology and public health laboratory, College of Medicine, Strasbourg, France; Damien Jolly, University Hospital of Reims; Sylvain Causeret, Centre Georges Franc¸ois Leclerc; Jean Cuisenier, Centre Georges Franc¸ois Leclerc; Mariette Mercier, Cellular and molecular biology laboratory, University Hospital of Besanc¸on; Franck Bonnetain, PhD, Centre Georges Francois Leclerc, France AIMS Time to quality of life (QoL) score deterioration (TD) is a method of longitudinal QoL data analysis that has been proposed for breast cancer (BC) patients (Hamidou et al. Oncologist 2011). As for RECIST criteria, the optimal definitions dealing with reference should be explored. This study aims to study the impact of changes in internal standards (CIS) of response-shift (RS) and the influence of baseline QoL expectancies on TD. METHODS A prospective multicenter study including all women hospitalized for a primary BC was conducted. The EORTC-QLQ-C30 and BR23 questionnaires were used to assess the QoL at baseline, at the end of 1st hospitalization, and 3 and 6 months after. CIS was investigated by the then-test method. QoL expectancy was assessed at baseline using Likert scale. Deterioration was defined as a decrease in QoL scores reaching at least the mean difference identified as minimal clinically important difference (MCID) using Jaeschke’s transition question. Sensitivity analyses were done using the then-test score as reference score, and considering 5 and 10 points as MCID. TD was estimated using Kaplan–Meier method. Cox regression analyses were used to identify factors influencing TD. RESULTS From February 2006 to February 2008, 381 women were included. For role functioning dimension, the median TD increased from 3.2 months [95 % CI: 3.1–3.36] to 4.76 months [3.3–6.2] when adjusting on CIS. For body image when adjusting on CIS, sentinel lymph node biopsy became significantly associated with longer TD (HR: 0.64 [0.43–0.94]) as compared to axillary lymph node dissection, radiotherapy to a shorter TD (HR: 0.63 [0.42–0.95] and the type of surgery had no effect on TD. For global health, cognitive and social functioning dimensions, patients expecting deterioration in their QoL had a significantly shorter TD. For fatigue and breast symptom scales, patients expecting no change had a significantly shorter TD, as compared to patients expecting an improvement. Sensitivity analyses using a MDCS of 5 or 10 confirmed these results. CONCLUSIONS Our results suggest that it would be more accurate to take into account CIS component of RS as well as QoL expectancies to estimate TD of QoL scores in patient with BC.
(303.2) Understanding response shift in multiple sclerosis patients: application of Random Forest method Mohamed Boucekine, Self-perceived Health Assessment Research Unit, France; Anderson Loundou, EA3279, Self-perceived Health Assessment Research Unit, Marseille, France; Rima L’Mouaci, Timone University Hospital, France; Patricia Minaya Flores, MA, Laboratoire de Sante Publique EA3279, France; Helmut Butzkueven, MS Unit, Melbourne, Victoria, Australia; Karine Baumstarck, France; Badih Ghattas, De´partement de Mathe´matiques, Faculte´ des Sciences de Luminy, Marseille, France; Pascal Auquier, PhD, Laboratoire de Sante Publique, France AIMS Response shift refers to a change in meaning of a quality of life (QoL) measure due to changes in internal standards (recalibration), values (reprioritization), and/or conceptualization. The aim of this study is to analyze the component of reprioritization using the Random Forest method in a multiple sclerosis (MS) population. METHODS Patients with MS were enrolled into a multicenter,
Qual Life Res (2012) 21:1–132 multiregional, longitudinal observational study. QoL was assessed using the 36-Item Short-Form (SF-36) instrument at baseline and every 6 months thereafter up to month 24 (INC, M6, M12, M18, M24). Patients were defined as worsened or non-worsened based on change in EDSS score from baseline to month 24. Statistical analysis: random forest method (R-RandomForest package) was performed to produce variable importance measures for predictor variables of the general health dimension score of SF-36, the predictor variables were the other dimension scores of SF-36. RESULTS Among the enrolled population, 407 individuals were not worsened and 107 worsened at 24-month. The variable importance measures identified changes in the order of domains (‘psychologicallike’, ‘physical-like’ or ‘social-like’ dimensions) over time in the worsened patients (additional material 1). These changes are different to the changes identified in the non-worsened patients. CONCLUSIONS The Random Forest method appears as an interesting approach of understanding the response shift phenomena. These findings should be compared to the other current statistical techniques.
(303.3) Assessing the measurement invariance of quality of life across age and aging perceptions: a MIMIC modeling approach Bellinda L. King-Kallimanis, PhD, TILDA, Trinity College Dublin, Ireland; Roxanne E. Jensen, PhD, Georgetown University; Rose Anne Kenny, Trinity College Dublin, Ireland AIMS To investigate differences in item responses according to age, aging perceptions, sex and comorbidities that could not be accounted for by differences in Quality of Life (QoL). METHODS Data come from the Irish Longitudinal Study on Ageing (TILDA), which is a nationally representative sample of community dwelling adults aged 50 and older. There were 5,519 respondents included. Using confirmatory factor analysis, a measurement model of the CASP-19 measure of QoL was fit to the data. This was then extended to a multiple indicators, multiple causes (MIMIC) model to investigate whether differences in item responses (bias) were present. Aging perceptions were measured using the subscales of the Aging Perceptions Questionnaire (APQ) and comorbidities were based on the Charlson Comorbidity Index. Bias was detected using modification indices (MI) and expected parameter changes (EPC). RESULTS A two factor, Control/Autonomy and Self-Realization/Pleasure measurement model fit the data well. In the next step the MIMIC model was fit; MIs and EPCs revealed a total of 19 biases associated with 5 of the CASP-19 items. Seven biases were associated with the item ‘‘My age prevents me from doing the things I would like to’’. This indicated that the item is too general as age, the APQ subscales, awareness of one’s age being constant, awareness of one’s age being cyclical, negative impact of aging on one’s life and emotional responses to aging, and number of comorbidities had undue influence on this item. Only APQ subscales had undue influence on the items ‘‘I feel that what happens to me is out of my control’’ and ‘‘I feel left out of things’’. A combination of age, APQ subscales and comorbidities had undue influence on the items ‘‘Shortage of money stops me from doing things I want to’’ and ‘‘My health stops me from doing things I want to’’. CONCLUSIONS This study demonstrated that age, aging perceptions and comorbidities led to biased responses to items on the Control/Autonomy common factor of the CASP-19. As a result care should be taken when investigating age and aging perceptions differences in relation to Control/Autonomy. The Self-Realization/ Pleasure common factor was invariant in relation to the variables investigated.
33 (303.4) Identifying individuals with inflammatory bowel disease who experienced response shift: a latent class analysis Richard Sawatzky, PhD RN, Trinity Western University, Canada; Anne M. Gadermann, PhD, University of British Columbia, Canada; Pamela A. Ratner, PhD RN FCAHS, University of British Columbia, School of Nursing, Canada; Bruno Zumbo, University of British Columbia, Canada; Lisa M. Lix, PhD, School of Public Health, University of Saskatchewan, Canada AIMS People with inflammatory bowel disease (IBD) often experience changes in disease activity that could lead to response shift (RS). Using latent class analysis, we sought to: (a) identify people with IBD who demonstrate RS in SF-36 physical and mental health scores, (b) determine whether response shift is associated with disease activity, and (c) evaluate the extent of measurement bias that may result if RS is ignored. METHODS The Manitoba IBD Cohort Study (N = 389) is a longitudinal prospective study that investigates psychosocial predictors of health outcomes in people with IBD. Health was measured with the SF-36 at two times, 6 months apart. Participants were assigned to one of three disease activity groups based on responses to a validated single-item measure of active symptoms in the previous 6 months: (a) consistently inactive disease (referent), (b) change in disease activity, and (c) consistently active disease. A repeated-measures mixture analysis was used to classify respondents into two latent classes. For the RS class, the model parameters of the SF-36 were allowed to vary across the two observations. The model parameters were held constant for the non-RS class. Potential bias in physical- and mental-health scores was examined by comparing the latent variable scores of the two-class model (adjusted for RS) to those of a one-class model (ignoring RS). RESULTS Relative to the one-class model, significantly improved model fit was obtained when two classes were specified. This suggests evidence of response shift. Interestingly, people in the non-RS latent class (51 % of the sample) were more likely to have reported disease activity during the 6-month period relative to those in the RS class. The likelihood of being in the non-RS latent class was 2.9 and 6.1 greater for the change in disease activity and the active disease groups, respectively, in comparison with the inactive disease group (Nagelkerke R-squared = 154 %). Ignoring RS resulted in substantial bias in latent mental and physical health scores. CONCLUSIONS This study demonstrates how latent class analysis can be used to identify groups of individuals with or without RS, which, in people with IBD, appears to be associated with disease activity.
(303.5) Examining response shift and true change in self-reported physical and mental health status in homeless and vulnerably housed individuals Anne M. Gadermann, PhD, University of British Columbia, Canada; Richard Sawatzky, PhD RN, Trinity Western University, Canada; Anita Palepu, University of British Columbia; Anita Hubley, University of British Columbia, Canada; Bruno Zumbo, University of British Columbia, Canada; Tim Aubry, University of Ottawa, Canada; Stephen Hwang, MD MPH, St. Michael’s Hospital, Canada AIMS Observed changes in longitudinal studies using self-report measures may reflect true change in the construct or may be due to response shift (RS), or a combination of both. RS includes changes in (a) the meaning of the target construct (reconceptualization), (b) the importance of an item (reprioritization), and (c) the respondents’ internal standards of measurement (recalibration). The aims of this study were (a) to evaluate whether homeless or vulnerably housed individuals (HVHIs) experienced RS over a 1 year time period in
123
34 their self-reported physical and mental health status, and (b) to determine the extent to which any found RS resulted in measurement bias. METHODS Data were taken from the Housing and Health in Transition Study. HVHIs were recruited in Vancouver, Toronto, and Ottawa (N = 1,191) using a two-stage sampling method and were reinterviewed 12 months later (N = 968). Structural equation modeling, employing Oort’s approach, was used to examine whether the different forms of RS were present with regard to the SF-12. A probit link function with mean and variance adjusted weighted-least squares estimation was used to accommodate the ordinal and binary distributions of the SF-12 variables. RESULTS A measurement model was established based on the recommended scoring of the SF-12, exploratory factor analyses with our data, and substantive considerations. The overall test of RS was statistically significant. Reconceptualization, reprioritization, and recalibration RS were detected for several of the items. Although there was statistically significant evidence of RS, the bias was relatively small. Effect sizes of the observed change in SF-12 scores when ignoring RS were Cohen‘s d = -.02 for physical health and d = .16 for mental health, whereas the true change, after accommodating for RS, was d = .003 and d = .20, respectively. CONCLUSIONS The results suggest that HVHIs may experience RS in their self-reported health status and that this could lead to underestimated change scores with respect to their mental health status. Considering that some HVHIs experience more RS than others, further research is recommended to identify subgroups of HVHIs who are most likely to experience different forms of RS.
304: Health Outcomes in Children (304.1) Weight status, quality of life and cigarette smoking among adolescents in Washington State Hongmei Wang, PhD, Social Medicine and Family Medicine, Zhejiang University, China, United States; Janice F. Bell, Department of Health Services, University of Washington; Todd C. Edwards, PhD, University of Washington, United States; Yi Huang, Department of Health Services, University of Washington; Donald Patrick, PhD MSPH, Seattle Quality of Life Group AIMS Overweight and obesity among children and adolescents continue to be public health concerns in the United States. Results of studies of the association between weight status and smoking are inconsistent yet studies have included other predictors of smoking as covariates in analyzing this relationship. This study aims to examine age- and sexspecific associations between weight status and intensity of cigarette smoking with and without controls for quality of life (QOL) and weight control behavior in a large sample of adolescents. METHODS We used cross-sectional data from the 2010 Washington State Healthy Youth Survey collected in grades 8, 10 and 12 (n = 11,222). Multinomial logistic regression was used to model cigarette smoking as a function of weight status, weight control behavior and QOL by sex and age. Models were also repeated with and without missing covariates imputed with the mode with no difference in findings. All statistical analyses were conducted in 2011 using Stata, version 10, and a probability of 0.05 was considered statistically significant. All estimates accounted for the complex survey sampling design. RESULTS Overweight/obesity was not associated with smoking. Weight control behavior was associated with smoking among girls only. Trying to stay the same weight relative to having no weight control behavior was associated with lower odds of light smoking versus no smoking for younger girls (RRR = 0.25; 95 % CI = 0.08, 0.84); whereas, trying to lose weight relative to having no weight control behavior was associated with higher odds of light smoking versus no smoking for older girls (RRR = 1.73; 95 % CI = 1.11, 2.70). Low QOL was associated with higher odds of light and frequent smoking compared to no smoking for both girls and boys (p \ 0.001). CONCLUSIONS QOL and weight control behavior are
123
Qual Life Res (2012) 21:1–132 important correlates in analyzing associations of weight status and smoking.
(304.2) Comprehensive assessment of health-related quality of life measures amongst Hungarian children and adolescents living with cancer, diabetes or juvenile rheumatoid arthritis Dalma Hosszu´, Ph.D. Candidate, University of Pe´cs Institute of Psychology; Ba´tor Ta´bor Foundation, Hungary; Szabolcs To¨ro¨k, Semmelweis University, Institute of Mental Health, Budapest, Hungary; Gyo¨ngyi Ko¨ko¨nyei, Eo¨tvo¨s Lora´nd University, Faculty of Education and Psychology, Budapest, Hungary; Andrea Be´ke´si, Ba´tor Ta´bor Foundation, Budapest, Hungary, Hungary; KIDSCREEN Group, University Clinic Hamburg-Eppendorf, Hamburg, Germany AIMS The purpose of our study was to comprehensively assess health-related quality of life, in a sample of Hungarian diabetic, juvenile rheumatoid arthritis and oncology pediatric patients, using self-report and parent-report questionnaires, and to compare the results with the national norm data. METHODS Hungarian children and proxy version of Kidscreen-52 questionnaire was used to assess health-related quality of life (n = 259). Multivariate analysis of variance was used to reveal differences of means of the subscales of the KIDSCREEN-52. For determining the level of agreement between parent proxy reports and child self-reports we used the intraclass correlation coefficient for absolute agreement. Differences between our sample means and the Hungarian population norms were analyzed using one-sample t-tests. RESULTS Differences were found not only considering disease groups, but both age and gender within the selfreports. Although agreement between parents and their children was moderate, parents view on their child’s physical health, psychological well-being, parental relations, peer relations was found worse in comparison with how children evaluate these issues on their own. In the light of the national norm data our participants had higher autonomy scores in self-reports, the same time lower social acceptance scores in proxy reports were found. CONCLUSIONS The merit of this study is not only the comprehensive assessment of quality of life within chronically ill Hungarian children, but that it is also a constituent part of an outcome effectiveness research program of a Hungarian non governmental organization—called Ba´tor Ta´bor Foundation. Ba´tor Ta´bor offers therapeutic recreation camping programs for children with chronic and life threatening illnesses. Considering its methods and its attitude to children and health, it is a high standard and unique initiative in Central Europe. Key words: health-related quality of life; cancer, diabetes, juvenile rheumatoid arthritis, KIDSCREEN-52.
(304.3) Trajectories of overweight and indicators of health-related quality of life and well-being in 5–7 year old children; the ‘Be active, eat right’ study Hein Raat, MD, PhD, MBA, Erasmus MC—Univ Medical Ctr Rotterdam, Netherlands; Amy Van Grieken, MSc, University Medical Center Rotterdam, Netherlands AIMS We evaluated whether the presence of overweight/obesity at age 5 years, and trajectories of weight development between age 5 and 7, affect general health perceptions and indicators of wellbeing at age 7 years. METHODS In the analyses we included 3,912 children/parents from a nation-wide general population sample, a subgroup in the ‘Be active, eat right’ study. At age 5 and 7 years, height and weight were measured; children were classified as underweight/normal-weight, and as overweight/obese (BMJ 2000: 1240–3); 3 trajectories (stable weight, decreasing weight, increasing
Qual Life Res (2012) 21:1–132 weight) were established. At 5 and 7 years, parents completed the general health perceptions scale (CHQ-PF50) and a single-item on child’s happiness; at age 7, 3 items on ‘child feeling insecure because of his/her weight’, ‘child being bullied because of weight’, and ‘parental worries because of child’s weight’. Linear regression models with weight category at 5-years and weight-trajectory as predictors were fitted to estimate outcomes at age 7-years, with gender and ethnic background as confounders; with/without correction for baseline values (5-years). Cohen’s effect sizes ‘d’ were calculated (mean difference/SD). RESULTS 88 % of respondents were mothers (mean age 36 years; SD: 4). Mean child age at baseline 5.8 years (SD:0.4); 50 % boys; 595 had underweight, 2,956 normal weight, 288 overweight, 73 obesity. Weight status and trajectories of weight development had no impact on child’s happiness (p [ 0.05). Compared to normal weight, underweight (d = 0.2), overweight (d = 0.0), and obesity (d = 0.5) at age 5 were associated with lower general health at age 7 (p \ 0.01); associations disappeared after including baseline CHQ-PF-rating (p [ 0.05). Overweight (d = 0.8/ 0.7) and obesity (d = 1.3/1.3) at age 5 had a negative impact on ‘feeling insecure’/’being bullied’ at age 7 (p \ 0.01); impact on parental worries was even higher (p \ 0.01). Increasing weight between age 5 and 7 was associated with ‘feeling insecure’/’being bullied’/’parental worries’ in the fully corrected models compared to stable weight (p \ 0.05). CONCLUSIONS Overweight and obesity, especially increasing weight, has a negative impact on children’s well-being as measured by parent reports on feelings of insecurity/ bullying and parental worries. We recommend longitudinal studies to assess impact of growth trajectories on various domains of child wellbeing, including children’s self-reports.
(304.4) Linking child and adolescent quality of life scales in youth with epilepsy Nora Fayed, University Health Network, Canada; David Streiner, Baycrest; Gabriel Ronen, McMaster University; Aileen M. Davis, PhD, Toronto Western Research Institute and University of Toronto, Toronto, Canada, Canada AIMS Monitoring quality of life over the course of youth requires that there is some valid means of measuring with continuity across developmental stages. However, conceptual and metric equivalence of these scale versions must be established before their results can be compared or contrasted. The aim of this study is to describe and discuss a scaling procedure for linking child and adolescent versions of a quality of life self-perception scale in a clinical sample of children with epilepsy. METHODS 327 child and 115 teen respondents from a Canadian national sample of children with epilepsy completed the respective versions of the Harter Self Perception Scale. Items from the child and adolescent version of each scale were assessed for conceptual equivalence using a validated content analysis method based in linking to the International Classification of Functioning, Disability and Health (ICF). Items with identical or proximal ICF categories were used to predict unidimensionality. Each scale version was independently fit to the Rasch model using fit statistics of local dependency and item difficulty. The overlapping or equivalent items from each version were used to anchor the adolescent Rasch model to the child model. RESULTS The content analysis method was useful for predicting which items would fit the Rasch model. Items deemed conceptually equivalent also demonstrated unidimensionality with a Rasch model within age versions. Anchoring using items in common for both scale versions was a useful strategy for linking child and adolescent age versions of the quality of life scale. CONCLUSIONS Content analysis combined with anchoring and Rasch modelling procedures can be suitable for linking child and adolescent versions of a latent construct scale such as quality of life based on self-perception
35 in cross-sectional applications. Further exploration of the implications of this procedure of scale linking across developmental stages is needed.
(304.5) The relationships between pediatric asthma control, excessive daytime sleepiness, and health-related quality of life I-Chan Huang, PhD, University of Florida; Zheng Li, University of Florida, United States; Lindsay Thompson, University of Florida; Sanjeev Tuli, University of Florida; Dennis Revicki, PhD, United BioSource Corporation; Elizabeth Shenkman, PhD, Univ of Florida College of Medicine AIMS This study primarily aimed to examine the relationships between asthma control, excess daytime sleepiness, and asthmaspecific health-related quality of life (HRQOL) in children with asthma. The second aim was to examine whether excess daytime sleepiness mediates the relationship between asthma control and asthma-specific HRQOL. METHODS This cross-sectional study utilized data collected from a sample of asthmatic children (n = 160) aged 8–17 years and their parents (n = 160) who visited four pediatric clinics of a university medical center. Excess daytime sleepiness and asthma-specific HRQOL were self-reported by children using the Cleveland Adolescent Sleepiness Questionnaire (CASQ) and NIH’s Patient-Reported Outcomes Measurement Information System (PROMIS) Pediatric Asthma Scale, respectively. CASQ is comprised of the aspects of sleepiness in school, sleepiness in evening, and sleepiness in transportation. Asthma control was reported by parents using a 5-item standardized questionnaire. Regression analyses were conducted to investigate the associations between asthma control and daytime sleepiness, asthma control and asthma-specific HRQOL, and daytime sleepiness and asthma-specific HRQOL, respectively. Structural equation modeling (SEM) was performed to test the mediating effect of excess daytime sleepiness on the relationship of asthma control with asthma-specific HRQOL. Parents’ age and gender, children’s age and gender, and physicianreported pediatric comorbid conditions were adjusted in the analyses. RESULTS Regression analyses suggest poorer asthma control was associated with greater excess daytime sleepiness (p \ 0.001) and poorer asthma-specific HRQOL (p \ 0.001). In addition, greater excess daytime sleepiness was associated with poorer asthma-specific HRQOL (p \ 0.001). In the SEM without including excess daytime sleepiness, children with poorer asthma control reported significantly poorer asthma-specific HRQOL compared to children with better asthma control (p \ 0.001). The inclusion of excess daytime sleepiness attenuated the effect of asthma control on asthmaspecific HRQOL (p \ 0.05). Two aspects of excess daytime sleepiness were identified which significantly mediate the relationship of asthma control with asthma-specific HRQOL: sleepiness in school (p \ 0.05) and sleepiness in evening (p \ 0.05). CONCLUSIONS Children’s asthma control status influenced their asthma-specific HRQOL. Excess daytime sleepiness mediated the relationship between asthma control and HRQOL. To better improve asthmatic children’s HRQOL, clinicians are encouraged to screen for sleep problems among asthmatic children in addition to achieving asthma control.
305: Symptom Monitoring in Clinical Research and Practice (305.1) The psychometric development of multiple asthma quality of life item banks Brian D. Stucky, RAND Corporation, United States; Cathy Sherbourne, PhD, The RAND Corporation; Maria Orlando Edelen,
123
36 PhD, RAND Corp.; Nicole K. Eberhart, RAND Corporation; Marielena Lara-Greenberg, RAND Corporation AIMS A collection of asthma-specific quality of life (QOL) domains and item banks are being developed for use in adult asthmatic populations. These item banks provide unique information regarding how asthma symptoms bother or impact the patient’s QOL (e.g., Functional Impairment, Interpersonal Impact, Trigger Avoidance, etc.). This paper focuses on the state-of-art psychometric steps that enable the development of these item banks. Emphasis is placed on initial assessments of scale dimensionality and identification of sources of local dependence via multidimensional item response theory (MIRT) models made possible through recent statistical software advances. Final item selection criteria and techniques are discussed using unidimensional IRT analyses. We conclude with a presentation of the psychometric properties of the item banks. METHODS By June 2012, more than 2000 asthmatic adults will have completed a computer-administered field test of over 100 asthma-specific QOL tryout items. Using subsets of items, various hierarchical confirmatory factor analysis (CFA) and MIRT models will be evaluated to identify and isolate violations of local independence for each asthma-related QOL domain. Based on these MIRT analyses per domain, a locally independent set of items will be selected which is subsequently analyzed using unidimensional IRT. Following the creation of these domain-specific banks, a general asthma-QOL bank will be developed that recasts the various domain-specific banks in a bifactor IRT model framework and identifies a unidimensional item subset that characterizes the degree of item-level multidimensionality. These recently developed psychometric techniques enable a unidimensional general asthma QOL bank to be developed from multidimensional data. Recommendations will follow regarding how these analytic techniques can be applied to other QOL constructs. RESULTS Results will disseminate the psychometric properties of the asthma QOL item banks. For each item bank, IRT-based scaled score precision estimates will be compared across the complete bank and potential computer adaptive tests (CAT). CONCLUSIONS This study will lead to the development of new item banks for use in measuring domain-specific and general asthma-related QOL. The complex nature of these domains creates a variety of possible approaches to scoring that can be tailored to meet the goals of the end user and minimize respondent burden.
(305.2) Minimally important difference (MID) in scores on the European organization for research and treatment of cancer (EORTC) Quality of Life Questionnaire Core 15 for Palliative (QLQ-C15-PAL) Kikuko Miyazaki, PhD, Schl of Pblc Hlth—Kyoto University, Japan; Yoshimi Suzukamo, PhD, Tohoku Univ, Dept of Physical Med & Rehab, Japan; Kojiro Shimozuma, Dept. of Biomedical Sciences, College of Life Sciences, Ritsumeikan University, Japan; Shozo Ohsumi, National Hospital Organization Shikoku Cancer Center, Japan; Mari Saito, Yokohama City University Medical Center; Takeo Nakayama, MD PhD, Kyoto University School of Public Health, Japan AIMS To determine the minimally important difference (MID) for the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire Core 15 for Palliative (QLQ-C15-PAL), which is a questionnaire used for evaluation of palliative care. METHODS This study was performed in a multicenter, prospective observational design. Subjects were palliative care patients with cancer, diagnosed by an attending physician as having a year or less left to live. The patient registration period
123
Qual Life Res (2012) 21:1–132 spanned from September 2009 to December 2011. The QOL evaluation was conducted using a self-reporting survey. During the second evaluation conducted 2 weeks later, responses of ‘‘improved,’’ ‘‘unchanged’’ or ‘‘worsened’’ were obtained on seven items from the QOL evaluation and Global Rating Scale (GRS). Based on the difference in GRS and two QOL scores, we calculated the patient’s subjective MID. RESULTS Among the 240 registered patients, 181 (92 male; average age, 67 years) who completed two QOL evaluations and GRS were subject to analysis. The number of subjects with ECOG PS of 0, 1, 2, 3, or 4, was 10, 30, 60, 64 and 17, respectively. On GRS, EORTC QLQ-C15-PAL responses for physical functioning (PF), emotional functioning (EF), global quality of life (GL), fatigue (FA), pain (PS) were ‘‘unchanged’’ in 63, 87, 77, 83, and 85 subjects, respectively. MID for ‘‘somewhat improved’’ was 4.1, 10.6, 12.1, 15.3, and 14.9. MID for ‘‘somewhat worsened’’ was 3.6, 6.8, 1.2, 6.1, and 4.6. CONCLUSIONS In the two QOL evaluations, ‘‘unchanged’’ accounted for almost half of the responses. As for the EORTC QLQ-C15-PAL scores, improving MID was 4.1–15.3, whereas worsening MID was 1.2–6.8. These results indicate that, even if the change is small, there is greater sensitivity to perceive worsening outcomes than improving outcomes. Outcomes for palliative care that aims to maintain or improve patient QOL can be obtained by measuring QOL. QOL measurement incorporating MID as a criterion will likely be performed more frequently in the clinical practice of palliative care.
(305.3) PROMIS version 2 pain quality item bank discriminates between neuropathic and other types of pain Dagmar Amtmann, PhD, University of Washington; Cindy Nowinski, MD PhD, Northwestern University; David Cella, PhD, Northwestern University; Rana Salem, University of Washington; Dennis Revicki, PhD, United BioSource Corporation; Fred Wolfe, National Data Bank for Rheumatic Diseases; Kaleb Michaud, National Data Bank for Rheumatic Diseases; Robert L. Askew, BA, UT—M.D. Anderson Cancer Center AIMS The assessment of pain sensation and quality is a key component in understanding the experience of individuals living with chronic pain. Understanding the underlying pain mechanism is essential for effective treatment. Patient descriptors of pain quality can be used to identify different types of pain mechanisms (e.g. neuropathic, nociceptive). METHODS PROMIS developed and tested 37 pain quality items that clustered into 6 factors that have clinical value in discriminating different kinds of chronic pain. Version 2 harmonized the PROMIS adult and pediatric pain quality items and added new items. Revised items were administered to 215 adults with rheumatoid arthritis (RA) and 109 with diabetic neuropathy (DN). RA participants were recruited from the National Data Bank for Rheumatic Diseases and from the Northwestern University Rheumatology clinic. DN participants were recruited from the University of Washington Diabetes Care Center. Exploratory factor analysis (EFA) was used to identify a unique factor with items that measure neuropathic pain. Differences in endorsement levels to the items that loaded on the neuropathic factor were compared between RA and DN. We hypothesized that items that load on the neuropathic factor would be more frequently endorsed by patients with DN than patients with RA. RESULTS EFA identified a factor with descriptors typical for neuropathic pain. The descriptors included itchy, cold, burning, numb, tingly, stinging, like pins and needles, electrical, and ‘‘weird’’ pain or discomfort. DN patients endorsed all descriptors at higher levels than patients with RA (all comparisons were statistically significant p \ 0.01). The
Qual Life Res (2012) 21:1–132 biggest difference between groups was on the ‘‘like pins and needles’’ item for which 60 % of respondents with DN endorsed somewhat to very much compared to 30 % of patients with RA. CONCLUSIONS Version 2 PROMIS pain quality items appear to differentiate between neuropathic pain and other pain, however DN and RA patients differ on a series of factors, not just their endorsement of neuropathic pain descriptors. The pediatric and adult pain quality items are identical and can be used to compare individuals across the life span.
(305.4) Assessing the Association of Cancer-Related Fatigue Thresholds on the Piper Fatigue Scale-12 with sexual functioning in a population-based cohort of breast cancer survivors: a HEAL study Angela M. Stover, University of North Carolina—Chapel Hill, United States; Bryce B. Reeve, Ph.D., University of North Carolina at Chapel Hill, United States; Barbara Piper, University of Arizona; Catherine M. Alfano, National Cancer Institute; Ashley Wilder Smith, PhD, National Cancer Institute; Sandra A. Mitchell, PhD CRNP AOCN, National Cancer Institute; Leslie Bernstein, City of Hope National Medical Center; Katherine B. Baumgartner, University of Louisville; Anne McTiernan, Fred Hutchinson Cancer Research Center; Rachel Ballard-Barbash, National Cancer Institute AIMS A 12-item Piper Fatigue Scale (PFS-12) was recently developed as a parsimonious yet reliable and psychometrically robust version of the original 22-item PFS. Our aims with this analysis were to identify clinically meaningful cut-points on the PFS-12 representing none, mild, moderate, and severe fatigue, and to evaluate the association of these cut-points with decrements in sexual functioning in a population-based cohort of breast cancer survivors. METHODS 799 breast cancer survivors, stages 0-IIIa, were recruited from three SEER registries (New Mexico, Western Washington, L.A.) as part of the Health, Eating, Activity, and Lifestyle (HEAL) study. Health-Related Quality of Life (HRQOL) data were collected a mean of 3 years post-diagnosis and included the 22-item PFS (alpha = .93) on a 0–10 rating scale where higher scores indicate more fatigue; and the Sexual Functioning subscale (alpha = .92) of the Cancer Inventory of Problem Situations (CIPS) which is scored from 0 (not a problem) to 3 (serious problem) with higher scores indicating worse sexual functioning. MANCOVA was used to empirically determine optimal cut-points for the PFS-12 using the SF-36 vitality items as dependent variables. Hierarchical regression was used to examine the amount of post-diagnosis sexual functioning that could be explained by cancer-related fatigue, after controlling for baseline sexual functioning, race, age, education, stage, treatment, and number of comorbidities that limit activities. RESULTS MANCOVA identified optimal score ranges on the PFS12 as: none (0), mildly (1–3), moderately (4–6), or severely (7–10) fatigued (Wilk’s Lambda: 2.15(12), p = .021). Cancer-related fatigue explained 33 % of the variance in post-diagnosis sexual functioning (F(1) = 8.62, p \ .0001). For every step increase in fatigue on the PFS-12 (i.e., none, mild, moderate, severe), significant decrements in sexual functioning scores occurred (F(3) = 7.53, p \ .0001). Survivors reporting no CRF had the best post-diagnosis sexual functioning (mean = .25, SD = .35); mild CRF: mean = .75, SD = .76; moderate CRF: mean = 1.04, SD = .83; and survivors with severe CRF reported the worst post-diagnosis sexual functioning: mean = 1.17, SD = .84. CONCLUSIONS Cumulative decrements in sexual functioning scores occurred for every threshold increase in cancer-related fatigue for breast cancer survivors approximately 3 years after diagnosis. Future research should examine whether interventions targeting CRF may also improve sexual functioning for breast cancer survivors.
37 306: Caregiver Quality of Life and Health Evaluation (306.1) Experienced burden of informal care among caregivers in the Netherlands Renske Hoefman, Erasmus University Rotterdam; Job Van Exel, Erasmus University Rotterdam, Netherlands; Werner Brouwer, Erasmus University Rottedam AIMS Informal caregivers provide a significant part of the total care needed by ill or disabled persons. Although informal care is often the preferred option of those who provide and those who receive informal care, caring can nevertheless be very straining. This study provides insight into the burden experienced by informal caregivers, and shows which characteristics are related to this subjective burden. In addition, we discuss a method to measure and value the effect of providing informal care that can be applied in (economic) evaluations of health care programs, including support programs for caregivers. METHODS Data was collected with a survey distributed to a representative sample of the adult population in the Netherlands in October 2010 (n = 1,244). Subjective burden was measured with the CarerQol instrument containing seven burden dimensions (fulfillment, relational problems, mental health, daily activities, physical health and support; CarerQol-7D) and a visual analogue scale (0–10) measuring general well-being (CarerQol-VAS). CarerQol-7D dimensions were analyzed with logistic regression and CarerQol-VAS with ordinary least squares regression, using three groups of independent variables: (1) informal care situation, (2) care recipient, (3) caregiver characteristics. RESULTS Problems often encountered were physical and mental health problems and problems combining care tasks with daily activities. These problems were more often experienced among caregivers caring for their partner or parent, among caregivers caring for a care recipient receiving professional care and among caregivers of relatively young care recipients. The mean CarerQol-VAS score was 7. This happiness score was lower among caregivers of younger age and lower health, caregivers experiencing depressive symptoms, caregivers providing care to more than one care recipient, caregivers combining informal care with care for a child, caregivers without a paid job, and caregivers lending care to persons with relatively low mental health. CONCLUSIONS While caregivers are relatively happy, many of them experience problems in their caregiving situation. This research shows that not only the health need of care recipients influences caregiving burden, but also characteristics of the informal care situation and of informal caregivers. This is valuable information for policymakers to effectively support caregivers in their important, but often burdensome task.
(306.2) Quality of life among parents of children with autism spectrum disorders: a comparison of generic instruments Nalin Payakachat, BPharm, MS, PhD, Univ of Arkansas for Med Sci; Renske Hoefman, Erasmus University Rotterdam; Erica Kovacs, Columbia University Medical Center Division of Child and Adolescent Psychiatry; Job Van Exel, Erasmus University Rotterdam, Netherlands; Jeffery M. Pyne, University of Arkansas for Medical Science College of Medicine; Karen A. Kuhlthau, PhD, Center for Child and Adolescent Health Policy; John Tilford, Univ of Arkansas for Medical Sciences; Werner Brouwer, Erasmus University Rotterdam AIMS Children with autism spectrum disorders (ASDs) can impact family functioning as well as broad quality of life (QOL) of parents and other family members. This study provides insight into the impact of caregiving on the QOL of parents of children with ASD by reporting outcomes of generic instruments. METHODS Parents were recruited from two sites of the Autism Treatment Network
123
38 (ATN): a developmental center in Little Rock, Arkansas and an outpatient psychiatric clinic of the Columbia University Medical Center in New York. Families were included if the child with the ASD was between 4 and 17 years of age and the parents spoke English. Eligible families were contacted by mail to see if they would be interested in participating in the study. A cross-sectional self-administered survey was mailed to parents of children with ASDs, with 224 parents completing surveys. Outcome measures for the parents included the EQ-5D and SF-6D, Center for Epidemiologic Studies Depression Scale (CES-D), Care-related Quality of Life (CarerQol), Self-Rated Burden scale (SRB). The HUI-3 was used to measure QOL of children with ASDs (proxy-reported). We studied the associations among impact of informal care on parents’ happiness and QOL using regression analysis. We hypothesized that high caregiving burden would lead to low happiness and low QOL among parents. RESULTS The majority of parents were biological mothers, white, with an average age of 40.3 ± 8.2 years. The parents’ average SF-6D and EQ-5D scores were 0.74 ± 0.12 and 0.85 ± 0.14. The average caregiving burden score (CarerQol-7D), happiness score (CarerQol-VAS) were 4.8 ± 2.9 (0 = no burden; 12 = maximum burden) and 7.4 ± 1.9 (0 = completely unhappy; 10 = completely happy), respectively. 97 % of parents reported that they had some or a lot of fulfillment with care tasks. The average SRB score was 36.1 ± 29.6 (0 = not at all straining; 100 = much too straining). 40 % of the parents had clinical depressive symptoms (CES-D = 16, depressed group); parents in the depressed group had significantly lower SF-6D and EQ-5D scores than parents in the non-depressed group. The children’s average HUI-3 was 0.67 ± 0.23. Results from path analysis showed that caregiving burden negatively impacts happiness of the parents (B = -0.614, p \ 0.05) and leads to lower SF-6D and EQ-5D scores (B = 0.377, -0.422, p \ 0.05). In contrast, happiness scores positively impact SF-6D and EQ-5D scores (B = 0.337; p \ 0.05, 0.152; p = 0.05). Child’s QOL was negatively associated with parent’s caregiving burden (p \ 0.05). CONCLUSIONS While providing informal care to children with ASDs can be a fulfilling activity, many parents also experienced challenges in their role as an informal caregiver, especially combining their caregiving tasks with daily activities. Parents’ self-reported happiness and QOL are directly associated with caregiving tasks for their children. Thus, interventions that improve the QOL of children with ASDs will decrease parents’ caregiving tasks and improve parents’ QOL.
(306.3) Health status of informal caregivers of cancer and prostate cancer patients in Europe Amir Goren, PhD, Kantar Health, United States; Isabelle Gilloteau, Bristol-Myers Squibb, United States; Ave Mori, Bristol-Myers Squibb; Marco DiBonaventura, Kantar Health AIMS Cancer imposes a burden on patients and on those who provide care for them. This study measures the health status of caregivers of patients with cancer with a specific focus on caregivers of patients with prostate cancer compared to non-caregivers. METHODS Data were obtained from 2010 to 2011 EU National Health and Wellness Surveys (NHWS), self-administered online surveys of a demographically-representative sample of adults in France, Germany, Italy, Spain, and the United Kingdom (n = 114,962). Respondents who reported providing informal care for a patient with cancer (‘‘cancer caregivers’’) or with prostate cancer specifically (‘‘prostate cancer caregivers’’) were compared with respondents not providing care (‘‘non-caregivers’’) on health status from the Short Form-12v2-based scores (including the SF-6D health utility measure) and on diagnosed comorbid illness. Regression models predicted outcomes of caregivers versus non-caregivers, controlling for demographics, health risk
123
Qual Life Res (2012) 21:1–132 behavior, and Charlson comorbidity index. RESULTS Cancer caregivers (n = 1,713) and prostate cancer caregivers (n = 95) were significantly more often female than non-caregivers (n = 103,868) (58.7 and 67.4% vs. 50.4 %, respectively). Adjusting for covariates, cancer caregivers reported significantly worse health status than noncaregivers (physical component summary: -1.32 points; mental component summary: -3.26 points; health utilities: -0.043; all eight domains: -0.68 to -3.65 points) (all p \ .001) and experienced greater odds than non-caregivers of being diagnosed with depression (OR = 1.504), anxiety (OR = 1.967), insomnia (OR = 2.013), headache (OR = 1.369), migraine (OR = 1.663), and gastrointestinal problems (OR = 1.626) (all p \ .001). Analyses on prostate cancer caregivers revealed significantly worse health status (mental component summary: -2.88 points; health utilities: -0.040; Mental Health domain: -3.06 points; Social Functioning domain: -2.65 points) (all p \ 0.02) and higher odds of diagnosed anxiety (OR = 1.823), insomnia (OR = 2.113), and gastrointestinal problems (OR = 2.237), (all p \ 0.02) compared with non-caregivers. CONCLUSIONS The results of the study demonstrate that providing informal care to a person with cancer has major physical and mental consequences for caregivers in Europe. In the particular case of prostate cancer, special attention should be paid to the mental health of caregivers, as well as to stress-related comorbidities, which in turn may translate into better health and well-being for the cancer patients themselves and potentially into better control of healthcare resource utilization for society.
(306.4) Validation of the distress thermometer for parents with a chronically ill child Lotte Haverman, MSc, Emma Children’s Hospital—Academic Medical Centre, Netherlands; Hedy Van Oers, Emma Children’s Hospital/Academic Medical Center; Perrine Limperg, Emma Children’s Hospital/Academic Medical Center, Netherlands; Bregje Houtzager, MA, Pediatric Psych, Netherlands; Jaap Huisman, VU medical centre; Anne-Sophie Darlington, PhD, University of Southampton School of Medicine, United Kingdom; Heleen Maurice-Stam, MSc, Emma’s Children’s Hosp, Academic Medical Ctr, Netherlands; Martha Grootenhuis, PhD, Academic Medical Center, Netherlands AIMS Parents with a chronically ill child frequently report mood problems, anxiety, physical problems, cognitive problems and feelings of lack of control over daily events. A short screeningquestionnaire that identifies parents with serious distress, problems and need for additional support would be useful in medical settings. The present study aimed to develop a screening-questionnaire (Distress Thermometer for Parents; DT-P), to examine the psychometric properties, and to assess a clinical cutoff score. METHODS Parents with a chronically ill child (0–18 years) were informed with announcements or actively approached at the outpatient clinics of the Emma Children’s Hospital and VU medical centre. Based on the DT used in oncology medical care, literature and clinical experience, the DT-P was developed. It consists of a distress thermometer from 0 (no distress) to 10 (extreme distress), a problem list (practical, social, emotional, physical, cognitive and parenting domains) and a question about parental wish for referral. The DT-P was validated with the Hospital Anxiety and Depression Scale (HADS) and the Parenting Stress Index (PSI). RESULTS The mean thermometer-score of the 706 participating parents was 3.7 (SD 3.0). The thermometer-score (.55 = r = .61) and the scores in the practical, emotional, physical and cognitive problem domains (.55 = r = .72) were strongly related to anxiety, depression and the total score of the HADS. The thermometer-score and all problem domain scores were moderately to strongly related to the PSI (.38 = r = .63). Parents with a wish for
Qual Life Res (2012) 21:1–132 referral scored significantly higher on the thermometer than parents without a wish for referral; 5.5 (SD 2.8) versus 3.1 (SD 2.8), p \ .001. A cutoff score of 4 correctly identified 86 % of ‘clinical-HADScases’ (sensitivity) and 67 % of ‘non-clinical-HADS-cases’ (specificity). CONCLUSIONS The DT-P is a useful rapid screening tool for monitoring parental distress. The DT-P could help to focus at those parents who most need and want support. In pediatrics, Parent Reported Outcomes could be a valuable additional component in the implementation of electronic Patient Reported Outcomes in the care of chronically ill children.
(306.5) Detecting response shift in a longitudinal study of stroke caregivers Lisa M. Lix, PhD, School of Public Health, University of Saskatchewan, Canada; Tolulope T. Sajobi, PhD, University of Saskatchewan, Canada; Richard Sawatzky, PhD RN, Trinity Western University, Canada; Juxin Liu, University of Saskatchewan, Canada; Nancy E. Mayo, PhD, McGill University, Royal Victoria Hospital, Canada AIMS Longitudinal change in health-related quality of life (HRQOL) may be difficult to interpret because of response shift (RS), a change in internal standards, values, or conceptualizations of health. RS can occur in stroke patients, but it is not known whether caregivers may also experience RS. The study aim was to investigate RS in the HRQOL of stroke spousal and non-spousal caregivers, comparing individuals caring for patients with severe and less severe stroke symptoms. METHODS Data were from a Canadian multi-center stroke study for N = 409 caregivers. RS was investigated for baseline (i.e., immediate post-stroke) and 6-month post-stroke occasions. New methods recently proposed to detect reprioritization RS were used; they are based on tests of change in relative importance weights for standardized discriminant function coefficients (SDFCs) and standardized logistic regression coefficients (SLRCs) for two groups. They were applied to the eight domains of the SF-36. The Stroke Impact Scale (SIS) physical function score at baseline, with a cut-off of 75, was used to measure patient stroke severity. RESULTS Spouses comprised 70 % of the sample (i.e., n = 205); 40 % of caregivers (i.e., n = 163) were caring for patients with severe stroke. SF-36 physical health scores were lower for spousal than non-spousal caregivers at baseline (spousal mean = 49.9 [SD = 10]; non-spousal mean = 53.4 [SD = 8]), but mental health scores were similar. For spousal caregivers, there were no statistically significant changes in relative importance weights for the SF-36 domains. For the non-spousal caregivers, there were statistically significant changes (p \ .05) in relative importance weights for the bodily pain and role physical domains. CONCLUSIONS RS may not have the same effect on all stroke caregivers, which has implications for measuring the effectiveness of interventions targeted at this population. The results suggest that nonspousal caregivers of stroke patients with severe symptoms are more likely to experience a shift in the relative importance of bodily pain and physical role to their overall quality of life than caregivers of patients with less severe symptoms. For spousal caregivers, there was no evidence of reprioritization of quality of life domains. 307: Psychometric Approaches (307.1) Spatio-temporal Rasch analysis of quality of life outcomes in the French general population. Measurement invariance and group comparisons Jean-Benoit Hardouin, PhD, University of Nantes—EA4275, France; Etienne Audureau, Assistance Publique-Hoˆpitaux de Paris, Universite´ Paris-Descartes, France; Alain Leple`ge, University
39 of Paris Diderot, France; Joe¨l Coste, Assistance Publique-Hoˆpitaux de Paris, Universite´ Paris-Descartes, France AIMS Worsening trends in HRQoL have been reported in France between 1995 and 2003, using Classical Test Theory. The present work aimed at confirming those results within the framework of Item Response Theory (IRT) [with a model of the Rasch family model] so as to check the crucial property of invariance of the measurement instrument. We present the step-by-step methodology and the advantages to using such a method for monitoring HRQoL in population. METHODS Data were drawn from two large population-based surveys conducted in France in 1995 (N = 3,656) and 2003 (N = 23,018). HRQoL was measured through three dimensions from the SF-36 questionnaire (General Health, Mental Health and Physical Functioning). After checking assumptions relating to IRT, Differential Item Functioning (DIF) was assessed to examine potential item bias across age categories, genders, regions of residency, and years of study. Latent regression mixed Partial Credit Models were fitted to explore the impact of covariates on HRQoL (mean and variance), accounting for the potential DIF previously detected. RESULTS DIF had been detected across age categories and genders. By taking account this DIF, a general decrease of HRQoL was observed between 1995 and 2003, after adjustment on region, age and gender. HRQol decreased with age, and worse HRQoL outcomes were found for women. Disparities were also found across regions, indicating better mean values in Paris, the West and the South, and worst values in the North and the East. CONCLUSIONS Rasch family model offers several advantages compared to a classical approach based on scores. First, DIF can be taken into account. More, handling of missing data is easy, because these models do not required imputation of missing data. Indeed, the specific objectivity property of Rasch family models allows obtaining consistent estimations of the parameters linked to the latent trait, whatever the set of items having, or not, been observed for each individual. Last, analysis using Rasch family model is more powerful than analysis based on scores, and allow highlighting a most important number of effects.
(307.2) The evaluation of the patient rated elbow evaluation using Rasch analysis Joshua I. Vincent, MPT, The University of Western Ontario, Canada; Joy MacDermid, PhD, McMaster University, Canada; Ruby Grewal, University of Western Ontario AIMS The Patient Rated Elbow Evaluation (PREE) is a joint specific measure of pain and disability and validated with classical psychometric methods. The purpose of this study is to: (1) To determine aspects of validity of the PREE using the Rasch model to assess the overall fit of the PREE data, the response scaling, individual item fit, differential item functioning (DIF) and person separation index (PSI). (2) To evaluate gender differences in answering different items of the questionnaire using DIF analysis. METHODS This is a cross-sectional study. A sample of 236 patients (Age range 21–79 years; M:F—97:139) with elbow disorders where recruited from The Hand and Upper Limb Centre, London, Ontario, Canada. Patients completed the PREE 6 months post-surgery. Rasch analysis was performed using RUMM 2030 software to analyse the overall fit of the PREE data to the model, the response scale, individual item fit, DIF and reliability indices (PSI). Gender differences in responding to the items of the questionnaire were analysed using DIF analysis. RESULTS The PREE initially showed disordered thresholds (16 out of 20 items), uniform DIF for gender with two items (‘‘Carrying a 10lbs object’’ from specific activities subscale; and ‘‘household work’’ from the usual activities subscale) and a highly significant Chi-square value of
123
40 103.81 (df = 60; p = 0.0003) indicating misfit to the Rasch model. DIF for gender in the two items indicated that men and women answered those questions differently. Later after rescoring the items, they were regrouped into three super items namely pain, light activities and heavy activities, then the PREE data fitted the Rasch model perfectly with an item fit residual of -0.308 and S.D. of 0.927 and a person fit residual of -0.386 and S.D. of 0.776, with a non-significant Chi square value of 14.07 (df = 6) p = 0.029. As well the PREE was well targeted (Mean location = -0.75), unidimensional (Per C \ 5 % = 2.7 %) and had high person separation index (PSI = 0.85) and the DIF for gender was eliminated. CONCLUSIONS The PREE has been proven to be a viable selfreport measure through Rasch analysis. The DIF for gender exhibited by the questionnaire was eliminated by using sub-test analysis.
(307.3) Is Rasch model analysis applicable in small sample pilot studies for assessing preliminary item characteristics: an example using data from PROMIS pain behavior item bank Wen-Hung Chen, PhD, United BioSource Corporation; William Lenderking, PhD, United BioSource Corporation; Ying Jin, United BioSource Corporation; Dennis Revicki, PhD, United BioSource Corporation AIMS Rasch modeling has recently gained attention for assessing content validity of patient-reported outcome (PRO) instruments. It is generally understood that large samples (at least 100) are needed to obtain robust estimates of item parameters and person measures. More recently, it was suggested that a small pilot study should be conducted using Rasch analysis as the preliminary assessment of the dimensionality and Guttman scale properties of the items in developing a PRO instrument. The objective of this study is to evaluate the results of Rasch analysis using small sample sizes. METHODS Twenty items from the PROMIS pain behavior item bank were used for these analyses. Random samples of 30, 100, and 250 were drawn ten times each from the total sample of 1,637 subjects. Rasch analysis was conducted for each of the random samples, as well as the full sample. The numbers of items with incorrectly ordered item parameters or with large fit residuals were compared. RESULTS In the full sample, two items were identified as incorrectly ordered and four with fit residuals exceeding ±3. The average numbers of items with incorrectly ordered item parameters were 4.5, 5.5, and 8.1, for sample size of 250, 100, and 30, respectively. The average numbers of items with fit residuals exceeding ±3 were 0.7, 0.2, and 0.1, for sample size of 250, 100, and 30, respectively. The averages of the average fit residuals were 0.125, 0.198, and 0.250, respectively. CONCLUSIONS Rasch analysis based on small samples identified a greater number of items with incorrectly ordered item parameters than the larger sample (=250). Fewer items were identified as mis-fitting which may lead to the conclusion that these items were satisfactorily unidimensional. Rasch analysis based on small sample is not recommended for preliminary assessment of item properties.
(307.4) Rasch measurement: a mixed methods approach for assessing content validity Stacie Hudgens, MA, Adelphi Values; Patrick Marquis, MD, MAPI Values AIMS Patient reported outcomes are often comprised of multi-item questionnaires designed to measure latent constructs such as symptoms or impacts (i.e., physical, functional, emotional, or social). Development of reliable and valid clinician or patient reported
123
Qual Life Res (2012) 21:1–132 outcomes requires a deep understanding of the patient population and the underlying traits of symptomology and disease impact. In order to accomplish this, content must be fully explored during the qualitative phase of instrument development. This methods paper suggests a mixed methods approach for the assessment of validity and reliability which combines the rigor of qualitative and quantitative methodology in an iterative fashion during the qualitative phase of instrument development as well as the quantitative validation phase METHODS In addition to concept elicitation and cognitive debriefing with patients, along with the assessment of saturation, we recommend the Rasch measurement model (RMM) which can be implemented iteratively during instrument development to fully understand the content validity and the underlying latent structure and reliability of PRO measures prior to use in clinical trials. The RMM is an extension of item response theory that models the probability of a specific response based on item difficulty (severity) and person ability (severity) which can be used to assess pilot responses in as few as 30 patients. RESULTS It has been selected as the supplementary item analytic approach to classical statistics due to the following: ability to triangulate against the qualitative research findings, ability to assess the domain structure, rating scale structure and reliability prior to finalization and validation, feasibility of evaluating and selecting items that most appropriately target the patients symptom/impact severity, ability to assess the latent severity on the construct (identify construct saturation and deficiency in the measured items), robust estimation of item and person parameters in limited sample sizes, and the ability to evaluate and statistically compare subgroup latent severity. CONCLUSIONS For this study, a series of simulated and real world data examples will be presented to show the above methods for the application of the Rasch model in the ascertainment of content validity during the instrument development phase.
308: Mental Health Evaluation (308.1) Mindfulness: a systematic review of measures of an emergent PRO Taehwan Park, University of Minnesota; Maryanne Reilly-Spong, University of Minnesota; Cynthia Gross, PhD, Univ of MN, Pharmacy & Nursing, United States AIMS Mindfulness, a concept derived from Buddhist psychology, has been secularized and adapted for health purposes. A growing body of evidence shows that mindfulness-based interventions reduce symptom distress and improve health-related quality of life in patients with chronic disorders. The aims of this study are to critically assess and compare the measurement properties of mindfulness scales. METHODS The study was conducted in three stages: (1) selection of relevant studies; (2) evaluation of methodological quality for each study; and (3) determination of scale quality. First, PubMED, PsycINFO and Web of Science were searched to identify articles. Articles were selected if their primary purpose was scale development or evaluation of the measurement properties (validity, reliability or responsiveness) of a self-report mindfulness scale. Next, the COSMIN checklist was used to evaluate the methodological quality of the selected studies. Measurement properties were separately assessed by two independent reviewers. Discrepancies were discussed with a third reviewer, and final scores were obtained based on the discussion. Finally, each scale was judged using a level of evidence approach that synthesizes results and study quality. RESULTS The search strategy identified 717 articles and 26 were selected representing 10 mindfulness scales. Several scales were rigorously developed following a thorough explication of the concept of mindfulness, measured as a whole or by sub-domains. Definitions were mostly drawn from the writings of Jon Kabat-Zinn, who developed mindfulness-based stress
Qual Life Res (2012) 21:1–132 reduction, a program to assist patients adapt to the stressors of chronic illness. Several scales had multiple evaluations of internal consistency and validity (structural validity and hypothesis testing) assessed in adequately-sized clinical and healthy populations with positive results. Major gaps were identified. We found no evidence to support comprehensive coverage of the content domain and limited or no evidence of test–retest reliability, responsiveness or estimates of measurement error. CONCLUSIONS Current mindfulness scales have important conceptual differences, and none can be strongly recommended based on superior psychometric properties. Important limitations in the field are the absence of a gold standard or external referents to support construct validity. Investigators need to proceed cautiously before optimizing any mindfulness intervention based on the existing scales.
(308.2) How is depression valued when it co-occurs with a somatic condition? Pilot study using in depth interviews Katerina Papageorgiou, MSc, University Medical Center Groningen, Netherlands; Karin Vermeulen, PhD, University Medical Center Groningen, Netherlands; Maya J. Schroevers, University Medical Center Groningen; Fenna RM Leijten, Rotterdam Erasmus Medical Center, Netherlands; Erik Buskens, University Medical Center Groningen; Adelita V. Ranchor, University Medical Center Groningen, Netherlands AIMS Time-Trade-Off (TTO) based utility values can be used to evaluate psychological interventions offered to somatic patients with depression. Previously reported valuation tasks and values consider depression as a single state. Yet, depression might be valued differently when it co-occurs with a somatic condition. This pilot study examines • feasibility and accuracy of the TTO task to value depression • differences in values assigned to single state depression and depression co-occurring with a somatic condition • aspects participants consider during their valuations. METHODS During in depth interviews, ten healthy individuals (five women, six university educated, mean age: 36) completed TTO valuations, while Thinking Aloud. They also rated difficulty of the task (0–10). Eight states were valued: single state mild and severe depression (N = 10) and mild depression co-occurring with both moderate and severe cancer (N = 7), diabetes (N = 6), or heart disease (N = 7). RESULTS • The valuation task proved feasible (mean difficulty: 2). However, qualitative analysis of Think Aloud transcripts revealed particular ambiguities in the task perception. Some participants (a) considered the time frame suggested in the TTO task as unrealistic (N = 7), (b) questioned the ‘‘intervention to cure depression’’ (N = 5), and (c) regarded the option to live without depression but with a somatic condition as unrealistic (N = 3). • Compared to depression as a single state, depression was valued worse (difference [ .1) when it co-occurred (a) with moderate cancer, diabetes and heart disease (by four, three and two participants, respectively), and b) with severe cancer, diabetes and heart disease (by six, four and three participants respectively). • Personal experience with the disease (N = 5), perceived susceptibility (N = 4) and family circumstances (N = 3) were considered by participants during their valuations. CONCLUSIONS The TTO can be used for valuations of depression co-occurring with a somatic condition. In this context, revising the time frame and the reference to an intervention is expected to improve its validity. Compared to depression as a single state, depression co-occurring with a somatic condition seems to be valued worse. This is especially true when depression co-occurs with a) severe levels of somatic conditions and (b) cancer. Findings regarding perceived susceptibility indicate that attitudes towards depression might affect valuations.
41 (308.3) Screening accuracy of the Center for Epidemiologic Studies Depression Scale (CES-D): a systematic review with meta-analysis Gemma Vilagut, MSc, Institut Muncicipal d’Investigacio Medica, Spain; Carlos G. Forero, PhD, IMIM, Spain; Gabriela Barbaglia,, IMIM-Research Institute Hospital del Mar; Jordi Alonso, MD PhD, IMIM-Hospital del Mar Medical Research Institute, Spain AIMS The Center for Epidemiologic Studies Depression Scale (CESD) is a popular instrument to assess depression and psychopathology in general population surveys. We carried out a systematic review of all the studies evaluating the diagnostic accuracy of the CES-D to detect depressive disorder in the general population to perform a meta-analysis of the diagnostic accuracy of the CES-D. METHODS We searched the MEDLINE database from January 1st, 1990 through January 31st, 2012. Studies were eligible if: (a) they were validation studies of anxiety and depression screening questionnaires with information on the diagnostic accuracy of the instruments; (b) the sample was taken from the general population or primary care settings; (c) standardized diagnostic interviews following standard classification systems were used; and (d) the language of publication was either English or Spanish. We estimated the summary receiver operating characteristic (SROC) curve through Rutter and Gatsonis mixed effects models. For the studies using the common cut-off point of 16, we fitted bivariate mixed effects model to obtain pooled estimates of diagnostic accuracy indexes. Covariates like setting, age group, or gold standard, were included in the models to assess possible sources of heterogeneity. RESULTS Two hundred and eighty five studies including the CES-D questionnaire were identified, of which 26 (n = 10,375 individuals) fulfilled inclusion criteria after peer review. Twelve of the studies were conducted in primary care settings. Median prevalence of major depression was 8.7 % (IQ range 3.9–14.2 %). Eighteen studies provided test accuracy information at cut-off point of 16, and pooled estimates at this threshold were: sensitivity = 0.85 [95 % CI = (0.80, 0.89)], Specificity = 0.72 [95 % CI = (0.66, 0.76)], and Diagnostic Odds Ratio DOR = 14.45 [95 % CI = (9.07,23.02)]. None of the studied possible sources or heterogeneity was found to be statistically significant. CONCLUSIONS Diagnostic accuracy of the CES-D is only moderate. The low specificity found at the common threshold of 16 may limit its appropriateness for the assessment of depression in the general population. A higher cut off point with higher specificity might be more adequate for this purpose.
(308.4) An IRT mixture model to correct prevalence bias from censored binary indicators: application to lifetime mental health comorbidity in the ESEMeD surveys Carlos G. Forero, PhD, IMIM, Spain; Josue Almansa, Utrecht University; Gemma Vilagut, MSc, Institut Muncicipal d’Investigacio Medica, Spain; Jordi Alonso, MD PhD, IMIM-Hospital del Mar Medical Research Institute, Spain AIMS Cross-sectional surveys often estimate lifetime prevalence of mental outcomes which might be underestimated as individuals are not observed beyond the time of the interview. Censorship bias is even more serious for multi-morbid prevalences. Here we propose a finite-mixture Item Response-Theory approach to correct prevalence estimates from censored indicators. METHODS We describe a weight function of outcome realization and time of onset that is applied to the likelihood function of a finite-mixtures two-parameter logistic model. Estimates are obtained using Pseudo-Maximum Likelihood. Univariate and multimorbid prevalences are obtained as model marginal probabilities. Parameter standard errors are computed using the Delta Method. The
123
42 model was applied to lifetime data of mental health disorders (depression, dysthymia, GAD, PTSD, agoraphobia, specific and social phobia, and panic disorder) as diagnosed with the CIDI 3.0 interview in the ESEMeD surveys, a cross-national study of representative adult general population in 6 European countries (n = 8,796). Person scores were obtained to permit a model-based epidemiologic population description, using weight to adjust for individual population representativeness. RESULTS A unidimensional model fitted the data adequately (average BVR = 0.06). univariate model prevalences ranged from a low 1.6 % for agoraphobia to a high 17.9 %, while the raw prevalences ranged from 1.2 to 13.4 % for the same disorders. Univariate raw prevalences were on average 1.4 times lower than model based estimates, 1.9 times lower in the case of bivariate comorbid prevalences. Age of onset and relative prevalence ratios (model estimate/raw estimate) correlated highly (r = 0.86), indicating that as time of onset increased, IRT model yielded progressively higher corrections. Person scores showed women more vulnerable of developing a disorder and older ages, protective. CONCLUSIONS The weighted likelihood estimator provides an effective method to correct prevalence underestimation due to censorship. It is also able to compute person scores useful for population and individual inferences. This theory-driven method opens research opportunities in psychiatric epidemiology when applied to lifetime mental outcomes.
309: Psychometric Approaches (309.1) Empirical evaluation of multivariate meta-analysis for multiple outcomes of QOL measurement using simulated data Kazue Yamaoka, PhD, Teikyo University, Graduate School of Public Health, Japan; Toshiro Tango, Center for Medical Statistics, Japan AIMS Many QOL studies report several dimensions measured as subscales and involve multiple outcomes reported in the same studies. The usual approach is to perform separate univariate meta-analyses of each dimension. However, these outcomes are evaluated in the same participants, and are thus stochastically dependent (correlated). When the question of interest involves multiple outcomes, multivariate metaanalysis is used to synthesize the outcomes simultaneously taking into account the correlation between the outcomes. In this presentation, we examined how the use of univariate or multivariate analysis affects the results of meta-analyses through simulated data. The results will be useful for making methodological recommendations for meta-analyses of the QOL data in general. METHODS We performed a simulation study to understand how the univariate and multivariate meta-analyses yield different results by correlation structure. We consider simulated data by assuming various correlation structures of the subscales and number of studies. These motivate a simulation study where estimation properties from multivariate random-effects meta-analysis (MRMA) are compared with those from separate univariate random effects meta-analyses (URMAs), the traditional approach. By assuming complete data, scenarios were considered 3 types of the withinand between-study correlations, the sizes of the meta-analysis studies. RESULTS Our simulations reveal the discrepancy between MRMA and URMA commonly occurs when the number of studies is small or the correlation between QOL subscales is varied. We should carefully interpret the results especially when the correlations between QOL subscales were highly correlated. The results by MRMA model did not sometimes converge due to small sample size. CONCLUSIONS A MRMA model offers advantages over separate univariate synthesizes and it highlights some of benefits corresponding to the structure of correlations. Conceptually, when we assume that QOL subdimensions construct global QOL, MRMA will provide more appropriate aspects of the results compared to the results from URMA. The details will be shown in the presentation.
123
Qual Life Res (2012) 21:1–132 (309.2) Does the choice of the minimal important difference for individual subjects affect the clinically relevant difference at the group level? Georg Kemmler, PhD, Innsbruck Medical University, Austria; Johannes M. Giesinger, PhD, Innsbruck Medical University, Austria; Monika J. Sztankay, Innsbruck Medical University, Austria; Lisa M. Wintner, Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Austria; Bernhard Holzner, PhD BE, Univ. Hospital, Innsbruck, Austria AIMS There has been a large amount of research on the concept of the minimal important difference (MID) or minimal important change (MIC) in recent years. However, there is still controversy about its use in patient groups rather than individual subjects, e.g., in clinical trials. Often the same MIDs are used uncritically for individual subjects and groups. Aim of this study is to bring more light into this unsatisfactory situation. METHODS Based on the MID (or MIC) as a threshold for relevant change in the individual subject, we consider the proportion of subjects in a group experiencing relevant improvement and the proportion without relevant improvement. We then study the effect of the chosen MID on standard measures for comparing two groups (odds ratio (OR), number needed to treat (NNT)) and relate it to the difference in mean change between the two groups. Results are based on theoretical considerations and simulations. RESULTS Generally the size of the MID for individual subjects had only a fairly small effect on the derived thresholds for relevant mean differences on the group level. Thus, for normally distributed changes, an increase of the chosen MID from 0.2 SD to 1 SD (factor 5) led to an increase of the respective relevant mean difference on the group level by at most 20 % (factor 1.2) when based on the OR and usually by less than 50 % when based on the NNT. CONCLUSIONS The choice of the MID for individual subjects obviously has only a rather small effect on the derived relevant mean difference for the comparison of groups. Apparently, other aspects than the MID (e.g., issues related to clinical decision making) are of more importance for the determination of the relevant difference on the group level.
(309.3) Analysis of longitudinal patient-reported outcomes with intermittent missing data (informative and non informative): comparison of score and Rasch-based methods Elodie De Bock, Universite de Nantes, France; Jean-Benoit Hardouin, PhD, University of Nantes—EA4275, France; Myriam Blanchin, University of Nantes, France; Tanguy Le Ne´el, Nantes University, France; Gildas Kubis, Nantes University; Etienne Dantan, Nantes University; Ve´ronique Se´bille, University of Nantes, France AIMS The purpose of this study was to identify the most adequate strategy to analyze longitudinal PRO in the presence of possibly informative intermittent missing data. Models coming from classical test theory (CTT) and item response theory (IRT) were compared. METHODS It was assumed that the PRO had been previously validated using CTT and IRT, as it is frequently performed nowadays. The patients’ responses were simulated with a longitudinal Rasch model. Patient’s responses to dichotomous items with three times of assessment were simulated. Different cases were considered: presence of a time effect or not, 100 or 200 patients, 4 or 7 items, and three different values for the correlation coefficient of the latent trait between two consecutive times were used (0.4; 0.7 and 0.9). Two types of intermittent missing data were compared: Missing Completely At Random and Missing Not At Random at different rates (10; 20 and 30 %). These simulated data were analysed with CTT using Score and Mixed models (SM) method (with Personal Mean Score— PMS—imputation) and with IRT using Longitudinal Rasch Mixed model (LRM) method. The type I error, the power and the bias of the
Qual Life Res (2012) 21:1–132 time effect estimation were compared between the two methods used for the analysis of longitudinal PRO data with intermittent missing data. RESULTS As expected, power of the tests was greatly improved by PMS imputation but, the gain of power was linked to an overestimation of the effect. It could be noticed that, for the LRM method, no imputation was necessary and that the values of power were overall higher than those obtained with SM method. Moreover, there was no bias of the time effect estimation for the LRM method. Results were the same for informative and non-informative intermittent missing data. CONCLUSIONS This study showed that the LRM method performs better than the SM method when analyzing longitudinal PRO in the presence of intermittent missing data (informative or not). Indeed, the specific objectivity, a Rasch model property, may ensure that the latent variable may be estimated consistently even for patients with missing items. Moreover, this study also pointed out the limits of PMS imputation.
(309.4) Development and testing of standards for individualization of patient reported outcome (PRO) measures using EMPRO Jose M. Valderas, MD, PhD, MPH, University of Oxford, United Kingdom; Montserrat Ferrer, MD PhD MPU, Institut Municipal d’Investigacio Medica (IMIM), Spain; Joan Mendivil, Bayer Healthcare; Olatz Garin, MPH, IMIM-Hospital del Mar Research Institute, Spain; Luis Rajmil, PhD, Age`ncia d’Informacio´ Avaluacio´ i Qualitat en Salut (AIAQS), Spain; Michael Herdmann, IMIM University, Insight Consulting and Research; Jordi Alonso, MD PhD, IMIM-Hospital del Mar Medical Research Institute, Spain; Daniela C. Gonc¸alves, University of Oxford, United Kingdom AIMS Individualized Patient Reported Outcomes (PROs) are measures for which each individual respondent is actively involved in the specification of its content and/or scoring. The ‘‘Evaluating the Measurement of Patient-Reported Outcomes’’ (EMPRO) instrument is a well established tool for the modular structured assessment of PROs, but it does not at present assess individualization. The aims of this study were to identify evaluation standards for individualization and to apply them to the development of a module for the EMPRO tool. METHODS Individualized PROs were identified through a systematic review and experts’ consultation. These instruments were analysed and a framework for evaluation individualized PROs was developed. The framework was submitted to experts in the area of PROMs development, use, and evaluation who were invited to participate in a three-round Delphi study. For each round the panel was sent a questionnaire containing the evaluation framework for individualized PROs along with closed and open-ended questions, and responses were systematically summarized and integrated into the next round. At every stage participants had the opportunity to reappraise previous feedback. The Delphi study is still ongoing. RESULTS The evaluation framework for individualized PRO measurement included standards in relation to individualization of: (a) content nomination (4 items), and/or selection (4), e.g.: ‘‘The number of domains to be nominated is clearly stated and is appropriate’’); (b) response scale (2), e.g.: ‘‘If the patient is given the opportunity to provide input into the response scale, the rationale is clearly stated and is adequate’’); (c) score weighting (2), e.g.: ‘‘If the patient is given the opportunity to provide input into the weighting procedure, the methods are clearly described and are adequate’’; and intra-individual stability of individualization (2), e.g.: ‘‘Sufficient evidence is presented regarding the framing of the instrument for longitudinal use’’. Results of Delphi study will be presented. CONCLUSIONS Despite their ability to convey idiosyncratic aspects of the patient’s perspective, individualized PROs might present
43 specific structural and methodological limitations. The newly developed EMPRO module will assist in the standardized assessment of the relative merits of candidate individualized measures.
Thursday Poster Session (1001) Assessing health related quality of life in pediatric brain tumor patients treated with proton radiotherapy: a comparison with a published brain tumor cohort Karen A. Kuhlthau, PhD, Center for Child and Adolescent Health Policy; Sundeep K. Bhat, Stanford/Kiaser Emergency Medicine Residency; Bow Yeap, Massachusetts General Hospital; Jennifer Delahaye, Massachusetts General Hospital; Thomas DeLaney, Massachusetts General Hospital; Shannon MacDonald, Massachusetts General Hospital; David Ebb, Massachusetts General Hospital; Paul Fisher, Lucile Packard Children’s Hospital at Stanford; Nancy Tarbell, Massachusetts General Hospital; Torunn Yock, Massachusetts General Hospital AIMS The use of radiotherapy (RT) can impair Health Related Quality of Life (HRQoL) in survivors of childhood brain tumors. Proton radiotherapy (PRT) reduces normal tissue irradiation and therefore may partially mitigate negative effects on HRQoL To compare HRQoL in PRT and photon (XRT) pediatric brain tumor survivors.; METHODS HRQoL data was prospectively collected during proton radiotherapy and annually thereafter using the PedsQL survey for patients age 2–18 treated at Massachusetts General Hospital (MGH). Mean PedsQL HRQoL scores were calculated and compared to cross-sectional data obtained from a Lucille Packard Children’s Hospital (LPCH) study by diagnosis and by treatment type.; RESULTS The HRQoL scores in the PRT cohort compare favorably to the XRT cohort in many groups of patients. Patients treated with RT without chemotherapy (CT) demonstrated the largest differences, with mean total summary scores of 81.2 (PRT) versus 52.6 (XRT, P \ .001), physical health summary scores of 86.4 versus 55.2 (p \ .001) and psychosocial summery score of 78.5 versus 50.3 (p \ .001) respectively. There were no significant differences in HRQoL between PRT and XRT cohorts in children treated with combined RT and CT. Differences between the PRT and XRT group were correlated with diagnosis. Total summary scores were higher in the PRT cohort in patients with medulloblastoma, ependymoma/highgrade glioma and low grade glioma, but no different for patients with germ cell tumors.; CONCLUSIONS PRT is correlated with higher HRQoL scores in many patients but depends upon treatment type and diagnosis. We further comment on the challenges of a comparative study for a treatment with presumed benefits.
(1002) Transition from pediatric to adult care: a qualitative meta-synthesis Liv Fegran, University of Agder, Kristiansand, Norway, Norway AIMS To synthesize qualitative studies on how adolescents and young adults with chronic diseases experience transition from pediatric to adult care; METHODS Literature search in the databases PubMed, Ovid, Scopus, CINAHL, ISI, and Nordic and German databases covering the years from 1999 to November 2010 was conducted. In addition, forward citation snowballing search was conducted in the databases OVID, CINAHL, ISI WOS, SCOPUS and Google Scholar. Of the 1,143 screened records, 18 studies were included. Inclusion criteria were English, German or Scandinavian language qualitative studies on adolescents’ and young adults’ transitions experiences
123
44 transferring from pediatric to adult care. There was no age limit as long as focus was on the actual transfer process and participants had a chronic disease. Data were analyzed into metasummaries and metasynthesis following Sandelowski and Barroso’s synthesis of qualitative research.; RESULTS Transition from pediatric to adult ward was characterized by experiences of loss of familiar surroundings and relationships combined with insecurity and a feeling of being unprepared for what was ahead. Four subthemes were identified: facing changes in significant relationships, moving from familiar to unknown, timing readiness to transfer and achieving ownership of disorder and self.; CONCLUSIONS Transition experiences seem to be commensurable across diagnoses and cultures. Increased focus on adolescents’ and young adults’ need for support and involvement when they are in limbo between pediatric and adult care could improve their quality of life and protect them from additional health problems in a vulnerable phase.
(1003) The effects of survey administration modality on the assessment of child reported health Katherine Bevans, PhD, Children’s Hospital of Philadelphia; Ramya Pratiwadi, Children’s Hospital of Philadelphia; Christopher B. Forrest, MD, PhD, University of Pennsylvania, School of Medicine, United States AIMS Computerized administration of patient reported outcomes (PROs) has become commonplace, yet few studies have contrasted the effects of this mode-of-administration with conventional paperand-pencil, particularly for children. In this study, we present a systematic approach for evaluating mode-of-administration effects using both classical test and modern measurement methods and apply this to a study on pediatric PRO assessment.; METHODS 10,320 youth aged 8–17 years were administered one or more pediatric PROMIS item banks that included family belonging, positive affect, mobility, or psychological stress response—2,363 to 2,558 youth/bank. Of participating youth, 35 % (n = 3,644) completed a paper/pencil version of the measure at school. Remaining youth (n = 6,676, 65 %) completed an internet-based questionnaire. We evaluated between-group differences in missingness, scale scores, internal consistency, and item bank dimensionality, and used IRT to identify differential item functioning (DIF) by data collection modality.; RESULTS Youth in the school sample were older (school: M = 13.5, SD = 2.0; Internet: M = 11.5, SD = 3.1). Item missingness was infrequent, but greater for the school sample (1.7 %) than the Internet sample (0.5 %). After controlling for age, youth from schools were less physically active and had higher levels of psychological stress symptoms. For all banks, estimates of internal consistency were high (r [ .95) and comparable between groups. One-factor CFA model fit statistics supported item bank unidimensionality for both groups. Associations between factor loadings for the school and internet samples were high (FamBel: r = .87; PosAff: r = .86; Mob: r = .75; Psych: r = .66). Betweengroup comparisons of item information curves revealed similarities in the degree to which items discriminated among youth with varying levels of health. Only a single item was found to exhibit modality-based DIF to the degree that it significantly affected the item’s true score (mobility during recess).; CONCLUSIONS As target populations and endpoints are expanded and diversified in clinical research, so must the approaches for directly assessing patients’ health experiences. Yet, caution must be applied in combining PRO data when it is obtained from varied sources. Applying a rigorous evaluation of data collection modality effects, we found no substantive differences in paper-and-pencil versus internet-based survey administration for children.
123
Qual Life Res (2012) 21:1–132 (1004) Qualitative development of child-report family relationship item banks Anne Riley, Johns Hopkins University, United States; Jeanne Landgraf, MA, HealthAct CHQ Inc.; Rachel E. Teneralli, MS, RD, Children’s Hospital of Philadelphia, United States; Barbara H. Fiese, University of Illinois at Champaign Urbana; Lisa J. Meltzer, PhD, National Jewish Health; Katherine Bevans, PhD, Children’s Hospital of Philadelphia; Anna K. Ettinger, Johns Hopkins Bloomberg School of Public Health; Christopher B. Forrest, MD, PhD, University of Pennsylvania, School of Medicine, United States AIMS To describe the qualitative development of two pediatric item pools designed to assess family relationships from the perspective of children 8–17 years old. Carried out as a PROMIS initiative, Family Relationships are domains in Social Relationships in Social Health.; METHODS Three phases were involved: (1) Domain/item bank concept specification—a literature review and interviews with 10 content experts, 24 children, and 7 parents; (2) Item Concept Map Formation and Item Library Formation— review of the measurement literature on family relationships; identification of instruments; definition and classification of item concepts; specification of each domain concept map; writing of new items to fill gaps in concept maps. (3) Comprehensibility Assessment—3 sub-studies: Cognitive debriefing interviews—5 children per item (total n = 43); Translatability reviews by 4 professionals; and Flesch-Kincaid Reading Level analysis.; RESULTS Defined two positively oriented domain concepts, children’s sense of Family Belonging, the deeply experienced understanding, feelings and beliefs about being part of one’s family–feeling valued, accepted, and cared for and knowing one can depend on the family; Family Involvement, the predictability, frequency, engagement and enjoyment of interactions with parents and other family members; the mechanism through which children develop, deepen, and elaborate their sense of family belonging. Identified 106 family instruments, resulting in two large item pools which were tested in Phase 3. Average comprehension score of 2.9 (s.d. 0.21) on a scale of 1–3 in cognitive debriefing; identified items with poor comprehension. Translatability review identified problems such as vague words, need for an active voice. Problems were addressed, resulting in 2 item pools: Family Belonging (55 items) and Family Involvement (58 items).; CONCLUSIONS The two item pools are ready for large-scale administration in order to calibrate them for administration in computerized adaptive test format. As with all PROMIS tools they will be readily available to researchers. Expected by 2013. Once calibrated and validated these items banks will allow description and testing of the nature of family relationships as aspects of health.
(1006) New card system quality of life (QOL) Questionnaire For Children (Part 2) Rika Hayashida, MS, Siebold Univ of Nagasaki, Japan; Megumi Masuyama, Nagasaki University Hospital, Japan; Aya Sakai, Isahaya City, Nagasaki, Japan; Michiko Kobayashi, MD PhD, Kwassui Women’s College, Japan; Takashi Mandai, MD, Japanese Society of Quality of Life Research, Japan AIMS The purpose of this study was to develop an original new card system style of QOL questionnaire for children.; METHODS Onehundred-four children, age six and under, participated in this study. 24 six-year-olds, 48 five-year-olds, and 32 four-year-olds, respectively. The former QOL questionnaire used in 2010 (18 questions divided into 5 categories) had an insufficient reliability and validity for clinical use on under 4 years of age children. Therefore we have developed an original new card system style of QOL questionnaire
Qual Life Res (2012) 21:1–132 consisting of 22 questions (using 22 picture cards) divided into 5 categories.; RESULTS The Cronbach’s alpha coefficient of our questionnaire was excellent enough to accept for clinical use for children 6 years of age and under: 0.70 in social interaction, 0.65 in playing time, 0.62 in eating time etc. (6 year-olds), 0.75 in social interaction, 0.66 in family interaction, 0.58 in everyday life etc. (5 year-olds), and 0.61 in family interaction, 0.49 social interaction, 0.45 in playing time etc. (4 year-olds), respectively. Our original new card system style of QOL questionnaire (using 22 picture cards) contained 8 main factors which matched the 5 categories (6 yearolds), 6 main factors which matched the 5 categories (5 year-olds), and 7 main factors which matched the 5 categories (4 year-olds). Their cumulative contributions were 0.85 in 6 year-old, 0.57 in 5 year-old and 0.77 in 4 year-old children, respectively. There was significant positive correlation between family interaction and social interaction (r = 0.60–0.76, p \ 0.01). For the excellent QOL under 6-year-old children, family interaction and playing time were most important for children.; CONCLUSIONS These findings indicate that our original new card system style of QOL questionnaire has a sufficient reliability and potency of validity to use for over 5-year-old children, and it has somewhat sufficient reliability and potency of validity to use for 4-year-old children. In the future, we must pay more attention to the age of children, testing time and place, and the method of questioning. And, it will be important to consider the many kinds of problems that children and parents face and how parents take care of their children in various developmental stages.
(1007) Psychometric properties of the KIDDY-KINDL for early kindergarten populations Ester Villalonga, PhD, Medical Psychology and Medical Sociology, Universitity Medical Centre, Go¨ttingen, Germany; Claudia Witte, University Medical Centre, Germany; Erich Kasten, Medical Psychology and Medical Sociology, Universitity Medical Centre, Go¨ttingen; Christiane Kiese-Himmel, Medical Psychology and Medical Sociology, Universitity Medical Centre, Go¨ttingen; J.S. Gaier, Medical Psychology and Medical Sociology, Universitity Medical Centre, Go¨ttingen; Nicole Von Steinbuchel, PhD, University of Goettingen, Germany AIMS Few instruments exist to assess HRQoL in children and adolescents. However, there are no questionnaires yet validated for early kindergarten populations. Our aim was to investigate psychometric properties of the KIDDY–KINDL (KK) in German speaking children with the age of 3–5 years from an immigrant and native German speaking socio-economically weaker background.; METHODS 166 children from five catholic kindergartens in and around Frankfurt/ Main participated in this cross-sectional study. HRQoL was selfreported by the children in an interview format using the KK. We measured internal consistency, convergent and discriminative validity using comparisons f.e. regarding health status and relation with peers and family.; RESULTS Psychometric characteristics of the selfreported KK showed a Cronbach’s alpha concerning general HRQoL of 0.71 and 0.65 for disease-specific HRQoL. Construct and convergent validity ranged from 0.47 to 0.54 (p \ 0.05) in items related with felling ill and experiencing symptoms, and to have fun and good relations with peers and family.; CONCLUSIONS This is the first study to show that the KK is—with moderate to good psychometric properties—a suitable instrument to measure self-reported HRQoL in 3 year old children with native German and immigrant background; however the dimension ‘‘feeling proud of one self’’ should be taken out. This concept does seem too difficult to be understand in 3 year olds. Our preliminary results should be reproduced in representative German samples. Funding: NIFBE (NMK), v. Metzler & HertieStiftung.
45 (1008) Headache and health related quality of life among Norwegian children and adolescents Kristin Haraldstad, PhD, University of Agder, Norway; Sølvi Helseth, Oslo and Akershus University College AIMS This study is part of the Norwegian Kidscreen study. The overall aim of the study was to investigate HRQOL in a sample of Norwegian children and adolescents. The specific aim of this part of the study is to investigate health-related quality of life in a school sample of children and adolescents aged 8–18 years, and to examine the relationship between health-related quality of life, and headache. Previous studies have identified pain as a significant factor associated with poor HRQOL in children and adolescents; METHODS A cross-sectional study was designed and data were collected among children in the Eastern part of Norway with around 1.7 million inhabitants (approximately 36 % of the total Norwegian population), and a child population (8–18 years) of around 230,000. Included in the study were 1,675 children and adolescents aged 8–18 years. HRQOL was measured with the Kidscreen-52 instrument. Pain was measured with Lubeck Pain questionnaire; RESULTS Eligible for analysis in this part study is 1,238 children/adolescents. 60 % of the children reported pain during the previous 3 months, and both boys and girls reported headache as the most common pain. Among the total sample, the prevalence of headache was 34 %, and 12 % reported chronic headache. Girls aged 16–18 years reported the most headaches. The children who experienced headache reported significantly lower HRQOL than did children without pain. The dimensions physical activity, school and self-perception were most affected.; CONCLUSIONS Headache is a common problem in children and should be regarded as a significant health problem. The experience of pain has negative effects on many aspects of the lives of children, and children with headache reported significantly lower HRQOL than children without headache measured with KIDSCREEN52 questionnaires. Data will be further analyzed for presentation at the ISOQOL conference in Budapest.
(1009) Health-related quality of life and self-worth in 10-year old children with congenital hypothyroidism diagnosed by neonatal screening Liesbeth Van Der Sluijs Veer, Emma Childrens Hospital; Marlies Kempers, University of Nijmegen; Tom Vulsma, Emma Childrens Hospital; Bob Last, Emma Childrens Hospital; Martha Grootenhuis, PhD, Academic Medical Center, Netherlands AIMS Although much is written about cognitive and motor development, little is known about emotional and social consequences for children growing up with CH. The objectives of the study were: (1) to compare health related quality of life (HRQoL) and selfworth of 10 year old CH patients with the general population; and (2) to explore whether severity of CH was related to these outcomes.; METHODS A total of 82 10-year old children with CH and their parents completed several questionnaires. Patients were classified to subgroups ‘severe CH, n = 41’ or ‘moderate/mild CH, n = 41’ based on their pre-treatment FT4 concentration. Differences between children with CH and the general population were analyzed by analysis of variance and one sample t-tests. To create a clinically meaningful distinction between children that can be considered ‘at risk’ or ‘not at risk for problems, two groups were formed and percentages at risk were compared using Chi-square tests and binomial tests.; RESULTS CH patients reported lower HRQoL on several domains (cognitive functioning; motor functioning; autonomy; social functioning; positive moods) and are more often at risk for HRQoL impairment and impaired self-worth on two scales of the CBSK (school competence and athletic
123
46 competence). There were no significant differences between the severity groups for HRQoL and self-worth.; CONCLUSIONS Negative consequences in terms of HRQoL and self-worth are prevalent in children with CH. Health care physicians should to be attentive to these consequences and provide attention and supportive care for these children if necessary.
(1010) Assessing the impact of brain cancer in caregivers’ quality of life (QoL) Patricia Minaya Flores, MA, Laboratoire de Sante Publique EA3279, France; Olivier Chinot, Timone University Hospital, Department of Neurooncology, Marseille; Julie Berbis, Department of Public Health, Medicine University, France; Pascal Auquier, PhD, Laboratoire de Sante Publique, France AIMS The aim of the study was to analyze the impact of brain cancer in caregivers QoL and to compare this specific population to other oncology caregivers and to normative population in order to find differences and understand which aspects of QoL are more impacted.; METHODS The sample was composed of caregivers of patients with brain cancer from the Neuro-Oncology department of Timone University Hospital of Marseille, France. Control caregivers were selected from different oncology departments and were matched with caregivers of patients with brain cancer on age, sex and relationship with the patients (spouses, son/daughter, parent or sibling). We used the specific CareGiver Oncology Quality of Life questionnaire (CarGOQoL), a self-administered instrument comprising 29 items which is based on the exclusive point of view of caregivers to assess the impact of cancer and its treatment on caregivers’ QoL. Caregivers also completed the Short Form 36 (SF36) for comparison to the French normative sample (a national representative sample of 3,617 healthy French adults).; RESULTS The study sample included 50 caregivers of patients with brain cancer, aged 30–77 years, 28 % of whom were men. Comparing specific CarGOQoL scores, significant differences were found for Burden and Leisure Time dimensions between cases and controls. Brain cancer caregivers had significantly lower scores for both dimensions, having an effect size of 0.4. Using SF36, no significant differences between cases and controls were observed. Compared to French age- and sex-matched normal controls, caregivers of patients with brain cancer had significantly lower mean scores for Social Functioning, Role limitation due to Physical health, Role limitations due to Emotional problems, Mental Health, Vitality and Bodily Pain.; CONCLUSIONS Caregivers of patients with brain cancer showed increased burden and lower scores for leisure time dimension. This could be explained by their unique care situation where patients become more limited physically and cognitively. Often changes in cognitive abilities, personality and behavior of the patient impact negatively in caregivers’ QoL, changes that are less frequently presented in other cancers.
(1011) Physical and mental health of family caregivers: findings from 60,569 adults in Thailand Vasoontara Yiengprugsawan, The Australian National University, Australia; Sam-ang Seubsman, Sukhothai Thammathirat Open University; Adrian Sleigh, The Australian National University AIMS Increased longevity and population ageing accompany socioeconomic development, now emerging rapidly in many parts of the world. As people live longer the demands for family caregivers increase. This study investigates prevalence and health
123
Qual Life Res (2012) 21:1–132 outcomes among adult caregivers in Thailand.; METHODS This study derive from the 2009 self-administered questionnaire completed by long-distance Open University adult students residing throughout Thailand (n = 60,569). We report covariate-Adjusted Odds Ratios (AOR) for the association between caregivers and their health outcomes. Covariates adjusted were age, marital status, household income, work status, urban–rural residence, health-risk behaviours (smoking and drinking), and body mass index. Analyses were reported separately by sex.; RESULTS Among respondents in 2009, 27.5 % reported being part-time and 6.6 % reported being full-time caregivers. Caregivers tended to be older age, married, unpaid family members, and rural residents. Caregivers were more likely to engage in health-risk behaviours such as smoking and drinking, and also reported higher body mass index compared to non caregivers. Among males, being a part- or full-time caregiver was associated with lower back pain (AOR 1.36 and 1.67) and poor psychological health (AOR = 1.16 and 1.68). Among females, being a part- or full-time caregiver was associated with poor selfassessed health (AOR = 1.21 and 1.34), lower back pain (AOR = 1.47 and 1.84) and psychological distress (AOR = 1.32 and 1.52).; CONCLUSIONS Overall, the magnitudes of the family caregiver effect on health were not large but were consistent across health outcomes and thus indicate a considerable burden imposed on caregivers. Our findings raise awareness and facilitate early identification of vulnerable family caregivers. This information will be useful for health, social service providers, and related policymakers.
(1012) Impact of culture on the quality of life of Mexican–American family caregivers of patients with advanced cancer Joan J. Branin, PhD, University of La Verne; Gloria Juarez, PhD, RN, MSN, City of Hope AIMS Although Hispanics are the fastest growing ethnic group in the United States and cancer the second leading cause of death among Hispanics, few studies have documented the impact of cancer diagnosis on the quality of life and caregiving experiences of Hispanic family caregivers. The influence of culture has been seen as important in the development and delivery of culturally congruent care and support services for cancer patients and their caregivers by healthcare and community-based providers. The aim of this study was to explore the influence of culture on the caregiving experiences and perceptions of QOL among Mexican–American (MA) family caregivers of adults with advanced cancer.; METHODS This qualitative descriptive study was conducted in the ambulatory care setting of an NCI-Comprehensive cancer care center in the United States. Twenty MA family caregivers of patients with advanced cancer participated in qualitative interviews. Thematic analysis of the interviews and the conceptual model of Quality of Life Well-being developed by Ferrell, Wisdom, and Wenzel (1989) were used to characterize the caregiving experience and the effects of caregiving on QOL of MA cancer family caregivers.; RESULTS In general, the caregiving experience and the meaning of QOL of these participants centered on spiritual or religious beliefs and practices, concern for providing the best home care, support of extended family, and the desire to fulfill cultural role and responsibility expectations. Overall, MA caregivers emphasized that their caregiving experience had more positive than negative outcomes. The caregiver’s overall QOL was primarily affected by the patient’s advanced disease, the consequences of the cancer treatment and its related side effects, and the patient’s emotional/psychological needs. Secondarily, the caregiver’s overall perception of quality of life was affected by their physical, social, emotional and spiritual needs.; CONCLUSIONS These findings provide insight into the
Qual Life Res (2012) 21:1–132 cultural dimension inherent in the family caregiving experience and perceptions of QOL of MA cancer family caregivers. It is imperative to recognize the influence of culture on the overall QOL and caregiving experiences of MA family caregivers in developing culturally relevant interventions specifically designed to reduce the burden of cancer and improve overall caregiver QOL.
(1013) Development of 9 item quality of life assessment scale for caregivers of children with atopic dermatitis in Japan Yukihiro Ohya, MD PhD, Nat’l Ctr for Child Hlth & Dvlpmnt, Japan; Toshio Katsunuma, The Jikei University School of Medicine AIMS QPCAD (Quality of life in Primary Caregivers of children with Atopic Dermatitis) is a questionnaire with 19 items and 4 subscales, which was developed in Japan to evaluate the Quality of life(QOL) of primary caregivers of children with Atopic Dermatitis (AD). We developed a short version of QPCAD (named as QP9) to reduce the burden of respondents when it will be applied in clinical practice.; METHODS In the first part, we applied factor analysis to the data of QPCAD answered by 529 primary caregivers of AD children at the age of 0–6, and estimated the factor structure consisting of 4 subscale factors and 19-item questions. Nine item questions were selected based on the factor structure of QPCAD and named as QP9. In the second part, we applied QP9 to 148 participants of this study to test the validity and reliability of this QOL assessment scale for caregivers of AD children.; RESULTS Children whose caregivers joined this study was male dominant (62.8 %) and their caregivers were mostly mothers (96.6 %). One hundred and thirty eight children of 148 participants experienced topical steroids therapy before their first visit to the authors. Total score of QP9 and severity of AD children evaluated by using SCORAD system was positively associated. Test retest reliability of each QP9 item collected from 141 participants showed high weighted k from 0.58 to 0.71. Therapeutic responsiveness was evaluated before and after our treatment (at the first time visit to outpatient unit and at the second time visit). Improved severity scores (SCORAD, Investigator general assessment, itchiness VAS, and insomnia VAS) was positively associated with the improvement of QP9 scores.; CONCLUSIONS Newly developed QP9 showed satisfactory features as a QOL assessment scale for caregivers of children with AD as a short version of QPCAD.
(1014) The role of metacognition and reflexivity in QoL: a philosophical analysis Julien Blanc, Aix-Marseille University, France; Laurent Boyer, PhD, Aix-Marseille University, France; Pascal Auquier, PhD, Laboratoire de Sante Publique, France; Pierre Le Coz, Aix-Marseille Universite´, France AIMS The concept of QoL (quality of life) in research applied to health seems to be characterized by the lack of definition agreement. However, an implicit general tendency to define QoL as ‘‘subjective’’ is detectable. ‘‘Subjective’’ refers to the idea that QoL is function to perception from within, and that the quality of a life depends on what an individual sees as valuable. This approach underlies QoL instruments based on the point of view of the patient (what he declares as important) and his perceptions. Some of its limits are explored and a alternative approach is proposed.; METHODS Philosophical analysis provides resources for such a critic. An alternative approach based on the non-subjective QoL theory of the philosopher J. Griffin is proposed; RESULTS Recent findings in QoL research for patients with schizophrenia show an inverse relation between QoL results and metacognitive capacities. These findings reveal the limits of the
47 approach to QoL based on the point of view and perception: it makes no difference between comfortable delusion experienced by the patient with metacognitive impairment, and bitter truth faced by the patient not impaired. Moreover, the QoL of the patient not impaired is evaluated on the criteria of the impaired patient: pleasantness of mental states and/or the level of satisfaction. These criteria are seen as paradigmatic. However, there is a significant qualitative difference between those two kinds of patients, that we explain in terms of ‘‘reflexivity’’, a philosophical concept equivalent to metacognition. We show, in reference to the work of J. Griffin, that reflexivity is a formal condition to QoL as it plays a determinant role in the evaluative capacities of the individual. QoL should be considered then as paradigmatically constituted by reflexive values not by bare pleasantness or satisfaction.; CONCLUSIONS The concept of QoL underlying QoL instruments should reflect the role of reflexivity. We suggest conceptual tools and perspectives for the qualitative development of QoL instruments coherent with our conceptual proposal.
(1015) Effect of facial emotion recognition on subjective psychosocial functioning in bipolar patients Omer Aydemir, Celal Bayar University, Turkey; Cengiz Akkaya, Uludag University Dept. of Psychiatry Bursa Turkey, Turkey; Burak Uykur, Celal Bayar Uni. Dept. of Psychiatry Manisa Turkey; Almila Erol, Izmir Ataturk Training Hospital Dept. of Psychiatry Izmir Turkey AIMS Facial emotion perception and recognition have been consistently demonstrated to be impaired in bipolar patients. The aim of the study is to investigate the relationship of facial emotion recognition along with clinical features and mood symptoms with subjective psychosocial functioning in bipolar patients.; METHODS The mean age of the bipolar patients was 39.7 and 58.7 % were female. Duration of the illness was 13.2 years with a mean of 6.8 episodes. The study was carried out with a total of 109 patients with bipolar disorder, 19 patients with major depressive episode, 20 patients with manic episode and 70 patients in remission, and 51 healthy controls. In the assessment of psychosocial functioning, Bipolar Disorder Functioning Questionnaire (BDFQ) was administered to the patients. The subscales of BDFQ are emotional functioning, intellectual functioning, sexual functioning, sense of stigmatization, social withdrawal, household activities, relations with friends, participation to social activities, daily activities and hobbies, taking initiative and self-sufficiency, and occupation. For the evaluation of facial emotion recognition, Facial Emotion Identification (FEI) and Facial Emotion Discrimination (FED) Tests were used. Each of the two tests contains six emotions: happiness, surprise, fear, sadness, shame, and anger. Regression analyses were performed.; RESULTS On all of the eleven domains of BDFQ, HAM-D score had significantly negative impact. YMRS score significantly affected intellectual functioning. Facial emotion identification had significantly positive impact on taking initiative in routine daily life. Having more previous episodes had negative impact on occupation. Facial emotion discrimination had positive but non significant (p = 0.051) impact on social activities.; CONCLUSIONS Impairment in facial emotion recognition seems to affect social aspects of daily life of bipolar patients.
(1016) Interpretation of response categories in patient-reported rating scales: a test–retest study among people with neurological disorders Peter Hagell, Kristianstad University, Sweden; Ida Knutsson, Lund University; Karin Samuelson, Lund University AIMS Unambiguous use and interpretation of rating scale data assumes that response categories are interpreted and work as intended. This
123
48 study investigated the stability of interpretations of commonly used patient-reported rating scale response categories among people with neurological disorders.; METHODS Forty-six people with neurological disorders (26 men; mean age, 57; Parkinson’s disease, 50 %; multiple sclerosis, 41 %) indicated their interpretation of 21 response categories (representing frequencies, intensities and levels of agreement) on 100-mm visual analog scales (VAS) on two occasions, = 2 weeks apart. Data were analyzed using intraclass correlation (ICC; should be [0.4), mean (95 % CI) differences, and the standard error of measurement (SEM).; RESULTS Fourteen ICC values were \0.4 (overall mean ICC, 0.279; frequencies: mean ICC, 0.224; intensities: mean ICC, 0.265; levels of agreement: mean ICC, 0.362). There were no systematic differences in VAS values between time points according to 95 % CIs. The mean difference across all 21 response categories was 0.43 mm (mean absolute difference, 3.36 mm). The overall mean SEM was 16.1 (range, 12.3–20.1) mm (frequencies, 17.8 mm; intensities, 14.8 mm; levels of agreement, 15.3 mm). The average proportions of absolute difference scores exceeding the SEM were 32.2 % (frequencies), 36.3 % (intensity), and 29.8 % (agreement).; CONCLUSIONS Stabilities in interpretations of patient-reported rating scale response categories were generally low. Categories expressing levels of agreement showed best reproducibility, suggesting that these may be preferable when appropriate with respect to the scale and its items. Future studies should consider response category interpretations in relation to various contexts. These observations suggest caution when interpreting raw rating scale data and argue for the use of modern rating scale methodologies such as the Rasch measurement model.
(1018) Inferior quality of life and psychological well-being in immigrant cancer survivors Phyllis Butow, PhD, University of Sydney, School of Psychology, Australia; Lynley Aldridge, University of Sydney, Australia; Melanie Bell, PhD, MSc, AB, University of Sydney, Australia; Ming Sze, University of Sydney; Maurice Eisenbruch, Monash University; Michael Jefford, Peter MacCallum Cancer Centre; Penelope Schofield, Peter MacCallum Cancer Centre; Afaf Girgis, University of New South Wales; David Goldstein, University of New South Wales; Madeleine T. King, PhD, Psycho-oncology Cooperative Research Group (PoCoG), Australia AIMS Immigrants with cancer are known to have poorer cancer outcomes. Social suffering, language difficulties and cultural factors may all contribute. This study aimed to compare quality of life (QoL) and psychological morbidity, and variables associated with these, in a population-based sample of first generation immigrants and AngloAustralians who had survived cancer.; METHODS Participants were recruited via Cancer Registries in Australia. Eligible cancer survivors had a new diagnosis of one of 12 most incident cancers (all stages) 1–6 years earlier, were aged 18–80 years at diagnosis and had a treating doctor who approved approach. Eligible immigrant participants and both parents were born in a country where Chinese, Arabic, or Greek is the primary language spoken and they spoke one of these languages. A random sample of English-speaking Anglo-Australian born controls frequency matched for cancer diagnosis was recruited. QoL was assessed with the Functional Assessment of Cancer Therapy—General (FACT-G) and psychological morbidity was assessed with the Hospital Anxiety and Depression Scale (HADS). Multiple linear regression was used to adjust for sociodemgraphic variables.; RESULTS 596 patients of whom 277 were immigrants were recruited to the study. Since 20 % of data on cancer stage was missing, stage was excluded from analyses (but estimates for impact of immigrancy were unaffected when stage was included in sensitivity analyses). After adjusting for age, sex, education, marital status, socio-economic status, time since diagnosis and type of cancer, immigrants had clinically significantly worse QoL
123
Qual Life Res (2012) 21:1–132 (5.4–8.5 FACT-G points, p \ 0.0001) and higher depression (p \ 0.0001) than Anglo-Australians. Understanding the health system partially mediated the relationship between immigrant status and depression (p = 0.0004) and QoL (p = 0.001), accounting for 55 % of the total effect for depression, and 25 % for QoL. Participants with confidence understanding English had higher QoL by 5.0 points (p = 0.06).; CONCLUSIONS Immigrants who are survivors of cancer, several years after initial diagnosis, have worse psychological and QoL outcomes compared with Anglo Australians. Potential targets for intervention to reduce disparities include assistance in navigating the health system, translated information and cultural competency training for health professionals.
(1019) PHQ-9 use in clinical practice Thomas E. Elliott, MD, Essentia Institute of Rural Health; Colleen M. Renier, BS, Essentia Institute of Rural Health; Jeanette A. Palcher, Essentia Institute of Rural Health AIMS To discover the prevalence of PHQ-9 use in primary care provider (PCP) clinical practice over a 7-year timeframe. To determine the effect of PHQ-9 data on PCP diagnosing clinical depression and prescribing antidepressant drugs in clinical practice.; METHODS We conducted a retrospective case series study analyzing data in the electronic medical records of a large healthcare system in the U.S. from 01/01/2005 to 12/31/2011. All patients age = 18 managed by a PCP were included. Data collected: patient demographics, provider descriptors, PHQ-9 data, ICD-9CM depression codes, and antidepressant drugs. Analytical tests included Cochran-Armitage test for trend, Cochran-Mantel–Haenszel test and logistic regression modeling for year of maximum PHQ-9 score. PHQ-9 tests were considered independent only if they were: initial events; 10+ months after the most recent test; or immediately following a prior test with a total score of less than 5(normal). The logistic regression model included any depression diagnosis (DEP), year, sex, age, maximum PHQ-9 score-5 categories from Normal to Severe Depression, any PCP measured PHQ-9, and any behavioral health measured PHQ-9.; RESULTS Study population included 294 PCP (*181 annually) and 216,494 patients (*105,240 annually). The measurement of independent PHQ-9 in patients managed by primary care providers increased from 0.7 % (2005) to 13.5 % (2011); Z = 153.34, p \ 0.001. Patients with an independent PHQ-9 test were 17.87 (95 % CI: 17.43–18.31) times more likely to have a diagnosis of depression. After adjustment in the logistic model, patients with DEP were more likely to be prescribed an antidepressant drug, rate ratio = 3.56 (3.33–3.80). As PHQ-9 scores increased, associated rate ratios (compared to normal PHQ-9) increased incrementally, 1.26 (1.82–1.35), 2.07 (1.90–2.25), 3.05 (2.79–3.34), and 4.18 (3.72– 4.70), for PHQ-9 scores 5–9, 10–14, 15–19 and 20+, respectively.; CONCLUSIONS Our study found PHQ-9 use in primary care clinical practice increased significantly over 7-years. Patients having PHQ-9 testing in primary care were more likely to have a diagnosis of depression. Patients with increased severity of PHQ-9 scores were substantially more likely to be prescribed an antidepressant drug. PHQ-9 data appear to increase the diagnosis of clinical depression and prescribing of antidepressant drugs in primary care.
(1020) Accentuate the positive: coping with the burden of treatment and complex self-care David T. Eton, PhD, Mayo Clinic, United States; Djenane R. Oliveira, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil, Brazil; Jason S. Egginton, Mayo Clinic, Rochester, MN USA;
Qual Life Res (2012) 21:1–132 Jennifer L. Ridgeway, Mayo Clinic, Rochester, MN USA; Laura Odell, Mayo Clinic, United States; Carl R. May, University of Southampton, United Kingdom; Victor M. Montori, Mayo Clinic, Rochester, MN USA AIMS Burden of treatment (BoT), or the workload of healthcare and its impact on patient functioning and well-being, can negatively influence adherence to treatment, clinical outcomes, and quality of life. There are however ways to alleviate BoT, and thereby lower the potential for adverse outcomes. In our work to understand the meaning of BoT, we explored how patients attempt to cope with it.; METHODS We conducted semi-structured, qualitative interviews with 32 patients (62 % response rate) seeking medication therapy management services at a large academic medical center. All patients had a complex regimen of self-care and were living with one or more chronic health conditions. Three trained interviewers conducted the interviews which were audiotaped and transcribed. We used framework analysis to identify and code themes independently. Codes were harmonized via discussion and consensus.; RESULTS Twenty women and 12 men were interviewed (age range 26–85 years). Number of health conditions ranged from 1 to 16 (median = 5) and all patients were taking several medications daily. Patients reported making use of social, personal and technological resources to cope with and mitigate BoT. Themes emerging from the interviews included (1) social support from family, friends, and coworkers, (2) relationships and communication with providers, (3) problem-focused strategies (e.g., preparing for medical appointments, seeking information), (4) emotion-focused strategies (e.g., focusing on jobs/hobbies/family, positive reframing, social comparisons to others, spirituality/meditation), (5) adaptation and normalization of routine, and (6) virtual connections to providers (e.g., e-mail).; CONCLUSIONS Maintaining a complex regimen of treatment and self-care is a daily challenge. Some patients call upon social and personal resources to cope with the challenges imposed by complex care. Resilient patients may find care to be more manageable and easier to adhere to. Ultimately, a balanced clinical approach of reducing patient workload and teaching patients to use multiple coping strategies could help solve the problem of treatment burden.
(1021) Health related quality of life: concerns about clinical practice integration and clinical utility Augusta Silveira, University Fernando Pessoa, Portugal; Joaquim Gonc¸alves, Polytechnic Institute of Ca´vado & Ave, Portugal; Ana Maria Almeida, Instituto Portugueˆs de Oncologia, Portugal; Cla´udia C. Ribeiro, Universidade Cato´lica Portuguesa, Portugal; Lurdes Carvalho, Portuguese Oncology Institute, Porto, Portugal; Eurico Monteiro, Portuguese Oncology Institute, Portugal; Francisco Luis Pimentel, MD PhD, Universidade de Aveiro, Portugal; Teresa Sequeira, Portugal AIMS Health Related Quality of Life (HRQoL) is a multidimensional and subjective concept and considered a major issue in the assessment of patient-reported outcomes in oncology. Both informatics technology and a Knowledge Management System constitute an essential demand for data integration into clinical practice, improving clinical research methodology. The incorporation process of HRQoL measurements into routine clinical practice is described and healthcare provider’s perspectives are discussed.; METHODS Routine HRQoL assessment was implemented in the Portuguese Oncology Institute-Porto. Methodology was previously approved by the local Ethics Committee and involved the Otorhinolaringology and Head & Neck Oncology Units. All patients gave their informed consent. The Portuguese version of EORTC QLQ-C30 and EORTC
49 QLQ-H&N35 PC software questionnaires were used. Results were analyzed in real time and translated in a graphic form. Individual cumulative printouts were automatically generated including results of previous questionnaires assessments; unfavorable scores were graphically highlighted. For 3 months, an individual ‘‘graphic report’’ (n = 178) was given to physicians and nurses before consultation. Healthcare professionals’ considerations concerning measurements incorporation into routine clinical practice were assessed by semi-structured interviews and by a checklist questionnaire administered after each consultation.; RESULTS Consultation time was not affected by the incorporation measure into routine clinical practice. The graphical report appearance was appreciated and its information was considered relevant. Physicians found the information particularly useful and relevant (p value = 0,028) when patients with low HRQoL scores were considered and considered it particularly helpful for patient-physician communication (p value = 0,031). The ‘‘time since diagnosis’’ parameter did influence the interest reported and when long-time cancer survivors were considered physicians identified the HRQoL report as most useful. Considering the first year after diagnosis, a peak of interest was reported by physicians 3 months after treatment (40 %). The information was then used in 38 % from all assessments, being more solicited again in the fourth and fifth years (50 %).; CONCLUSIONS HRQoL implementation and measurements incorporation into routine clinical practice are feasible and practicable. This study was among the firsts to adopt the complete procedure of routine HRQoL implementation in clinical practice. Attitudes of healthcare team were positive and the barriers encountered towards successful implementation were subsequently resolved.
(1022) Communicating clinical and quality of life outcomes to patients before surgery for oesophageal cancer: what’s most important? Rhiannon C. Macefield, University of Bristol, United Kingdom; Natalie Blencowe, University of Bristol; Marc Jacobs, MSc, Amsterdam Medical Center, Netherlands; Sara Brookes, University of Bristol; Mirjam AG Sprangers, PhD, Academic Medical Center, Netherlands; Jane Blazeby, MD, University of Bristol, United Kingdom AIMS Surgeons talk to oesophageal cancer patients before operating on them, but little is known about what is routinely discussed. Disclosure of information informs decision-making and should include clinical and quality of life (QOL) issues, however, standards for information provision are lacking. Consensus on a core information set to be communicated as a minimum in pre-surgical consultations is needed. The current study aimed to identify all outcomes for discussion and compare patients’ and health care professionals’ (HCPs) views on their relative importance.; METHODS Clinical and QOL domains were identified from systematic literature reviews, hospital patient information leaflets and qualitative analyses of pre-surgical consultations. Patients and HCPs rated the importance of discussing each domain prior to surgery, on a scale of 1 (not essential) to 9 (absolutely essential) in a postal questionnaire. Ratings were categorised as (a) low importance, 1–3, (b) moderate importance, 4–6 and (c) high importance, 7–9. The proportion of patients’ and HCPs’ responses in each group were compared for all information domains. Domains of high importance to the majority (rated high by [50 % and low by \15 % of respondents) were examined.; RESULTS Some 66 domains were identified (34 clinical and 32 QOL) and 183 patients and 123 HCPs completed questionnaires. Anastomotic leak, in-hospital mortality and type of surgery were rated high by the greatest proportion of HCPs ([90 %), followed by information about
123
50 immediate post-operative care and QOL. Significantly fewer patients rated these domains as high, with the exception of QOL (rated high by 152/180 patients compared to 108/122 HCPs, p = 0.315). Instead, patients considered information about disease recurrence as most important (rated high by 86 %), followed by survival, general health, QOL and fitness for surgery. Domains qualifying as highly important to the majority of patients were predominantly QOL issues (24/44 domains), whereas for HCPs these were mostly clinical (25/37 domains, p = 0.046).; CONCLUSIONS Patients and HCPs have differing views on the relative importance of clinical and QOL information, although both consider QOL an essential outcome to discuss prior to oesophageal cancer surgery. These findings will inform the development of a core information set and emphasise the need for input from all stakeholders.
(1023) Pediatrician’s capability, willingness, and barriers of measuring patient-reported outcomes in clinical practice: a national survey in the United States I-Chan Huang, PhD, University of Florida; Onyekachukwu Osakwe, University of Florida; Lindsay Thompson, University of Florida; Darren DeWalt, MD MPH, University of North Carolina; Dennis Revicki, PhD, United BioSource Corporation; Elizabeth Shenkman, PhD, Univ of Florida College of Medicine AIMS Little is known about the use of patient-reported outcomes (PROs) measures by pediatricians in clinical practice. This study aimed to examine general pediatricians and pediatric subspecialists’ capability, willingness, and barriers of using pediatric PROs measures in the United States. METHODS This cross-sectional study utilized data collected from a sample of general pediatricians and pediatric subspecialists randomly selected from the American Medical Association’s Physician Masterfile. A survey questionnaire was designed and postal mailed to 938 physicians of general pediatrics (n = 376) and 7 selected pediatric subspecialties (n = 562) between 05/2011 and 04/2012. A total of 453 physicians (182 general pediatrician and 271 subspecialties) returned the survey with 48.3 % response rate. The questionnaire is comprised of 59 items measuring physician characteristics and their capability, willingness, self-efficacy, and barriers in measuring pediatric PROs in clinical practice. Frequency of the item response was reported and linear regression models were conducted to investigate the relationships between physician characteristics (including demographics, year since practice, practice setting, etc.) and the capability, willingness, self-efficacy, and barriers in measuring pediatric PROs.; RESULTS Item level analyses revealed 47.8 % of participants had no intention to use pediatric PROs, 36.0 % reported potential or occasional use, and 16.2 % reported regular use. The intention of measuring pediatric PROs was for evaluating treatment effectiveness (23.2 %), developing treatment plan (22.2 %), and involving patients in the consultation process (17.4 %). Only 26.1 % of participants reported a moderate to strong confidence in measuring pediatric PROs. The major barriers to implementing PROs include limited time/manpower (61.7 %), a lack of formal training (56.4 %), limited scoring skill (54.5 %), and a lack of clinically meaningful cutoffs for results interpretation (52.8 %). Participants who were subspecialists, worked in the academic setting, held greater knowledge in PROs measures, and had higher self-efficacy reported greater capability for and less barriers in measuring PROs than those who were general pediatricians, worked with community care/private practice, possessed limited PROs knowledge, and had lower selfefficacy (all p \ 0.001).; CONCLUSIONS In this national survey, a significant number pediatricians reported limited capability, low selfefficacy, and important barriers in using pediatric PROs measures in clinical practice. Providing systematic training and appropriate resources is important to overcome this issue.
123
Qual Life Res (2012) 21:1–132 (1024) A randomized controlled study on the effect of a group therapy for parents of youths with type 1 diabetes on health-related quality of life Elisabetta Petracci, Department of Statistics; Giulia Cavrini, Free University of Bolzano/Bozen, Italy; Laura Nardi, Department of Pediatrics; Stefano Zucchini, Department of Pediatrics; Elena Tomba, Department of Psychology University of Bologna; Silvana Grandi, Department of Psychology University of Bologna; Franco D’Alberton, Department of Pediatrics: AIMS Group interventions for youth’s parents with type 1 diabetes (T1DM) can improve Health-related Quality of Life (HRQoL). This study investigates how a group therapy, carried out on parents of diabetic children and adolescents, may influence quality of life and indirectly metabolic control of youths with T1DM.; METHODS A randomized study was conducted in 2010 the Pediatric Diabetes Centre of the University Hospital in Bologna. A total of 147 patients aged 2–20 were originally enrolled in the study, 69 of which agreed to participate and were randomized into two groups: 49 controls and 20 cases. Every 3 weeks, 10 psychodynamic group meetings of around 15 parents were held by a trained psychotherapist. Questionnaires and HbA1c were evaluated on all the parents on four occasions (before, during, at the end of the therapy and after 6 months). HRQoL was evaluated by administering the EQ-5D-proxy questionnaire to youths’ parents. This explores five areas related to mobility, care of oneself, habitual activities, pain and/or distress, anxiety and/or depression and includes a visual analogue scale for measuring global health status. Due to the presence of patients dropping out of the study before completion, an intention-to-treat analysis was used. The justification of the use of this controversial method was the absence of statistically significant differences between dropout and not dropout patients at each evaluation time for many variables. To analyze the data a multilevel model for repeated measures was used.; RESULTS Patients’ mean age was 13.2 (range 3.4–19.0), disease duration 6.3 years (range 0.5–15.7). Initial HbA1c was 7.9 (range 5.8–13). The results from the multilevel analysis showed a positive association between the group intervention therapy and EQ-5D-VAS proxy as reported by parents, after adjusting for patients characteristics and clinical and psychological information. Although HbA1c values unchanged during the study, a significant negative correlation between HRQoL and HbA1c was observed.; CONCLUSIONS Group psychological intervention for youths’ parents affected by diabetes seems to improve quality of life of youths as reported by their parents. Although clinical randomized interventions provide the best effect evidence, the matter of dropouts is this still one of the major problems.
(1025) An EORTC phase III trial of adjuvant whole-brain radiotherapy versus observation in patients with 1–3 brain metastases from solid tumors after surgical resection or radiosurgery: quality-of-life results Riccardo Soffietti, University of Torino and San Giovanni Battista Hospital, Neurology/Neuro-Oncology, Torino, Italy; Martin Kocher, University of Cologne, Radiation Oncology, Koeln, Germany; Mehmet Ufuk Abacioglu, Marmara University Hospital, Radiation Oncology, Istanbul, Turkey; Savador Villa, Hospital Germans Trias i Pujol ICO, Radiation Oncology, Barcelona, Spain; Franc¸ois Fauchon, Centre Haute Energie, Nice, France; Brigitta G. Baumert, RadiationOncology (MAASTRO), Maastricht University Medical Centre (MUMC), GROW (School for Oncology), Maastricht, Netherlands; Laura Fariselli, Fondazione Istituto Neurologico ‘‘Carlo Besta’’, Milano, Italy; Tzahala Tzuk-Shina, Rambam Medical Center, Haifa, Israel; Rolf-Dieter Kortmann, University Hospital, Leipzig, Germany; Christian Carrie, Department of Radiation Oncology, Centre Leon
Qual Life Res (2012) 21:1–132 Be´rard, Lyon, France; Mohamed Ben Hassell, Department of Radiotherapy, Centre Euge`ne Marquis-Rennes, Rennes, France; Mauri Kouri, Radiotherapy and Oncology, Helsinki University Central Hospital, Finland; Egils Valeinis, Neurosurgery Clinic, Pauls Stradins Clinical University Hospital, Riga, Latvia; Dirk van den Berge, Radiation Oncology Department, Univesitair Ziekenhuis Brussel, Belgium; Rolf-Peter Mueller, University of Cologne, Radiation Oncology, Koeln, Germany; Gloria Tridello, Azienda Ospedaliera Universitaria Verona, Verona, Italy; Laurence Collette, EORTC Headquarters, Brussels, Belgium; Andrew Bottomley, PhD, on behalf of the EORTC Radiation Oncology and Brain Groups EORTC Headquarters, Brussels, Belgium AIMS This phase III trial compared adjuvant whole-brain radiotherapy (WBRT) to observation (OBS) after either surgery or radiosurgery of a limited number of brain metastases in patients with stable solid tumors. Here, we report the Health Related Quality of Life (HRQOL) results.; METHODS HRQOL was a secondary endpoint in the trial. HRQOL was assessed at baseline, at 8 weeks and then every 3 months until 3 years with the EORTC QLQ-C30 and Brain Cancer Module (QLQBN20). Six primary HRQOL scales were considered: global health status, physical, cognitive, role and emotional functioning, and fatigue. Statistical significance required p = 0.05 and clinical relevance a = 10 point difference.; RESULTS Compliance was 88.3 % at baseline and dropped to 45.0 % at 1 year, thus only the first year was analyzed. Overall, patients in the observation only arm reported better HRQOL scores than did those who received WBRT. The differences were statistically significant and clinically relevant mostly during the early follow-up period: for global health status at 9 months, physical functioning at 8 weeks, cognitive functioning at 12 months, and fatigue at 8 weeks. Exploratory analysis of all other HRQOL scales suggested worse scores for the WBRT group but none was clinically relevant; CONCLUSIONS This study shows that adjuvant WBRT after surgery or radiosurgery of a limited number of brain metastases from solid tumors may negatively impact some aspects of HRQOL, even if these effects are transitory. Consequently, observation with close monitoring with MRI (as done in the EORTC trial) is not detrimental for HRQOL.
(1026) Effects of high intensity resistance training on quality of life in patients with prostate cancer treated with androgen deprivation therapy Sally Wilson, NHS, United Kingdom; Nicholas Stuart, Gwynedd Hospital; Samuele Marcora, University of Kent, United Kingdom AIMS Androgen deprivation therapy (ADT) is an effective treatment for prostate cancer. However, it leads to loss of muscle, increased fat mass and reduced strength, which lead to impaired physical function, reduced independence and ultimately reduced quality of life (QoL). The primary aim of this study was to examine the effects of high intensity progressive resistance training (PRT) on QoL in elderly prostate cancer patients receiving ADT. Secondary aims included the effects of PRT on body composition and physical function.; METHODS Seventeen patients receiving ADT were randomised to 12 weeks of PRT (n = 9) or to usual care (n = 8). PRT consisted of 3 sets of 8 repetitions (designed to replicate 80 % of maximal effort) using 8 different muscle groups, repeated 3 times a week for 12 weeks. QoL (FACT-P), fatigue (Chalder Fatigue Scale), and anxiety and depression (HADs) were assessed at weeks 0 and 12. Body composition (appendicular lean mass (APLM) and fat mass measured by DXA), physical function (senior fitness test battery) and activity (pedometers) were also measured at weeks 0 and 12.; RESULTS Compared with the control group, QoL improved in those patients who underwent PRT (p = 0.030). Prostate specific worries contributed most to the improvement in QoL (p = 0.026). Total fatigue reduced in the PRT
51 group (p = 0.038), with reductions in both the mental (p = 0.025) and physical aspects (p = 0.089). Increases in APLM (+1 kg, p \ 0.05) and reductions in fat percent (p = 0.044) were observed in the PRT group. Physical function significantly improved in the PRT group, with increases in the number of chair stands (+25 %), and ‘Up and Go’ time (15 %).; CONCLUSIONS This small randomised controlled trial demonstrated that PRT is safe and effective in patients with prostate cancer undergoing ADT. Importantly, PRT improved quality of life in these patients. The reasons for this can be explained partly by improvements in body composition, strength, physical function and fatigue. Further studies should evaluate the cost-effectiveness of this approach in routine care in men with prostate cancer.
(1027) Item response theory and factor analysis as mean to characterize occurrence of response shift for longitudinal quality of life study in breast cancer patients Ame´lie Anota, Plateform ‘Quality of Life and Cancer’, France; Caroline Bascoul-Mollevi, Biostatistic Unit, Montpellier, France, France; Francis Guillemin, Clinical epidemiology and evaluation department, Nancy; Thierry Conroy, Centre Alexis Vautrin, France; Michel Velten, Epidemiology and public health laboratory, College of Medicine, Strasbourg, France; Damien Jolly, University Hospital of Reims; Mariette Mercier, Cellular and molecular biology laboratory, University Hospital of Besanc¸on; Sylvain Causeret, Centre Georges Franc¸ois Leclerc; Tienhan Sandrine Dabakuyo, PharmD PhD, Centre Georges Francois Leclerc, France; Franck Bonnetain, PhD, Centre Georges Francois Leclerc, France AIMS Health-related quality of life (HRQoL) is a dynamic process which depends on the adaptation of the patient and reflected by a Response Shift (RS) effect. RS results in a recalibration, a reprioritization and a reconceptualization of key HRQoL domains. Longitudinal analyses of HRQoL have to take into account the possible occurrence of RS. However, there is no standard of statistical analysis to characterize RS. Two complementary methods are investigated to characterize RS.; METHODS This work builds on data of a prospective multicenter study including all primitive breast cancer patients or suspicion. HRQoL was evaluated using the EORTC QLQC30 and QLQ-BR23 at baseline, after surgery, at 3 months and 6 months, according to the ‘‘then-test/post test’’ design: the retrospective assessments done after surgery and at 3 months refer to baseline HRQoL; the retrospective measurement done at 6 months refers to HRQoL at 3 months. The order then-test and post test of HRQoL questionnaires was randomized. Recalibration was explored by Multiple Correspondence Analyses (MCA) and the Linear Logistic Model with Relaxed Assumptions (LLRA) of Item Response Theory (IRT). LLRA gives trend of item easiness parameters. Reprioritization and reconceptualization were explored by Principal Component Analyses (PCA).; RESULTS Between February 2006 and February 2008, 381 patients were included, 90 % had a confirmed breast cancer. PCA show a secondary reprioritization of the QLQ-C30’s dimensions. Fatigue and pain remain priority symptoms. Secondary symptoms are insomnia at baseline, diarrhea after surgery, nausea and vomiting at 3 and 6 months. A stronger and stronger link between functional scales reflects a reconceptualization. Main recalibration’s profiles reflected by MCA are from one modality to an adjacent one. A lower or upward recalibration of each dimension is reflected by the IRT model. Based on retrospective assessment at 6 months of HRQoL at 3 months, arm and breast symptoms were overestimated with trend parameters equal to -0.59 and -1.05 (p \ 0.001).; CONCLUSIONS IRT models have mainly been used to validate HRQoL questionnaires. This work shows their interest to characterize occurrence of RS. Further analyses should be lead to validate their abilities to characterize all RS components.
123
52 (1028) IRT-scaled checklists of mental disorder symptoms for diagnosis and severity assessment at the patient’s first contact Carlos G. Forero, PhD, IMIM, Spain; Gemma Vilagut, MSc, Institut Muncicipal d’Investigacio Medica, Spain; Jordi Alonso, MD PhD, IMIM-Hospital del Mar Medical Research Institute, Spain AIMS Mental health diagnoses are time demanding and frequently unfeasible at patients’ first contact. Many short-PRO measures are specific for just some diagnostics but not others. Conversely, assessing DSM-IV symptom criteria is straightforward and can be conducted succinctly. Here, we study the feasibility of direct assessment of DSMIV symptom endorsement for initial diagnostic of common mood and anxiety disorders and as indicators of severity.; METHODS We applied a multidimensional IRT 2-parameter logistic model to the occurrence of DSM-IV Major Depressive Episode (9), Generalized Anxiety (8), and Panic disorder (13) symptoms. Disorders status was assessed with the CIDI 3.0 interview in a sample representative of the adult population in 6 European countries (ESEMeD, N = 21,425). Model fit was assessed based on least-squares estimators in an n = 10,713 subsample. IRT Individual scores were obtained in a cross-validation sample (n = 10,713). Score diagnostic accuracy was studied using disorder status according to all DSM-IV criteria (symptoms, duration and impairment) as the gold standard. Finally, we explored cut-off scores for disorder presence.; RESULTS The symptom model showed excellent fit (RMSEA = 0.01, CFI = 0.98, TLI = 0.98, WMVR \ 0.06), and yielded maximum information in a theta-score range 1.5 and 2, indicating moderate to high severity cases. The standard error of measurement in the maximum information region was roughly 0.05 for MDE and GAD, and 0.18 for Panic disorder. Scores showed good diagnostic accuracy for all disorders (MDE AUC = 0.92; GAD AUC = 0.90; PAT AUC = 0.92). The maximum-information scores showed high sensitivity and specificity when detecting disorder status in the cross validation sample for MDE (sens = 0.99, spec = 0.87) and GAD (sens = 0.97, spec = 0.86). Panic disorder symptoms performed worse (sens = 0 .90, spec = 0.81).; CONCLUSIONS Straightforward assessment of symptom endorsement is quite informative of disorder status and informs reliably of its severity, in particular in medium through high pathology levels. IRT scoring of diagnostic symptom checklists might be an excellent tool for first-contact diagnoses and severity assessment of depression and anxiety disorders in non-specialized care settings.
(1029) Development of an item bank measuring economic and financial quality of life for individuals with disabilities David Tulsky, PhD, University of Michigan; Pamela A. Kisala, University of Michigan; Noelle Carlozzi, University of Michigan; Jin-Shei Lai, PhD, Northwestern University; Allen W. Heinemann, Northwestern University, United States AIMS Research has shown that financial factors are associated with better well-being, life satisfaction, and quality of life across a variety of different populations. While there is less literature on economic and financial QOL (ECQ) in individuals with disabilities, there is still clinical evidence that supports this important construct. Currently available measures of ECQ tend to include very few ECQ specific items, tend to be generic in nature, or targeted towards individuals with specific medical conditions other than physical disabilities. To address these shortcomings, the current project aims to develop a patient reported outcome measure of ECQ using contemporary measurement development methodology, (i.e., item response theory and applications of item banking).; METHODS The ECQ item development process included a thorough literature review, a series of semi-structured interviews with individuals with SCI (n = 44) and
123
Qual Life Res (2012) 21:1–132 TBI (n = 20), and focus groups with individuals with SCI (12 groups, n = 65) and TBI (7 groups, n = 33). The initial item pool was then revised through expert review and cognitive debriefing interviews with individuals with SCI, TBI, and stroke (n = 5 per item). The final item pool was pilot tested in a cross-sectional sample of individuals with SCI (n = 105), TBI (n = 100), and stroke (n = 100) and confirmatory factor analyses (CFA) were then used to evaluate the unidimensionality of items, and potential secondary factors via evaluation of local dependence among items. Rasch analysis was used to evaluate response patterns of items.; RESULTS The item development process resulted in a preliminary item pool of 67 items. These items were reviewed further using expert reviews and cognitive interviewing with many being revised or deleted resulting in a final pool of 37 items. An additional 5 items were removed after CFA and Rasch analyses (deleted items exhibited low item-total correlations, local item dependence, resulting in 35 items.; CONCLUSIONS The final ECQ item pool has been developed directly with individuals with disabilities and measures a unique component of quality of life. Calibration using Item Response Theory modeling is currently underway. The final measure will be a component of a larger Environmental Factors measurement system and will also be included in the SCI-QOL (Tulsky et al., 2011) and TBIQOL measurement systems (Carlozzi et al., 2011). Economic and Financial factors can serve as significant barriers or facilitators of quality of life and as such this new item bank is a much needed addition to new quality of life measurement systems for individuals with disabilities.
(1030) Factors associated with asthma control and health-related quality of life in children: a path analysis Pranav Gandhi, PhD, South College, United States; Kelly Kenzik, MS, University of Florida; Lindsay Thompson, University of Florida; Sanjeev Tuli, University of Florida; Darren DeWalt, MD MPH, University of North Carolina; Elizabeth Shenkman, PhD, Univ of Florida College of Medicine; I-Chan Huang, PhD, University of Florida AIMS This study primarily aimed to assess the relationship between control of asthma symptoms and asthma health-related quality of life (HRQoL) in children with asthma. The secondary aim was to examine how parental health literacy, perceived self-efficacy with patient-physician interaction, and satisfaction with shared decision making (SDM) contribute to children’ asthma control and HRQoL.; METHODS This cross-sectional study utilized data collected from a sample of asthmatic children (n = 160) aged 8–17 years and their parents (n = 160) who visited four pediatric clinics of a university medical center. Asthma HRQoL was self-reported by children using the NIH’s Patient-Reported Outcomes Measurement Information System (PROMIS) Pediatric Asthma Impact Scale. Satisfaction with SDM, perceived self-efficacy with patient-physician interaction, parental health literacy, and asthma control were reported by parents using standardized scales. Structural equation modeling (SEM) was performed to test the hypothesized pathways by accounting for the measurement properties of individual scales (e.g., asthma HRQoL). Parents’ age and gender, children’s age and gender, and physician-reported pediatric comorbid conditions were included as covariates in SEM.; RESULTS SEM analyses revealed a satisfied model fit of the data (b2 (df) = 1,036.69 (624), RMSEA (90 % CI) = 0.064 (0.057–0.071)). Asthma HRQoL was significantly associated with asthma control status, where children with better asthma control status reported higher asthma HRQoL compared to those with poor asthma control (b = 0.4, p \ 0.001). Parents with higher health literacy and greater perceived self-efficacy with patient-physician interaction had higher satisfaction with SDM (b = 0.16, p \ 0.05; b = 0.58, p \ 0.001, respectively) and, greater perceived self-efficacy
Qual Life Res (2012) 21:1–132 with patient-physician interaction was indirectly associated with higher asthma HRQoL through greater satisfaction with SDM and better asthma control (b = -0.06, p \ 0.05). In addition, greater satisfaction with SDM was indirectly associated with higher asthma HRQoL through better asthma control (b = -0.11, p \ 0.05).; CONCLUSIONS Children’s asthma control status influenced their HRQoL. Additionally, several factors contributing to asthma control indirectly affected asthma HRQoL. Parents with greater perceived self-efficacy with patient-physician interaction were more likely to be satisfied with SDM, which in turn was likely associated with better skills in managing their children’s asthma control, leading to better pediatric asthma HRQoL.
(1031) Application of a universal translation method to PROMISÒ pediatric item banks Janine Devine, University Medical Center Hamburg-Eppendorf, Germany; Philippa L. Gerigk, University Medical Center Hamburg; Franka Metzner, University Medical Center Hamburg; Michael Herdmann, IMIM University, Insight Consulting and Research; Margarita Hurtado, MPH Consulting; Graciella Castillo, American Institutes for Research; JeanHee Moon, Children’s Hospital of Philadelphia; Christopher B. Forrest, MD, PhD, University of Pennsylvania, School of Medicine, United States; Ulrike Ravens-Sieberer, University Medical Center of Hamburg AIMS The Patient Reported Outcome Measurement Information System (PROMIS) is a federally funded initiative in the US to build efficient, precise, valid, and cross-culturally acceptable item banks for measuring health in adults and children. The publicly available item banks are administered via fixed, short forms or Computer Adaptive Test (CAT) assessments. This study aimed to apply a universal method for translation of item banks into German and Spanish for 9 PROMIS Pediatric measures: Positive Affect, Life Satisfaction, Meaning/Purpose, Family Belonging, Family Involvement, Stress Response: Somatic, Stress Response: Psychological, Physical Activity, and Sedentary Behaviors.; METHODS We used a Forward– Backward-Forward translation technique in adherence to the latest recommendations of the ISPOR PRO Translation and Linguistic Validation Task Force. First, 10 independent professional and experienced translators generated National Forward Translations (NFTs) of each item bank (6 German NFTs: 2 for German, Swiss, and Austrian each; 4 Spanish NFTs: 2 for US-Spanish and 2 for European Spanish). The NFTs were compared with one another and to the original English version by translators and subject matter experts. Next, expert consensus on conceptual equivalence and comprehensibility for children in each sub-population was reached, and the revised versions were back-translated. The Backwards Translations were reviewed again, compared to the original English and the NFTs and reconciled versions (one per language) were generated. These were cognitively-tested for comprehensibility and conceptual equivalence in 120 children (55 German, 3 Austrian, 2 Swiss, 40 US-Spanish, and 30 Spanish).; RESULTS During the universal translation process of the 9 PROMIS Pediatric item banks, a number of items were deleted because of inability to translate into a target language or inapplicable target concept. A final conceptually-equivalent item pool easy to understand by German- and Spanish-speaking children as young as age 8 was achieved.; CONCLUSIONS We have applied a universal translation process for pediatric item banks and successfully applied it to 9 PROMIS Pediatric item banks. The translation efforts assure conceptual equivalence. Next steps in this study will be to link the item banks to the English versions to place them on the same scale, thereby creating a single, cross-culturally harmonized set of PRO metrics.
53 (1032) How do respondents perceive health state valuation tasks? A ‘think aloud’ study investigating time trade off and discrete choice experiments Brendan Mulhern, University of Sheffield, United Kingdom; Aki Tsuchiya, School of Health and Related Research, University of Sheffield; John E. Brazier, PhD, Univ of Sheffield, United Kingdom; Donna Rowen, PhD, University of Sheffield, United Kingdom AIMS We report a qualitative ‘think aloud’ investigation into the completion of the health state valuation exercises Time Trade Off (TTO) and Discrete Choice Experiments (DCE). The aim is to investigate the acceptability of the exercises, the strategies used to complete the tasks, how external factors impact on responses made and whether these issues impact on the validity of the exercises.; METHODS The valuation tasks included DCE, DCE incorporating duration (DCETTO), iterative TTO and iterative Lead Time TTO. Hypothetical EQ-5D-5L health states were used for all tasks. A ‘think aloud’ interview protocol including semi structured follow up questions was used to investigate respondent completion of the health state valuation exercises. A convenience sample of 30 respondents were asked to complete between 5 and 7 tasks whilst talking out loud about the strategies they were using to answer the question, and any related thoughts or opinions about the health states or tasks in general that may impact on their answer. Interviews were transcribed verbatim and analysed using framework analysis. An initial coding frame for each valuation exercise was developed based on existing literature and the participant responses were allocated to the framework.; RESULTS Respondents report using a range of strategies to complete the DCE and DCETTO tasks, and the addition of duration generally increases the difficulty of the task. Across all tasks, respondents report that a range of external influences such as personal health experiences and ability to cope with the health states impact on their responses. In terms of impact on validity, respondents report that the tasks are generally complex and that some of the health states are not realistic. There were also some difficulties perceiving the difference between the ‘severe’ and ‘extreme’ EQ-5D-5L severity levels.; CONCLUSIONS Health state valuation exercises are widely used in the generation of utility values for the estimation of QALYs. However, little qualitative evidence is available regarding how respondents perceive the task and complete the exercise. Qualitative evidence can inform the choice of health state valuation techniques, and the design of valuation studies.
(1033) Valuing a vision ‘bolt-on’ item for the EQ-5D Louise Longworth, Jr., Health Economics Research Group, Brunel University; Yaling Yang, PhD, Brunel University, United Kingdom; John E. Brazier, PhD, Univ of Sheffield, United Kingdom; Donna Rowen, PhD, University of Sheffield, United Kingdom; Aki Tsuchiya, School of Health and Related Research, University of Sheffield; Tracey A. Young, PhD MSc BSc CStat, University of Sheffield, United Kingdom AIMS Generic preference-based measures have been criticised for not being sufficiently sensitive for use in some conditions. There has been growing interest in developing ‘bolt-on’ items by combining additional aspects of health with existing generic preferencebased measures such as the EQ-5D. However, it is not clear what additional information such bolt-ons can provide. The aim of this study was to develop a ‘vision’ bolt-on to EQ-5D and to assess its impact on the health state values.; METHODS An item on vision to bolt on to the EQ-5D was developed following a systematic review of the literature on the performance of EQ-5D in a range of
123
54 conditions. The bolt-on item was developed following a review of existing instruments in vision, and has three levels as in EQ-5D (‘no problems’ for level 1, ‘some problems’ for level 2 and ‘extreme problems’ for level 3). The bolt-on item was tested in an exploratory valuation study. An orthogonal design of a 6-dimension 3-level instrument includes 18 mostly severe states. Starting from these, 20 health states each for EQ-5D+ Vision and EQ-5D were selected for valuation. The set of EQ-5D+ Vision states adds two mild states to the orthogonal set. The set of EQ-5D states removes the sixth vision dimension from the 20 EQ-5D+ Vision states. This process maintained the orthogonal property of health state selection for both instruments. A valuation study was undertaken using faceto-face interview to obtain time trade-off values from members of the generic public in South Yorkshire, UK.; RESULTS Three hundred people completed the interviews. Mean health state values ranged from 0.05 to 0.96 for EQ-5D, and -0.04 to 0.95 for EQ-5D+ Vision. Various econometric models have been fitted to the data, and the results indicate that the inclusion of a vision bolton has a statistically significant impact on the valuation of EQ-5D health states.; CONCLUSIONS This study has demonstrated that a vision bolt-on item significantly impacts on EQ-5D health state values. However, the methodology to develop and implement bolton items is still in development and issues remain as to whether and how bolts–ons should be incorporated for health technology assessment.
(1035) Quantifying the top region of the QALY scale with multidimensional scaling A.M.M Arons, MSc, Radboud University Nijmegen Medical Centre, Netherlands; Paul F.M. Krabbe, PhD, University of Groningen, Dept. of Epidemiology, Netherlands AIMS A new methodology, multidimensional scaling (MDS), is introduced to scale health-state values on a cardinal scale. MDS is based on the ranking of similarity data of health states and seems especially suited to calibrate difficult regions on the health-state continuum (0 = dead, 1 = full health) such as the top-end, the bottom-end and states around ‘dead’. A comparison will be made with discrete choice (DC) modelling, a method that is based on preference data, for the top-end.; METHODS A convenience sample of clinical experts in the field of dementia were invited to participate in the study. They first completed a DC experiment on the dementia quality of life instrument (DQI). Subsequently, they valued the top 7 health states (health states: 111111, 211111, 121111,…, 111112) of the DQI based on their similarity. The similarity data was analyzed and transformed using a nonmetric MDS algorithm.; RESULTS In total 108 experts in the field of dementia participated in the study. The DC experiment provided the following utilities for the health states: 111111 = 1, 211111 = 1, 121111 = 0.945, 112111 = 0.951, 111211 = 0.968, 111121 = 0.931, 111112 = 1. The unidimensional MDS solution provided the following rescaled utilities of the health states: ‘111111’ = 1, ‘211111’ = 1.009, ‘121111’ = 0.980, ‘112111’ = 0.933, ‘111211’ = 0.963, ‘111121’ = 0.956, and ‘111112’ = 0.931. The stress-1 value was 0.243, which indicated a poor fit. The intra-class correlation (ICC) between the DC model and the MDS unidimensional solution was 0.39 (p [ 0.1). Rescaling the Euclidean distances of the two-dimensional solution into utilities provided: ‘111111’ = 1, ‘211111’ = 0.956, ‘121111’ = 0.0.957, ‘112111’ = 0.913, ‘111211’ = 0.952, ‘111121’ = 0.918, and ‘111112’ = 0.931. The stress-1 value was 0.106, which indicated a poor to fair fit. The ICC between the DC model and the twodimensional solution was 0.58 (p = 0.64).; CONCLUSIONS The DC model appeared to provide the most easily interpretable results, however, not all attribute levels were significant. The
123
Qual Life Res (2012) 21:1–132 two-dimensional MDS solution seemed interpretable if one is willing to allow two-dimensional utilities. The results of the current study do not allow forming a preference for either methodology. More respondents would most likely lead to more stable estimates with both methodologies, allowing for a more valid comparison.
(1036) New series of tools for comprehensive symptom profile assessment in patients with hematological malignancies Tatiana I. Ionova, Multinational Center for Quality of Life Research, Russian Federation; Denis A. Fedorenko, PhD, Nat’l Pirogov Med Surgical Ctr, Russia; Tatiana Nikitina, PhD, Multinational Ctr for QoL Research, Russia; Nikita Mochkin, MD, National Medical Surgical Center, Russia; Kira A. Kurbatova, Multinational Center for Quality of Life Research, Russian Federation; Gary I. Gorodokin, New Jersey Center for Quality of Life and Health Outcomes Research, NJ, USA; Andrei Novik, MD PhD, National Medical Surgical Center, Russia AIMS Comprehensive symptom assessment before and during treatment in patients with hematological malignancies is worthwhile. We aimed to develop a new series of symptom assessment tools—Comprehensive Symptom Profile (CSP) which unites questionnaires aimed at providing a comprehensive assessment of the severity of symptoms specific to a certain hematological malignancy. The goal of this study was to test applicability of CSP-Lymphoma, SCP-Multiple Myeloma, and CSP-Chronic Leukemia in patients with malignant lymphomas (ML), multiple myeloma (MM), and chronic myeloid leukemia (CML).; METHODS 190 patients were included in the study: ML— 108, mean age—35 years old, male/female—43/65; MM—46, mean age—54 years old, male/female—28/18; CML—36, mean age— 48 years old, male/female—18/18. The CSP questionnaires are selfreported tools which allow physicians to assess the severity of symptoms specific to ML, MM, and CML patients, respectively. They consist of numerical rating scales, scored from ‘‘0’’ (no symptom) to ‘‘10’’ (most expressed symptom). Utility of the tools was demonstrated by means of interviewing both patients and physicians. Applicability of the tools with preliminary analysis of psychometric properties was tested. Patients filled out the questionnaires before and at different time-points of treatment.; RESULTS The utility of the tools was demonstrated. All of the items were easy for the patients to read and understand. The patients acknowledged the comprehensiveness of the tools. Their completion in paper and pencil format took 5–7 min. The data produced by the tools were clear for interpretation by physicians and were used by them in day-to-day decision making. Statistically significant differences between the groups with different clinical characteristics (p \ 0.05) were found for the majority of symptoms. Reliability of the tools was satisfactory. Sensitivity to changes was demonstrated by comparison of symptom severity before treatment and after chemotherapy. The majority of symptoms after treatment were less severe than at base-line.; CONCLUSIONS The instruments of the CSP series are appropriate and practical tools to assess the symptom severity in patients with hematological malignancies. Satisfactory psychometric properties of the tools were demonstrated. Further studies are needed before the wide-spread use of the instruments of the CSP series in clinical practice and clinical trials.
(1037) A study of the QOL criteria of nursing college students Keiko Saitou, Kwassui Women’s University, Omura, Nagasaki, Japan; Michiko Kobayashi, MD PhD, Kwassui Women’s College, Japan; Takashi Mandai, MD, Japanese Society of Quality of Life Research, Japan
Qual Life Res (2012) 21:1–132 AIMS This study aims to identify the main factors of QOL improvement in students receiving highly specialized education closely related to people’s QOL, by measuring their QOL and improving the QOL questionnaire.; METHODS We prepared a questionnaire with 9 domains and 42 question items, developed by the nursing students themselves, based on the World Health Organization (WHO) concepts and the Japanese Society of QOL Research QOL questionnaire.; RESULTS The response rate was 66 out of 80 female sophomore students. Looking at the 9 domains’ average scores, the highest were in Family and Social Life, followed by Well-Being. The lowest average score was in Physical. Also, comparing by domain the 13 highest QOLs with the 12 lowest, the graph is cylindrical, with the QOL of the highest 13 being better than the lowest 12 in all domains (p \ 0.01).; CONCLUSIONS There was a decrease of 0.3 points of the nursing students’ QOL average, comparing the first and the second survey (p \ 0.01). The Well-Being, Dietary Habits, Physical, Psychological Functioning, and Economical domains showed hardly any difference, but Sleeping, Living Environment, and Social fell sharply. As for the Daily Life and Social domains, the big difference can be explained by the fact that for the second survey we moved questions concerning family from the Daily Life domain to the Family and Social domains. The results of this study indicate that favorable family relations and the existence of somebody to talk to (somebody dependable) are factors needed for the improvement of QOL, but because the nursing students’ second average QOL score was lower compared to the first, we argue that for improvement of the QOL of nursing students, their education should help them to find a sense of calling, hope, meaning and significance as a nursing student, and be a guide to their future. To this end keeping regular hours is important, and it is especially necessary to put the living environment in order by ensuring a sufficiency of sleep, and mental and physical health.
(1038) Patient-reported outcome and quality of life instruments database (PROQOLID): evolution of content, structure, and functionalities (2001–2012) Marie-Pierre Emery, MSc, MAPI Research Trust, France; Martine Caron, MAPI Research Trust, France; Laure-Lou Perrier, MAPI Research Trust; Katrin Conway, MA, MAPI Research Trust, France AIMS PROQOLID was developed in mid-2001 to provide all those involved in health care evaluation with a comprehensive and unique source of information on Patient-Reported Outcome (PRO) and HealthRelated Quality of Life measures available through the Internet. The objective of this study is to review the evolution of content, structure, and functionalities of PROQOLID since its creation.; METHODS The archives of PROQOLID were searched to retrieve the database just before its launch and to compare its content and structure as it was in April 2012.; RESULTS The September 2001 database (then known as QOLID) was retrieved. It included 313 instruments (32 generic and 281 disease- or condition-specific). Instruments specific to oncology were the most frequent (54). The structure was simple, with only three categories in the membership level: contact, conditions of use, and a copy of the questionnaire. By comparison, the April 2012 database includes 714 instruments (an increase of 128 %, with 100 generic and 614 disease- or condition-specific), with an increase of almost 40 instruments each year (up by 401 in 10 years and half). Instruments specific to nervous system diseases are the most frequent (141). The information displayed for each instrument has been improved with the addition of five categories in the membership level: translations available, descriptive information, content validity documentation, measurement properties, references, and websites. A search engine was added to enable advanced searches with the following criteria: abbreviation, full
55 name, author, dimension, disease, type of instrument, population, and languages. In January 2005 the database was renamed PROQOLID to reflect the wider use of the term PRO. In April 2012, to celebrate PROQOLID’s decade of service, a new Google-type search engine was added to make browsing more user friendly. Very soon all users will be able to search for instruments by the domain assessed, and will have direct access to contact information, review copies, and user agreements for all instruments distributed by MAPI Research Trust.; CONCLUSIONS Over a period of 10 years, the PROQOLID database has considerably evolved in content and structure, and offers a range of information and services adapted to the evolution of the field.
(1039) Quality of life dimensions in elderly people and its relation to religiosity Neyde Cintra Santos, Faculdade Adventista da Bahia, Brazil; Gina A. Abdala, MPH, Brazil Adventist University, Brazil; Rafael BMO Vidal, PUC Minas Brazil, Brazil AIMS To evaluate the dimensions of quality of life (SF36) in a community-dwelling’s elderly and its relation to organizational religious affiliation (ORA), non-organizational religious affiliation (NORA) and intrinsic religiosity (IR).; METHODS Cross-sectional study with a random sample of 83 elders from Cachoeira/BA, Brazil.; RESULTS The female predominates with 61.4 %. The age and dimensions of quality of life according to the mean and standard deviation were: age (71 ± 9.39), functional capacity (70.6 ± 23:57), physical limitations (66.86 ± 36.23), pain (66.75 ± 30.62), general health (63.79 ± 22.85), vitality (64.59 ± 17.72), social functioning (71.07 ± 26.32), emotional limitations (77.26 ± 28.96) and mental health (74.31 ± 19:52). As to religion affiliation, 60.2 % are Catholic and 27.7 % are Protestants/evangelicals. The data also showed that 74.4 % attend church (ORA) two or more times per week, 89.1 % practiced religious activities (NORA) from twice per week to more than once a day. For questions about IR: feel the presence of God, act on their beliefs and strive to live the religion in all aspects of life, the elderly had a percentage of 95, 90.2 and 84.2 % respectively. Using the Pearson Chi-square test, it was found that the emotional limitations was the dimension that received the greatest influence of ORA (v2 = 11.539, p = 0.001, Cramer’s V = 0.372), NORA (v2 = 7.949, p = 0.005, Cramer’s V = 0.309) and IR (v2 = 5.126, p = 0.05, Cramer’s V = 0.249). This means that 14, 10 and 6 % respectively of the variation among those who had scores greater than 50 in emotional limitations can be explained by those who are attend church two or more times per week, are regular to private religious activities and strive to live the religion in all aspects of life. It was also found positive influence on physical limitations, pain, general, mental and social health. In other hand, no association between religiosity and the physical and vitality dimensions was found.; CONCLUSIONS It is inferred that there is a positive association between religiosity and quality of life of elderly people. Since aging is a reality in Brazil and in the world, to expand these findings will be of great benefit to their quality of life.
(1040) Tobacco and quality of life in secondary students of Mexico Carlos Alejandro Hidalgo Rasmussen, PhD, University of Guadalajara, Mexico; Claudia C. Marti Criollo, University of Guadalajara, Mexico; Gerardo Jimenez Haro, University of Guadalajara AIMS In Me´xico 25,000 people death each year by tobacco consume, and have a tremendous impact for passive consume too. Adolescents are affected especially not only for health problems if not because tobacco is a door for other drugs. The relationship between quality of life (QOL)
123
56 and tobacco consume has been explored in other countries and found worse QOL in those who smoke (Martinez, 2004, Cayuela, 2007). The aim of this study was to analyze the relationship between QOL and tobacco consume in adolescents students of Mexico.; METHODS Transversal study with analytic design carried out in 2010. Secondary school adolescents (n = 570, 11–17 years) were included, 48.1 % women, 20 % worked. An online self-administered questionnaire integrated by perceptual module of Youth Quality of Life-Research Instrument, and five items about tobacco consume, adapted from 2009 High School Youth Risk Behavior Survey. Analysis: Kruskal–Wallis, Mann–Whitney U with SPSS 17. Ethics: informed consent, voluntary, private and confidential.; RESULTS Lower QOL was for those who ever smoke in total index [TI] (p .042) and general and relational domains (p = 0.005 and p = 0.30 respective), those who had smoke 1 or 2 days during the past 30 days compared with those who had not smoke in personal domain (p .029), and those who during the past 12 months tried to quit smoking cigarettes in TI (p .008), general (p .005) personal (p .008) and relational domains (p .008).; CONCLUSIONS The age when they began to smoke was not different for QOL, the number of cigarettes neither shown differences for QOL and to smoke cigarettes daily every day for 30 days. The few cases in these questions may be was not permit the analysis, but may be the damages for the early age of this sample are not evident yet. Screening in secondary schools for tobacco consume is needed toidentify and intervene early with students because that can be a signal for deep problems.
(1041) Quality of work life of academicians in Ankara, Turkey Selda Tasdemir Afsar, Nigde University, Turkey AIMS This study has two aims: (1) To determine whether there is a meaningful difference in the quality of work life between the academicians working in the Foundation Universities and State Universities. (2) To examine whether the quality of work life of the academicians working in foundation and state universities differs according to the age, gender, marital status, academic position, salary, total working year in the organization, total working year in the related position and assignment type.; METHODS The samples of the study are consisted of 570 academicians in 2 foundation universities and in 2 state universities in Ankara, the capital city of Turkey. The data of the study was collected from the Scale of Quality of Work Life including questions in order to meet 7 needs (health and safety needs, economic and family needs, social needs, esteem needs, actualization needs, knowledge needs, and aesthetics needs) developed by Sirgy et al. (2001) and the questionnaire form including 14 questions.; RESULTS The data collected was processed with SPSS (Statistical Packet for Social Sciences 16.0) packet programme. It was found that the level of quality of work life of the academicians working in the foundation universities is higher than the quality of work life of the academicians working in the state universities as a result of ‘t’ test conducted in order to determine whether the quality of work life of academicians differs due to the type of university they work in (t = -2,949 p = 0,003 \ 0.01). It was also found that the quality of work life of the academicians working in both foundation and state universities differs due to the age, gender, marital status, academic position, salary, total working year in the organization, total working year in the related position and assignment type as a result of ‘t’ test and ‘One-Way Anova’ analysis conducted in order to determine whether the quality of work life of the academicians working in foundation and state universities differs according to the age, gender, marital status, academic position, salary, total working year in the organization, total working year in the related position and assignment type.; CONCLUSIONS The results found in the study to support the relevant literature that the quality of work life increases as the age, salary, academic position and total working year in the related position and organization increase and the quality of work life of men, married
123
Qual Life Res (2012) 21:1–132 people and permanent staff is higher than women, single people and temporary staff.
(1042) Etiological factors affected quality of life due to their association with traditional Chinese medicine syndrome Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Chen Zhihui, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yu Changhe, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Lv Meijun, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Ma Yuan, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Wang Yang, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS As we have discussed in a former study, etiological factors affect subjects in a way they performed in their quality of life (QOL) outcomes. However, TCM pays much attention to the association between the syndrome and its causes. In this study we make the hypothesis that the associations between etiological factors and Traditional Chinese Medicine (TCM) syndromes are the potential causes for the former conclusion. Here we aim at discussing this association and try to find how syndrome factors influence patients’ QOL.; METHODS The patients were diagnosed in different syndromes by TCM doctors and classified into three groups: Qi deficiency, Phlegm and Blood stasis. A descriptive analysis of the above variables was performed. Differences in etiological factors between groups are tested by Chi-square.; RESULTS All the 1,574 patients were diagnosed blood stasis, and 789 (age 59.5 ± 9.7 years; male, 45.6 %) were Qi deficiency, and 785 (age 59.1 ± 10.4 years; male, 51.2 %) phlegm. Among the three contributing factors, ‘‘overwork’’ was selected in the ratio 79.2 % in blood stasis, 83.9 % in Qi deficiency, 74.5 % in phlegm (p \ 0.05); ‘‘cloudy or rainy weather’’ 43.7, 38.8, 48.5 %, (p \ 0.05); ‘‘excessive emotions’’ 68.0, 68.6, 67.5 % (p [ 0.05). This reflects the TCM theory that ‘‘overwork’’ greatly related to patients disease status and can aggravate Qi deficiency in an more obvious trend; ‘‘cloudy or rainy weather’’ aggravate phlegm symptom; and ‘‘excessive emotions’’ may usually cause the Qi stagnation thus leading to blood stasis, which is not a special etiological factors for all the three syndromes considered in this study and show no difference.; CONCLUSIONS This study reveals evident correlations between etiological factors and the syndrome factors with which the subject presents. Different etiological factors indicate different syndromes, which accordingly leads to different QOL as we have discussed in a former study. This may remind the researchers giving considerations to the etiological factors in the QOL research in future.
(1043) Health-related quality of life (HRQL) after surgery for esophageal cancer: data of a systematic review and preliminary meta-analysis Marc Jacobs, MSc, Amsterdam Medical Center, Netherlands; Rhiannon C. Macefield, University of Bristol, United Kingdom; Jane Blazeby, MD, University of Bristol, United Kingdom; Ida Korfage, PhD, Erasmus MC, Netherlands; Mark I. Van Berge Henegouwen, Department of Surgery, Academic Medical Center, The Netherlands; Hanneke C.J.M. De Haes, Department of Medical
Qual Life Res (2012) 21:1–132 Psychology, Academic Medical Center, The Netherlands; Ellen M. Smets, Department of Medical Psychology, Academic Medical Center, The Netherlands; Mirjam AG Sprangers, PhD, Academic Medical Center, Netherlands AIMS To estimate change in HRQL and time needed for HRQL to recover after surgery for esophageal cancer; METHODS A systematic literature search to identify studies that measured HRQL with a validated multidimensional patient-reported outcome measure. Excluded were cross-sectional studies and studies published before 2000. Studies were arranged according to the HRQL measure used (e.g. EORTC QLQ C30), treatment evaluated (surgery alone [open surgery (OS); minimally invasive surgery (MIS); gastrectomy] or multimodal treatment (MM) [neoadjuvant chemo and/-or chemoradiation + surgery]), and timing of assessment applied (4–6 weeks, 3, 6, 9, and 12 months postoperative). Studies similar regarding HRQL measure, treatment, and timing of assessment were entered into meta-analyses (random effects model). For each treatment, raw HRQL scores were extracted or calculated to estimate the weighted mean difference and 95 % confidence interval between the post-operative HRQL assessment and baseline value. An absolute difference in HRQL scores of =10 was considered a clinically relevant change. Authors were contacted for missing data.; RESULTS Data from eight studies (out of 21 eligible) were available, reporting on OS (three studies), MIS (two studies), and MM (three studies) by use of the EORTC QLQ C30, or OES18/OES24. For OS and MIS, post-operative HRQL data was only available until 6 months, and studies displayed a significant between study variability (i.e. heterogeneity) for most HRQL domain/symptoms. Preliminary meta-analysis indicates an immediate clinically relevant decline in HRQL after surgery. After OS, HRQL scores for insomnia, constipation, and speech problems recovered at 3 months, and coughing recovered at 6 months. However, physical functioning, pain, dyspnoea, fatigue, and reflux did not recover within the first 6 months after surgery. After MIS, HRQL scores for insomnia recovered at 3 months, and for dyspnoea at 6 months. After MM, most HRQL domains/symptoms recovered at 3 months, although dyspnoea, diarrhea, and dysphagia did not recover within the first post-operative year. Coughing recovered at 9 months. Physical and role functioning, and fatigue recovered at 12 months; CONCLUSIONS The rate of HRQL recovery varies across treatment received and HRQL domain/symptom measured. For OS and MIS, heterogeneity in weighted mean differences across studies limits robust estimates of HRQL recovery.
(1044) Daytime sleepiness and sleep quality are associated with health-related quality of life in Norwegian dialysis patients Fredrik B. Brekke, Student of medicine, University of Oslo, Norway; Amin Amro, Oslo University Hospital; Tone B. Østhus, Oslo University Hospital, Norway; Helga Gudmundsdottir, Oslo University Hospital; Ba˚rd Waldum, University of Oslo; Ingrid Os, University of Oslo, Norway AIMS There is a paucity of data about sleep disturbances in Norwegian dialysis patients. The aims of the study were to assess associations between sleep complaints patients and health-related quality of life (HRQOL).; METHODS A cross-sectional population-based study of 301 prevalent dialysis patients (% 66 males) in 10 dialysis units from all parts of Norway. The questionnaires Kidney Disease Quality of LifeSF36 (KDQOL-SF36), Pittsburgh Sleep Quality Index (PSQI) and Epsworth Sleepiness Scale (ESS) were used. Mental (MCS) and physical component summary scores (PCS) were calculated from SF36. Depression was measured using Beck Depression Inventory (BDI).; RESULTS Sleep quality was strongly related to depression (rho = 0.49, p \ 0.001), MCS (-0.32, p \ 0.001) and PCS score (-0.33, p [ 0.001), and to all eight subscales of SF-36 (data not
57 shown). Effect of kidney disease (-0.41, P \ 0.001) and burden of kidney disease (-0,23, p \ 0.001) were associated with PSQI score. ESS was also related to all subscales of SF-36 except physical function. Associations were observed between ESS and vitality (-0.30, p \ 0.001), effect of kidney disease (-0.19, p = 0.002), burden of kidney disease (-0.20, p = 0.001) and other subscales of KDQOL. In separate multivariate regression analyses, BDI (logtransformed, p \ 0.001), MCS (p \ 0.001), and PCS (p \ 0.001) remained significantly associated with PSQI after adjustment for age, gender, time in dialysis and comorbidity. Using a similar model, ESS remained associated BDI (logtransformed, p \ 0.001), after adjustment for age (p = 0.049), comorbidity (p = 0.026), time in dialysis (p = 0.057), and gender. Similarly ESS remained related to PCS (p \ 0.001) and MCS (p = 0.001) after multiple adjustment in separate analyses.; CONCLUSIONS Poor sleep and daytime sleepiness are both strongly associated with reduced HRQOL and depression. Cause-relationship cannot be assessed due to the cross-sectional design and needs to be further investigated.
(1045) Effects of complementary health practices on perceived stress and quality of life in Japanese outpatients Tomoaki Kimura, PhD, MOA Health Science Foundation, Japan; Hiroshi Matsuo, Aiwakai Tatekawa Clinic, Japan; Naoharu Iida, Zuisenkai MOA Sapporo Clinic; Yoshiteru Maki, Korinkai Kagoshima Clinic, Japan; Kazuo Nitta, Gyokusenkai MOA Takanawa Clinic, Japan; Seiya Uchida, MOA Health Science Foundation; Yasutami Tsuda, MOA Health Science Foundation; Hiroyuki Sakaguchi, MOA Health Science Foundation; Kiyoshi Yamaoka, MOA Health Science Foundation; Kiyoshi Suzuki, Gyokusenkai MOA Takanawa Clinic AIMS In Japan and other countries, complementary health practices based on lifestyle modification such as exercise, diet, art and culture, gardening, and biofield therapy have received increasing attention. The objective of this study was to clarify the effects of these health practices on perceived stress and the quality of life (QOL) of outpatients.; METHODS Computer-assisted data collection was conducted with outpatients in ten clinics between April 2009 and December 2011. Participants completed questionnaires that measured socioeconomic status (SES) and frequencies of complementary health practices. Psychological stress was measured using the 10-Item Japanese version of the Perceived Stress Scale (JPSS-10). Generic QOL was measured using the 10-Item MOA Quality of Life Questionnaire (MQL-10). A cross-sectional analysis was conducted to examine associations of the variables of demographic variables, SES, and health practices with the QOL. A longitudinal analysis was also conducted to determine the associations of health practices with perceived stress and QOL at the follow-up assessment.; RESULTS Baseline data of 1,423 participants (366 men and 1,057 women; mean age 58.5) were available for the cross-sectional analysis, and data of 309 participants (98 men and 211 women; mean age 56.6) were available for the longitudinal analysis at the follow-up. Results indicated that most outpatients had chronic diseases (80.2 %) such as hypertension (19.6 %), dyslipidemia (17.8 %), and diabetes (7.1 %). Gender differences were not observed in stress and QOL, but age was positively correlated with QOL and negatively with stress. There were very little associations between SES and QOL, and weak associations between QOL and the frequency of each health practice (r = 0.15–0.27; p \ 0.001). The longitudinal analyses indicated that among the health practices, exercise had the strongest positive association with QOL (r = 0.31), whereas art and culture had the strongest negative association with stress (r = -0.30). There were also significant associations between other variables and QOL (r = 0.17–0.26), and stress (r = -0.14 to -0.24).; CONCLUSIONS
123
58 These results suggest that complementary health practices had a significant effect on maintaining, or improving QOL, as well as in reducing stress of Japanese outpatients.
(1046) Reliability and validity of the body image quality of life inventory: version for Brazilian burn victims Fla´via F. O. Assunc¸a˜o, Universidade de Ribeira˜o Preto; Nata´lia Gonc¸alves, University of Sa˜o Paulo; Jayme A. Farina-Junior, University of Sa˜o Paulo; Rosana Spadoti Dantas, PhD, University of Sa˜o Paulo, Brazil; Ma´rcia A. Ciol, University of Washington; Lidia A. Rossi, PhD, University of Sao Paulo, Brazil AIMS This study aimed to adapt the Body Image Quality of Life Inventory (BIQLI) to Brazilian Portuguese and to assess the reliability, construct validity, and dimensionality of the adapted version.; METHODS We performed the adaptation according to the following steps: translation, expert committee assessment, back-translation, semantic analysis, and pre-test. The reliability and construct validity were assessed, correlating the BIQLI scores with self-esteem measured by the Rosenberg’s Self-Esteem Scale, and with health status measured by the Burns Specific Health Scale-Revised (BSHS-R). Factor analysis with Varimax rotation was applied, establishing a minimum value of .40 to compare the items’ factor loadings in each factor or component.; RESULTS Study participants were 77 adult burn victims. Most of them suffered burns in 20 % or less of the total body surface area (62.3 %). Cronbach’s Alpha for the total scale was high (.90), and when excluding one item, alphas ranged from .89 to .90. Correlations between the scores of the adapted version of BIQLI and of the RSES were moderate for the entire sample (r = .30, p = .008), but when separating by sex, the correlation was moderate for women (r = .38, p = .01) but weak for men (r = .19, p = .28). Correlations between the adapted BIQLI scores and the scores of the affect and body image domain of the BSHS-R were strong for women (r = .51, p \ .001) and moderate for men (r = .37, p = .032); and between the adapted BIQLI scores and scores in the interpersonal relationships domain of the BSHS-R were moderate for the entire sample (r = .33, p = .003), moderate for women (r = .42, p = .005), and weak for men (r = .17, p = .34). The results of the factor analysis showed five components with eigenvalues greater than one, which accounted for 65 % of the data variance. However, six items loaded on more than one factor and no satisfactory explanation was found for the pattern found in our study.; CONCLUSIONS In conclusion, the adapted version of the BIQLI has shown statistically acceptable levels of reliability and validity for Brazilian burn patients and will continue to be studied for further validation. This study was sponsored by Sa˜o Paulo Research Foundation—FAPESP.
(1047) Linguistic differences between translations of quality-of-life questionnaires: a comparative study of the English and Polish versions of EORTC QLQ-C30 Dagmara Kulis, EORTC, Belgium; Maria Arnott, EORTC; Cheryl Whittaker, EORTC; Eva Elfriede Greimel, PhD, Medical University Graz, Austria; Andrew Bottomley, PhD, EORTC Headquarters, Belgium; Michael Koller, PhD, Center for Clinical Studies, Germany AIMS Currently the EORTC Quality of Life Department administers over 700 questionnaires in 60 languages for the EORTC Quality of Life Group. Each language has its own characteristics that influence the translation process. In this study, we compared the Polish translation of the EORTC QLQ-C30 questionnaire with its English original to find the key differences in these languages.; METHODS We analysed the source and target versions (English and Polish) of the QLQ-
123
Qual Life Res (2012) 21:1–132 C30 questionnaire. We compared and analysed all elements: instructions, time frames, response scale and items.; RESULTS We found the following differences: (1) In general, the English questionnaire is more informal in register (e.g. ‘‘Quite a bit’’). In a questionnaire it is impossible to use the same language register in Polish and so the translation is more formal (‘‘Znacznie’’—‘‘Markedly a lot’’, instead of a more literal translation of ‘‘Quite a bit’’—e.g. ‘‘Dosc sporo’’). (2) More polite, formal forms of address are used in Polish (‘‘Pan/Pani’’— ‘‘Mr/Mrs’’), raising the register even further. The direct equivalent of ‘‘you’’—plural ‘‘Wy’’—could be used but it is informal and rather archaic. (3) Using ‘‘Pan/Pani’’ triggers the issue of gender—all verbs, adjectives, pronouns, etc. have to be adapted accordingly. (4) Gender has to be taken into account in items 29 and 30: the response scales need adjustment of the gender of the item’s subjects: ‘‘Quality of life’’ is feminine, ‘‘health’’ is masculine—hence ‘‘bardzo zla/doskonala’’ and ‘‘bardzo zly/doskonaly’’ (in English, one version: ‘‘very poor/ excellent’’). (5) In Polish, ‘‘please’’ has to be translated as a verb in a personal form and becomes 1st person singular, making the text more personal. (6) In the Polish questionnaire, the question particle ‘‘czy’’ is placed in the time frame (‘‘Czy w ostatnim tygodniu’’—literally: ‘‘Whether in the past week’’) to avoid repetitions in each item.; CONCLUSIONS The main differences in the English and Polish questionnaires show that the translation process is more complex than simple, literal translation. In the context of questionnaires, certain aspects have to be different (like register), so that the comprehensibility, appropriateness and general validity are preserved.
(1049) Assessing the measurement invariance of the functional assessment of cancer therapy-colorectal quality of life instrument between self- and interviewer-administration Carlos King Ho Wong, Mphil, BSc, Department of Family Medicine and Primary Care, The University of Hong Kong, Hong Kong; Cindy L.K. Lam, MD, The University of Hong Kong, Peoples Republic of China; Brendan Mulhern, University of Sheffield, United Kingdom AIMS To test the measurement invariance of the Functional Assessment of Cancer Therapy-Colorectal (FACT-C) in patients with colorectal neoplasms between two modes of administration (selfand interviewer-administered). It is important to establish the measurement invariance of the FACT-C between different modes of administration to ascertain whether it is valid to pool or compare FACTC data collected by different modes. Unless measurement invariance between administration modes is confirmed, group comparisons of scores collected by different administrative modes may be biased and misleading.; METHODS A cross-sectional study of 391 Chinese patients with colorectal neoplasms recruited from colorectal specialist outpatient clinics of a regional hospital in Hong Kong between September 2009 and July 2010. Confirmatory factor analysis (CFA) was used to test the original five-factor model (physical well-being, social well-being, emotional well-being, functional well-being and colorectal cancer subscale) of the FACT-C on data collected by self- and interviewer-administration in single group analysis. Multiple-group CFA was then used to compare the factor structure by each mode of administration group using Chi-squares tests and other model goodness-of-fit statistics such as root mean square error of approximation, comparative fit index and Tucker-Lewis index.; RESULTS The hypothesized five-factor model of FACT-C demonstrated good fit for each group. The hypothesized factor model (configural invariance) and factor loadings (metric invariance) were invariant between administration groups in multiple-group CFA. Three violations of invariance were identified in three items: GP3, GS3 and GF1 of the physical, social and functional well-being, respectively. Intercepts and its corresponding error variances were not identical between administration groups, suggesting evidence of partial strict factorial invariance
Qual Life Res (2012) 21:1–132 CONCLUSIONS Our results confirmed that the five-factor structure of FACT-C was invariant in Chinese patients for both self- and interviewer-administration. It is appropriate to pool or compare data of emotional well-being and colorectal cancer subscales that demonstrated measurement invariance between these two modes of administration. Measurement invariance in three items, one from each of the other subscales, may be contaminated by response bias between modes of administration.
(1050) The development and content validation of a novel generic family quality of life instrument: the family reported outcome measure Catherine J. Golics, MPharm, Cardiff University, United Kingdom; Mohammad K.A. Basra, MD, Cardiff University, United Kingdom; Sam Salek, BSc RPh PhD FFPM, Cardiff University, Redwood Bldg, United Kingdom; Andrew Finlay, FRCP, Univ. of Wales College of Medicine, United Kingdom AIMS Research into family quality of life (QoL) has usually focused on families of patients with one specific disease, or the effects on one specific family member. We previously identified the key QoL themes which impact family members of patients with diverse medical conditions. The aim of this study was to use these data to create a selfreported outcome measure to assess this impact on family members.; METHODS Transcripts from 133 interviews with family members of patients were coded into themes and subthemes. Any subtheme mentioned by greater than 5 % of participants was included as an item in the draft instrument. Each item was phrased simply and concisely measuring one concept. An expert panel of 12 clinicians, 7 nurses, 3 family members and 3 QoL experts provided written and oral feedback on the following four attributes for each item: language clarity, completeness, relevance, and scaling. Feedback was used when designing the language and layout of the measure. Intraclass correlation (ICC) was performed on the written feedback using a standardised tick box form; RESULTS The expert panel gave detailed and positive feedback concerning most questionnaire items. Nine items were re-worded or merged and the response options were renamed to simplify the questionnaire. The panel’s ratings of each item on a 4-point scale for the four attributes showed either ‘‘strongly agreed’’ or ‘‘agreed’’ (88 %), with an ICC value of 0.98 (CI = 0.97–0.99) suggesting a high correlation between the panel members’ responses. A 31-item generic family quality of life instrument, the Family Reported Outcome Measure (FROM), with a 5 point likert response scale was developed. Three utility questions, including hours of sleep lost per night, hours spent looking after the patient, and hours spent on housework were also included in the draft questionnaire.; CONCLUSIONS Being able to measure the impact of illness on family members will enable more complete assessment of the wider burden of disease and of targeted interventions, hence improving provision of care for patients and families. Further refinement and validation of the FROM is underway.
(1051) Application methodology of the focal groups technique in deaf people who communicate in sign language Soraya Bianca Reis Durate, Ms, Instituto Federal de Educac¸a˜o, Cieˆncia e Tecnologia -IFG, Brazil; Adriana Ribeiro De Freitas, Pontifı´cia Universidade Cato´lica de Goia´s; Neuma Chaveiro, Federal University of Goias, Brazil; Celmo Porto, Universidade Federal de Goia´s; Maria Alves Barbosa, Universidade Federal de Goia´s; Marcelo P. Fleck, PhD, Hospital de Clinicas de Porto Alegre, Brazil AIMS Describe the focal groups technique used in the analysis of the translation/validation of the WHOQOL-BREF and the WHOQOL-
59 DIS versions in Libras.; METHODS Qualitative research of descriptive-exploratory transversal typology carried out using the focal groups technique. The sample for convenience was composed by nine deaf people who use Libras. The meetings were video recorded and transcripted later on. The first one was for the explanation of the quality of life project of the Worldwide Health Organization—WHO; the second was for the discussion of each question of the WHOQOLBREF and the WHOQOL-DIS versions in Brazilian Sign. At the end of each question, the understanding and relevance were discussed and suggestions about changings were note taken. Analysis of the groups was based upon Bardin content analysis.; RESULTS The focal groups technique with deaf people will be more successful if composed by no more than six components and longer period to be accomplished. In this study, the focal group of deaf people meeting last 3 h and 30 min and its components faced some difficulties in respecting rules such as objectivity, one at a time speaking and pair conversation avoidance, proposed by the focal groups technique for listeners. It shows some peculiar characteristics of the culture and way of expression of the deaf population.; CONCLUSIONS The focal groups technique, when applied to deaf people groups who use sign language, requires some adaptations in its rules, taking into consideration the cultural and linguistics peculiarities of the deaf people. The focal group mediation must be carried out by people who are fluent in sign language. Considering the sign language as viso-spatial, it is necessary to fully video record all the meetings to the register all the data.
(1052) Calling pain pain: what should we ask MS patients about pain severity? Shahnaz Shahrbanian, McGill University, Canada; Nancy E. Mayo, PhD, McGill University, Royal Victoria Hospital, Canada; Pierre Duquette, University of Montreal; Ayse Kuspinar, BSc(PT), MSc., McGill University, Canada; Shang Yuan Teng, MSc, McGill University, Canada AIMS The purpose of this study was to identify the extent to which four 0–10 numerical pain rating scales (NRS) (current, lowest, highest and average) capture sufficiently unique information as to warrant them all to be included in standard pain rating. In particular, does rating of average pain, contribute information over and above what is provided by lowest and highest pain?; METHODS Study population was a centre-stratified random sample of patients registered at the 3 MS clinics in Montreal comprising 139 women and 49 men. Patients were asked to rate their pain severity on 0–10 NRS, with 0 indicating ‘no pain’ and 10 indicating ‘pain as bad as it could be’. Using 4 separate scales, patients rated their worst, least and average pain over the past week, as well as their current pain at the time of evaluation. Participants also completed the 36-item Short Form Health Status Survey (SF-36) which is a measure of health related quality of life and considered to be as a reference of valid and reliable measure. Composite pain intensity scores were created for possible combinations of the single pain ratings by computing their mean. Bivariate correlation coefficients were then calculated between these composite scores, individuals rating scales, and SF-36.; RESULTS Of the 188 persons (age 42 ± 10; duration of disease: 8 years ± 3; median EDSS: 2.0), 42 % identified pain as a symptom. The mean values for bodily pain subscale of SF-36 was 68, lower than age expected norms of 76. The mean value for rating of current pain at the time of evaluation was 3.3 ± 2.3; mean of lowest pain severity was 2.2 ± 2; worst pain severity was 6.8 ± 2; and pain average was 5.0 ± 2. The patient rating of average pain, differed from that calculated from their lowest and highest, suggesting that different but unknown information is used to estimate the average. All metrics were correlated including the calculated average of lowest and worst. Interestingly, it was indicated that of all ratings, the patients’ ratings of worst pain was the most closely associated with the rating of average pain (r = 0.76). In
123
60 addition, they all correlated similarly with an external pain rating scale (SF-36) and they all correlated with similar magnitude with all of the other dimensions of health-related quality of life (H-QOL).; CONCLUSIONS As patient estimate of average pain was highly correlated to the calculated average, we recommend not asking patients to rate ‘‘average’’ pain and for research purposes calculate the average. It yields a good summary measure reducing 4 values to 1.
(1053) Comparison of the Dermatology Life Quality Index (DLQI) and EuroQol-5D (EQ-5D) for their use in economic studies of dermatological treatments Sam Salek, BSc RPh PhD FFPM, Cardiff University, Redwood Bldg, United Kingdom; Mohammad K.A. Basra, MD, Cardiff University, United Kingdom; Pedro Cruz, Cardiff University; Vicent Piquet, Cardiff University; Andrew Finlay, FRCP, Univ. of Wales College of Medicine, United Kingdom AIMS The use of the DLQI, a simple practical measure for routine clinical use, and the EQ-5D, a utility measure, is increasing in dermatology clinical research and a number of studies have used both instruments together. A formula has been derived from psoriasis patients’ data to describe the relationship between the two measures: EQ-5D utility score =0.956-[0.02489 (DLQI score)]. The aim of this study was to review the current published evidence concerning the relationship between DLQI and EQ-5D scores derived from studies on different dermatoses and to provide guidance for the conversion of the DLQI scores to utility values.; METHODS A systematic online and manual literature search was conducted to identify studies published in English that used both the DLQI and the EQ-5D together. The search terms included: DLQI, EQ-5D, skin disease, dermatology, quality of life, preference-based measures, utility measures.; RESULTS Twenty articles and three abstracts were identified that described the use of both the EQ-5D and DLQI in the same study. Six articles were cross-sectional studies designed to evaluate the QoL impact of hidradenitis suppurativa, vitiligo, generalised psoriasis and nail psoriasis. Two articles and one abstract described the cost-utility and cost-effectiveness of interventions. Three articles and two abstracts examined the relationship between PROs and clinical outcomes. Eight articles assessed the QoL improvement following treatment for psoriasis and one article following treatment for venous eczema. Correlation coefficients between the DLQI and EQ-5D were described only in six studies. Data were identified from the 14 studies that gave both DLQI and EQ-5D values and these data were used to calculate the overall mean correlation between the DLQI and EQ-5D i.e. r = 0.68. This relationship is represented by the equation: EQ-5D utility score = 0.9297-[0.0219 9 (DLQI score)].; CONCLUSIONS The findings indicate a moderate level of correlation between the DLQI and EQ-5D based on the published data. Moreover, there appears to be close similarity between the conversion formula derived from original research and the secondary analysis of published values. However, until a large prospective study is carried out to clarify the exact nature of the relationship, we propose the continued use of the previously published formula.
(1054) Evaluation of a newly developed questionnaire to measure satisfaction with health among people living with type 2 diabetes (T2DM): The Current Health Satisfaction Questionnaire (CHES-Q) Shana B. Traina, Janssen Global Services; Susan D. Mathias, MPH, Health Outcomes Solutions; Hilary H. Colwell, Health Outcomes Solutions; Ross Crosby, PhD, Neuropsychiatric Research Institute AIMS Health satisfaction may help predict adherence to diabetes selfcare behaviors, a cornerstone of optimal diabetes management. For
123
Qual Life Res (2012) 21:1–132 example, if a person living with T2DM is satisfied with his/her weight, even if it is clinically unhealthy, s/he may not attempt weight loss. Historically medication satisfaction is measured in this population, and there has been little focus on the concept of health satisfaction. Therefore, the 14-item CHES-Q was developed, via literature reviews, input from people living with T2DM, and healthcare professionals to measure satisfaction with aspects of health (e.g., blood sugar, weight) and knowledge related to T2DM. This study evaluated some psychometric properties of the CHES-Q.; METHODS An online survey targeting US adults self-reported as living with and taking medication for T2DM was implemented. The survey included questions about demographics, the CHES-Q, and SF-36, and was completed at baseline and 3–7 days later. Since each item of the CHES-Q measures distinct concepts, items were scored individually. Test–retest reliability [using intra-class correlation coefficients (ICC) for those reporting no change in health], convergent validity [via Pearson correlations with the SF-36], known-groups validity [based on BMI calculated using selfreported height and weight], and minimum important difference (MID) [using distribution-based approaches] were assessed.; RESULTS Of the 1,015 respondents, (n = 529 at 3–7 day retest), 60 % were male, mean age 63 years, 84 % white/non-Hispanic, 64 % married, and 59 % had a BMI [ 30. ICCs for 11 items were above the acceptable threshold of 0.70 (satisfaction with appetite, social interactions, and mood [0.60). Correlations between CHES-Q and related SF-36 items were moderate to high (range 0.49–0.73, p \ 0.01). There were significant differences between BMI groups (p \ 0.001) on 12 of 14 items (with exception of two knowledge items). Combining distribution-based results yielded a 1-point MID estimate for each item.; CONCLUSIONS The CHES-Q has acceptable psychometric properties; however, if used as an outcome measure, its responsiveness should be evaluated. Measuring health satisfaction among people living with T2DM using the CHES-Q may be useful when personalizing treatment approaches that require adherence to self-care activities.
(1055) Assessment of healthcare needs among crime victims in Japan Kazuaki Hirai, Oita University, Japan; Takayuki Kageyama, Oita University of Nursing & Health Sciences; Namiko KAWAMURA, Oita university AIMS Crime victims are often physically, mentally, and socially invaded, being likely to suffer from mental disorders, such as depression and PTSD. To help them, the Japanese national government begins to organize crime victims support center (CVSC). Most of CVSC is managed by nonprofit organizations. Although previous reports from CVSC suggest that many crime victims have not visited doctors regardless of their healthcare needs, the detail has remained unclear. The aim of this study is to examine healthcare needs among crime victims who visited CVSC, and to know how the needs are assessed by supporters of CVSC.; METHODS We carried out semi-structured interview to the supporters of a CVSC in Japan. They are not public servants, but nonprofessionals with government-sponsored training. Their main roles are consultation, crisis intervention, defending clients against recidivism, and public relations. Eight supporters of CVSC, two males and six females aged 33–65, participated. They gave anonymous information about the crime victims with healthcare needs in the last 1 year. They also reported the reason why they judged the clients had healthcare needs.; RESULTS Out of 324 clients in the last 1 year, sixteen had healthcare needs. They perceived healthcare needs when they found clients’ suicidal thought or behaviors, physical injury, or mental symptoms such as depression and delusion of persecution. Approximately 44 % (seven cases) of clients with healthcare needs didn’t go to doctors, because of personal reasons such as reduced capacity to make decision, and environmental factors such as lack of
Qual Life Res (2012) 21:1–132 cooperation by healthcare institutions. This was also related to the fact that CVSC cannot force the clients to go to doctors. The supporters were, however, not confident in their judge for healthcare needs. Even if they referred clients to clinics, they received no feedback from the clinics concerning their judge for the reference or prognosis of the clients.; CONCLUSIONS To help the supporters of CVSC to find healthcare needs among crime victims, not only education for them but systematic countermeasures are important. Cooperation with medical institutions, particularly for providing the information about prognosis of referred clients, and supervision systems for the supporters should be established.
(1056) The Utrecht Scale for evaluation of rehabilitation-participation (USER-participation) and its psychometric properties Carlijn H. Van Der Zee, MSc, Rehabilitation Center De Hoogstraat, Netherlands; Marcel Post, PhD, Institute for Rehabilitation Research, Netherlands; Johanna MA Visser-Meily, University Medical Center Utrecht; Eline Lindeman, University Medical Center Utrecht, Netherlands AIMS Participation is an important outcome of outpatient rehabilitation programmes. It is defined as ‘the involvement in a life situation’ and is a multidimensional concept, consisting of an objective and a subjective dimension. Objective participation concerns observable behaviours and subjective participation concerns the subjective appraisal of the person involved. There are many instruments to measure participation, however, instruments measuring both objective and subjective participation are sparse. Therefore, the USER-Participation was developed.; METHODS The USER-Participation consists of 31 items measuring three aspects of participation; Frequency of behaviours (Frequency scale), experienced participation Restrictions (Restriction scale), and Satisfaction with participation (Satisfaction scale). All scales comprise items on vocational, leisure and social activities. For all three scales a separate score is calculated and there is no USER-Participation total score. In a series of studies in a multi-diagnostic group of rehabilitation outpatients we assessed reproducibility, construct validity, concurrent validity, discriminant validity, and responsiveness. The USER-Participation was compared to the Frenchay Activities Index, the Participation subscale of the ICF Measure of Participation and Activities: Screener, and the Participation Scale.; RESULTS Test–retest reliability was good for the Restriction and Satisfaction scales (intraclass correlation coefficient = 0.85 and 0.84, respectively) and agreement was good for all three scales (standard error of the measurement 5.5–6.8 % of the scale range; ratio smallest detectable change/standard deviation 0.16–0.24). Construct validity was good (Cronbach’s a 0.70–0.91; Spearman correlations between the USER-Participation scales 0.36–0.52), and the Restriction and Satisfaction scales showed a strong first factor explaining more than 50 % of the variance. Concurrent validity was shown by expected strong correlations between the USER-Participation scales and three other participation measures (range 0.59–0.75). Discriminant validity was shown by significant differences in USER-Participation scores between participants with different levels of independence (p range \ 0.001–0.063) and between participants with different health conditions (p \ 0.001). Responsiveness was moderate (ES range 0.19–0.49; SRM range 0.21–0.54), but better than responsiveness of the other participation measures.; CONCLUSIONS The USER-Participation is a psychometrically good measure to rate the multidimensional concept of participation in persons with physical disabilities and can be used both for clinical and research purposes.
61 (1057) Cost-effectiveness analysis of anti-viral treatments in patients with chronic hepatitis B infection in Hong Kong T.P. Elegance Lam, PhD, The University of Hong Kong, Hong Kong; Cindy L.K. Lam, MD, The University of Hong Kong, Peoples Republic of China; C.L. Lai, The University of Hong Kong; M.F. Yuen, The University of Hong Kong, Hong Kong; Sarah McGhee, The University of Hong Kong, Hong Kong; Daniel Fong, PhD, The University of Hong Kong, Hong Kong; Gabriel Leung, The University of Hong Kong, Hong Kong AIMS To estimate the cost-effectiveness, in terms of cost per quality adjusted life years (QALY), of anti-viral treatments in preventing the disease progression and mortality from the asymptomatic hepatitis B (AHB) carrier state to complications.; METHODS A cost-effectiveness analysis, using a computer-base micro-simulation model, of antiviral treatments for the prevention of the disease progression and mortality, from AHB carrier to cirrhosis and hepatocellular carcinoma (HCC) was developed and different therapeutic strategies were evaluated: (1) no anti-viral treatment; (2) lamivudine (LVD) monotherapy; (3) adefovir dipivoxil (ADV) monotherapy; and (4) entecavir (ETV) monotherapy.; RESULTS The cost per QALY gain from LVD treatment was HKD 29,010. Compared with LVD, the use of ADV and ETV were more associated with more QALYs gain at an incremental cost of HKD 223,604 and HKD 149,293 per QALY, respectively.; CONCLUSIONS This is the first study on the cost-effectiveness of anti-viral treatment of chronic hepatitis B (CHB) infection in Hong Kong. The results can provide evidence for or against anti-viral treatment for CHB carriers in clinical practice and provide information on the cost-effectiveness of different anti-viral treatments in preventing the disease progression from CHB carriers to cirrhosis or HCC. Antiviral treatment is recommended for the prevention of disease progression from CHB to liver-related complications to improve survival.
(1058) Optimizing satisfaction in breast reconstruction: the importance of the preoperative information and patient-physician communication Adelyn L. Ho, MD MPH, University of British Columbia, Canada; Anne Klassen, D Phil, McMaster University, Canada; Stefan Cano, BSc PhD, Peninsula College of Medicine & Dentristry, United Kingdom; Amie M. Scott, MPH, Memorial Sloan-Kettering Cancer Center; Andrea Pusic, MD MHS FRCSC, Memorial Sloan Kettering Cancer Ctr AIMS In breast reconstruction, we currently know very little about the extent to which process of care (POC) factors, such as satisfaction with information, with interactions with the plastic surgeon, and with the medical team, influence patient outcomes. Patient-reported POC measures are important for quality improvement efforts as they can highlight modifiable areas to improve patient outcomes. In this study we examined the relationship between satisfaction with POC and the outcome of surgery using the BREAST-Q.; METHODS A multicentre, cross-sectional study design was employed. The association between satisfaction with POC measures and both satisfaction with breasts and satisfaction with outcome were evaluated using linear regression analysis. Multivariate regression models were constructed to control for confounders of this relationship and to identify predictors of satisfaction with breasts and overall outcome.; RESULTS A total of 571 reconstruction patients completed the questionnaire (response rate 66 %). The 510 patients meeting inclusion criteria for analysis are characterized by a mean age of 54.3 ± 9.3 (21–81) and a mean BMI of 25.2 ± 4.3 (16.3–48.9). Satisfaction with information and plastic surgeon both predicted a higher satisfaction with breasts,
123
62 after controlling for age, BMI, complications, study centre, adjuvant radiotherapy, and procedure type (information p \ 0.001, surgeon p = 0.003; R2 = 0.29). Saline implants (vs. silicone) and adjuvant radiation therapy independently predicted a lower satisfaction with breasts. In the model that evaluated satisfaction with overall outcome, both satisfaction with information and plastic surgeon remained significant, after adjusting for relevant covariates (information p \ 0.001, surgeon p \ 0.001; R2 = 0.31). Advanced age, higher BMI, and the occurrence of major complications were independently associated with a lower satisfaction with outcome.; CONCLUSIONS The quality of information that a patient received and satisfaction with their plastic surgeon were important POC measures that influenced patients’ satisfaction with their breasts and overall outcome. The implications of these findings are that interventions to improve communication of information for breast reconstruction patients as well as factors that promote a strong patient-surgeon relationship can have an important impact on patient outcomes. Research is needed to delineate the factors that will improve satisfaction with information and the surgeon-patient relationship.
(1059) The use of secondary qualitative analysis to establish domains of patient-centred care in people with long term conditions Monica Hadi, PhD, University of Oxford, United Kingdom; Louise Locock, University of Oxford, United Kingdom; Sara Ryan, University of Oxford; Raymond Fitzpatrick, PhD, Univ of Oxford, Dept of Public Hlth, United Kingdom; Elizabeth J. Gibbons, MSc, University of Oxford, United Kingdom AIMS Secondary qualitative analysis (SQA) is increasingly used in social research. It involves using pre-existing data to investigate a new research question. However, the method remains problematic and ill-defined. This research aims to describe in detail the use of SQA to identify domains of patient-centred care from experiences of people living with long-term conditions and highlight the advantages and limitation of SQA.; METHODS The primary data is a collection of narrative interviews from Health Experience Research Group, University of Oxford. Five chronic conditions were chosen for analysis; Rheumatoid Arthritis, Type 2 Diabetes, Asthma, Heart Failure and Epilepsy (100 interviews). A literature review informed the development of a primary coding framework; a set of anticipated domains. By means of Heaton’s (2000) five categories of SQA, this study falls into the categories of Supra-analysis, Amplified analysis and Supplementary analysis.; RESULTS Interviews were coded using a modified grounded theory approach, looking for both anticipated and emergent themes. Constant comparison was used to refine the coding framework. This was done after each interview and included the addition of new codes in the framework and merging, expanding and relating codes. A codebook was created, and this, together with analytic memoing helped to determine the point at which data saturation had been reached. The second step of the SQA involved further in depth analysis of the codes extracted from interviews. A visual mindmap technique was used to seek analytic depth in the data and gather new insights and interpretations. The main concerns of SQA are ‘lack of contextual knowledge’, access to ‘the bank of knowledge’ and ‘fit of data’ to the new research question. However, there is great potential for re-analysing material and SQA saves considerable time, which allows for deeper analysis and understanding of the primary data. Ethical approval may also not be required and further investigation will not burden participants.; CONCLUSIONS SQA is a useful, cost-effective method which decreases respondent burden. It is beneficial for the researcher to be aware of and address limitations of this method before commencing analysis. Results from SQA will inform the development and evaluation of a Patient Reported Experience Measure for people with Long-term conditions.
123
Qual Life Res (2012) 21:1–132 (1060) Does the Hospital Depression and Anxiety scale tick the boxes in rheumatology: a Rasch measurement theory analysis Sophie Cleanthous, MSc BSc, University College London, United Kingdom; Stanton P. Newman, City University, United Kingdom; Michael Shipley, University College London Hospital; David A. Isenberg, University College London; Stefan Cano, BSc PhD, Peninsula College of Medicine & Dentristry, United Kingdom AIMS PRO instruments are increasingly being used to evaluate the impact of rheumatological conditions and their treatment. The general recommendation is that both generic and disease-specific instruments should be used in clinical research, and that psychometric evidence for their use is available. However, despite this being common place for disease-specific instruments, their generic counterparts are frequently used based on the often untested assumption that they satisfy psychometric requirements across different disease groups. In this study, we examined the psychometric properties of a widely used mood scale, the Hospital Depression and Anxiety scale (HADS), in rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE).; METHODS We carried out a cross-sectional observational study across five hospital trusts in the UK involving people with RA and SLE. The HADS was completed via a postal survey or during outpatient appointments. We analysed HADS data based on Rasch Measurement Theory (RMT) methods at the subscale-level (HADS Anxiety, HADS Depression) including examinations of: item fit, targeting, item dependency, reliability and stability. We used RUMM2030.; RESULTS A total of 383 patients were recruited (210 RA, 171 SLE; mean age 52; mean years since diagnosis: 12). The HADS Anxiety and HADS Depression scales performed adequately, in both RA and SLE samples, in relation to item dependency, reliability and stability. However, we found 3/14 misfitting items (HADS Anxiety restless, fright; HADS Depression book). This suggests that these items were poor at discriminating different levels of mood. In addition, we found scale-to-sample targeting to be sub-optimal for both HADS subscales. Essentially both subscales exhibited significant ceiling effects.; CONCLUSIONS Our RMT analysis of the HADS uncovered two key issues relating to the potential problems with the some of the content of the HADS subscales and content coverage in RA and SLE. Importantly, the sophisticated techniques and methods offered by RMT provide an initial evidence-base to guide any future improvement of this widely used instrument. Our study also further underlines the importance of establishing the psychometric properties of generic instruments before their use in clinical research.
(1061) Quality of life assessment in patients with idiopathic normal pressure hydrocephalus Hideki Origasa, PhD, Univ of Toyama at Sugitani School of Medicine, Japan; Masaaki Hashimoto, MD, Department of Neurosurgery, Noto General Hospital, Japan; Masakazu Miyajima, MD, Department of Neurosurgery, Juntendo University School of Medicine, Japan; Hajime Arai, MD, Department of Neurosurgery, Juntendo University School of Medicine, Japan AIMS Idiopathic normal pressure hydrocephalus (iNPH) is a syndrome characterized by disturbed gait, cognition, and urinary symptoms. Although the iNPHGS (iNHP grading score) consisted of the above three symptoms has been applied, quality of life (QOL) might be an essential feature for each patient and/or family. Since no challenges have been conducted in QOL for iNPH, we first investigated the orientation of QOL for this disease. METHODS We used EQ-5D as a measure of overall QOL and analyzed the relationship with nursing levels, modified Rankin scale (mRS), and iNHPGS with its three components. Each component was scored from 0 (none) to 4
Qual Life Res (2012) 21:1–132 (severe) that was scaled from 0 to 100 points for easy interpretation. We also investigated the difference between the response by patients and caregivers. Investigation was conducted at two clinical centers in late 2011. A total of 37 patients answered the questionnaires. RESULTS Mean QOL measured by EQ-5D was 0.71 (IQR: 0.62–0.85). Mean age was 76 (Range 60–89) and 57 % were women. Subjects included 10 independence, 5 requiring support, 22 needed long-term care on the nursing level. 85 % of them had mRS of 1–3 (mild to moderate). Mean gait score was 1.8 (45 points), cognition was 1.5 (38 points), and incontinence was 1.2 (30 points). QOL decreased according to older ages, yet there was no gender difference. QOL and its subscales were not associated with both nursing levels and iNPHGS. QOL was similar among the points 1–3 of mRS, yet remarkably decreased in patients with mRS of 4–5 (severe, necessary care). 44 % of patients differently responded with caregivers on the rating of QOL subscales. CONCLUSIONS Little attentions have been focused upon the QOL for iNPH. This study showed the QOL seemed to be independent of conventionally used severity score (iNPHGS). Discrepancy in responses between patients and caregivers implied that we should make a careful survey for both respondents.
(1062) Attitude of medical recipients (patients) and medical providers (nurses and doctors) toward life and death Yujiro Kuroda, the University of Tokyo Hospital, Japan; Keiichi Nakagawa, the University of Tokyo, Japan; Yumi Iwamitsu, Kitasato University; Midori Kotani, Dai-ichi Life research institute; Yutaka Kitazawa, Toyo Eiwa University; Yoichi Yamada, University of Tsukuba; Osamu Sakura, The University of Tokyo AIMS This study compared the attitude of health care providers, that is, physicians and nurses, and of recipients, that is, cancer patients, toward life and death.; METHODS From January to December 2008, a self-rated questionnaire survey was administered to 220 physicians and 470 nurses and to 450 cancer patients undergoing chemotherapy at a university-based hospital. Of these, 310 cancer patients (female patients, 41 %), 176 physicians (women, 14 %), and 366 nurses (women, 96 %) participated in the study (response rate, 68, 80, and 78 %, respectively). The questionnaire, which has been developed by a Japanese researcher (Hirai, 2007), contained 27 items and 7 domains (afterlife belief, death anxiety, death relief, death avoidance, life purpose, death concern, and supernatural belief) examining the attitude toward life and death. We calculated the score for each domain and then performed oneway analysis of variance (ANOVA) for the 3 groups; RESULTS The main cancer types were uterine cervical cancer (12 %) and pulmonary cancer (10 %), and 81 % of the patients had developed metastasis. ANOVA showed statistically significant differences in the following domains: The nurses had higher scores for afterlife belief and death concern than the doctors (F2,811 = 66.65, p \ .01) and patients (F2,807 = 13.36, p \ .01). This suggests that nurses often think about the deaths of their loved ones and their own deaths than doctors and patients. The scores for supernatural belief and death relief were higher for the nurses (F2,809 = 24.49, p \ .01) and patients (F2,800 = 7.49, p \ .01) than for the doctors. This suggests that nurses and patients often believe that death leads to relief from pain and suffering. The score for death avoidance was higher for the patients than for the doctors and nurses (F2,801 = 20, p \ .01).; CONCLUSIONS The results showed that the attitude toward life and death differ between doctors and nurses who provide medical treatment as well as between them and patients who receive medical treatment. This information is important for medical professional education and patient education in terms of understanding the different perspectives of others.
63 (1064) An improvement of the established TCM-SAQ: another syndrome-specific instruction to assess quality of life for Angina treated by traditional Chinese medicine. (QBS) Chen Zhihui, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Lv Meijun, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yu Changhe, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Ma Yuan, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Wang Yang, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS We have already developed a syndrome specific questionnaire to assess quality of life treated by Traditional Chinese Medicine (TCM): Traditional Chinese Medicine Syndrome of Angina Questionnaire, Phlegm and Blood Stasis syndrome (TCM-SAQ-PBS), which examined to be reliable and valid. However, there are kinds of syndromes for the same disease in TCM. Therefore, an alternative ‘‘syndrome’’ domain for Qi Deficiency and Blood Stasis (QBS) syndrome was chosen to replace that in TCM-SAQ-PBS. Thus another syndrome-specific questionnaire for QBS was developed to enrich TCM-SAQ.; METHODS The content of this questionnaire was developed on the basis of a former established syndrome-specific questionnaire TCM-SAQ-PBS. There are 30 items within 6 domains (angina, activity, sleep, worry, satisfaction, syndrome) in TCM-SAQPBS. We extracted the first 5 domains from TCM-SAQ-PBS entirely. Then we replaced the items in ‘‘syndrome’’ domain in PBS by items specifically for QBS syndrome. A draft version containing 32 items composed 6 domains was formed after cognitive interview of 20 inpatients. A cross-sectional study was conducted to test TCM-SAQQBS. The validity and reliability of the new domain were tested by measurement properties especially. The psychometric properties of the whole questionnaire also tested.; RESULTS 10 items were added into ‘‘syndrome’’ domain, including 3 facets (angina influenced by time change, Qi deficiency, blood stasis). In the cross-sectional study, 789 patients (age 59.50 ± 9.692 year; 360 male, 45.6 %) completed the questionnaire. The Spearman correlation coefficient (CC) between 3 facets ranged from 0.295 to 0.408 (p = 0.000), and CC with the global domain ranged from 0.556 to 0.943, which reflects adequate content validity of this new adding domain. The Cronbach’s alpha was 0.837 of the 10 items, which shows good internal consistency. The CC with the global scores reach to 0.702, and ranging from 0.085 to 0.490 (p \ 0.005) for other 5 domains, which displays relatively independence and significant importance. As to the whole questionnaire, test–retest CC of 240 (30.03 %) patients was 0.953. Cronbach’s alpha coefficient was 0.865, and ranging from 0.812 to 0.963 for each domain. The CC between TCM-SAQ-QBS and SAQ was 0.535. Nine factors were extracted in EFA, and were stratified into 6 domains as expected from the conceptual framework.; CONCLUSIONS The new developed syndrome-specific questionnaire TCM-SAQ-QBS has good psychometric properties, and can be used as an outcome for angina in TCM clinical practice. The method employed to develop questionnaire is properly reasonable.
(1065) Developing a core outcome set for oesophageal cancer surgery: the role of patient-reported outcomes Natalie Blencowe, University of Bristol; Rhiannon C. Macefield, University of Bristol, United Kingdom; Marc Jacobs, MSc, Amsterdam Medical Center, Netherlands; Sean Strong, Bmed,
123
64 MRCS, MSC, University of Bristol, United Kingdom; Sara Brookes, University of Bristol; Mirjam AG Sprangers, PhD, Academic Medical Center, Netherlands; Jane Blazeby, MD, University of Bristol, United Kingdom AIMS Accurate outcome measurement is necessary to allow crosscentre comparisons, inform data synthesis and to provide patients with information for decision-making. If outcomes are not accurately defined, measured and reported, outcome reporting bias may occur. There are however, many different clinical and PROs that may be measured in research studies. The development of a ‘core outcome set’: a minimum set of outcomes selected by key stakeholders, to be reported in all trials of a condition or procedure, may help improve outcome reporting. This study aimed to identify and classify all clinical and PROs of oesophageal cancer surgery, which will subsequently be used to develop a core outcome set; METHODS Systematic literature reviews identified clinical and PROs of oesophagectomy. These were supplemented with outcomes identified from a UK oesophago-gastric national audit database, patient information leaflets and consent consultations between surgeons and patients. Outcomes from all sources were independently grouped into domains by at least two clinicians or researchers. Outcome domains were separately and independently verified by patient advisers. All domains were then compared, duplicates removed and the final set developed into questionnaire items. RESULTS Literature reviews identified 380 different outcomes (248 clinical and 132 PROs). Clinical outcomes were extracted from 17 randomised and 105 non-randomised studies and PROs extracted from 4 randomised and 51 non-randomised studies respectively. A further 308 different outcomes were identified from the patient information leaflets and 13 from the national audit, meaning that 701 different outcomes were identified in total. These were all grouped into 67 outcome domains (24 clinical, 32 PRO and 11 process domains), and displayed as 67 separate items in the questionnaire. CONCLUSIONS Although more clinical outcomes of oesophageal cancer surgery were identified than PROs, these were categorised into fewer outcome domains meaning that questionnaire items were focused on PROs. Further work will reduce these domains to a core set to improve outcome reporting in clinical trials.
(1066) Cross-cultural adaptation and validation of the Portuguese version of the Pregnancy Physical Activity Questionnaire (PPAQ) Rui Soles Goncalves, PhD, Centre for Health Studies and Research of the University of Coimbra, Portugal; Filipa Pombo, Alcoita˜o Health College; Ana Cristina Silva, Alcoita˜o Health College; Anto´nio Fernandes Lopes, Alcoita˜o Health College; Luis Manuel Cavalheiro, PhD, University of Coimbra, Portugal; Joao Antonio Gil, PT PhD, University of Coimbra, Portugal; Pedro L. Ferreira, PhD, University of Coimbra, Portugal AIMS The objective of this study was to translate and culturally adapt the Pregnancy Physical Activity Questionnaire (PPAQ) to the Portuguese language and to test its reliability and validity.; METHODS This new version was obtained with forward/backward translations, consensus panels and a pre-test. The Portuguese PPAQ and International Physical Activity Questionnaire (IPAQ) questionnaires and a form for the characteristics of the participants were applied to 114 pregnant women.; RESULTS Reliability was good with intraclass correlation coefficients between 0.73 and 0.91 (0.89 for total activity), standard error of measurement ranging from 1.95 to 22.07 MET-h/ week (31.95 MET-h/week for total activity), and smallest detectable change ranging from 5.39 to 61.18 MET-h/week (88.57 MET-h/week for total activity) for the PPAQ subscales. Construct validity was supported by the confirmation of predefined hypotheses involving expected significant correlations between PPAQ and IPAQ subscales that represent similar constructs.; CONCLUSIONS The Portuguese
123
Qual Life Res (2012) 21:1–132 PPAQ exhibited suitable psychometric properties, in terms of reproducibility, measurement error and construct validity.
(1067) Reliability and validity of a new measure to evaluate mobility in the elderly (Functional Mobility Scale—elderly and very elderly) in the Portuguese population Irene Higgs, Alcoita˜o health college; Luis Manuel Cavalheiro, PhD, University of Coimbra, Portugal; Jose´ Milla´n Calenti, University of A Corun˜a; Rui Soles Goncalves, PhD, Centre for Health Studies and Research of the University of Coimbra, Portugal; Pedro L. Ferreira, PhD, University of Coimbra, Portugal; Joao Antonio Gil, PT PhD, University of Coimbra, Portugal AIMS The objective of this study was to test the reliability and validity of a measure designed to evaluate mobility in the elderly aged 75 years or older: Functional Mobility Scale—Elderly and Very Elderly (FMSEVE).; METHODS The FMS-EVE was obtained in a 3-phase study: (1) generation of a set of items based on a literature review; (2) Item analysis by a multidisciplinary panel of health care and human movement professionals with expertise in geriatrics, who identified the relevant items; (3) through a focus group methodology 10 experts (8 physiotherapists and 2 occupational therapists) selected and grouped 34 items related to mobility in the lying, sitting and standing positions and to mobility in locomotion and walking. The content validity was obtained by a Delphi panel composed of 15 experts with an average agreement of 90.3 % related to the items, the instructions and the scoring system. The FMS-EVE, the Katz Index and data from the individual, clinical and cognitive characteristics, were collected in a sample of 214 elderly (age = 86.69 ± 6.29 years; 78 % females) institutionalized at the Center for Social Support of Oeiras, Portugal.; RESULTS The factor analysis revealed one model of 2 components that explained 89.4 % of the total variance. One of these components was saturated with all items related to mobility in standing and walking (61.0 %), and it was called Standing and Motion Mobility. The other component, called Lying and Sitting Mobility, included all items related to the lying and sitting positions (28.4 %). This structure of the measure revealed high internal consistency (alpha = 0.98), discriminative capacity between independents and dependents in ADL’s (rho = -0.82, p \ 0.001) and high inter-observer reliability (Kendall’s tau between 0.94 and 0.96).; CONCLUSIONS The results suggest that the FMS-EVE is composed by a 2 factor model with good psychometric properties. Further studies need to be developed with different samples and contexts in order to assess the stability and strength of this factorial structure.
(1068) Psychometric validation of patient-reported outcome measures for assessing chronic constipation symptom severity: results from phase 2b and phase 3 trials of linaclotide Mollie J. Baird, MPH, Ironwood Pharmaceuticals; Robyn Carson, MPH, Forest Research Institute; Cheryl D. Coon, PhD, RTI Health Solutions; Sheri Fehnel, PhD, RTI Health Solutions; James MacDougall, Ironwood Pharmaceuticals; Lauren Nelson, RTI Health Solutions; Stavros Tourkodimitris, Forest Research Institute; Valerie Williams, PhD, RTI Health Solutions; Jeffrey M. Johnston, Ironwood Pharmaceuticals AIMS In accordance with the FDA guidance on patient-reported outcomes (PROs), psychometric analyses were conducted to establish the reliability, validity, responsiveness, and thresholds of a set of symptombased PRO measures assessing the severity of chronic constipation (CC). Data from one Phase 2b and two Phase 3 clinical trials of linaclotide in CC were analyzed.; METHODS Eight daily PRO measures
Qual Life Res (2012) 21:1–132 of abdominal and bowel symptoms and two weekly global PRO items were administered using interactive voice response system technology. Intraclass correlation coefficients (ICCs), Pearson correlations, factor analyses, test statistics, and effect sizes were computed to evaluate the reliability, construct validity, discriminating ability, and responsiveness of the PRO measures in a clinical trial context. Anchor- and distribution-based methods, along with cumulative distribution plots, were used to establish thresholds of meaningful change. The Phase 2b, multicenter, randomized, double-blind, dose-range–finding, parallelgroup study compared placebo and four doses of linaclotide taken once daily for 4 weeks (N = 310). The two Phase 3, multicenter, randomized, double-blind, parallel-group trials compared placebo and two doses of linaclotide taken once daily for 12 weeks (N = 1,272).; RESULTS The CC PRO items demonstrated satisfactory test–retest reliability (all ICCs [ 0.68). Correlations among the CC PRO items and with other measures were similar across trials and provided evidence of the convergent and divergent validity of the PRO items. Factor analyses produced two highly correlated factors reflecting abdominal symptoms and bowel symptoms (r = 0.414). Known groups analyses were statistically significant and in the expected direction, substantiating the ability of the CC PROs to differentiate between subgroups of interest. Responsiveness statistics for the bowel symptom PROs were large (0.79–3.29), and for the abdominal symptom PRO items, moderate to large (-0.54 to -1.21). While threshold estimates for defining responders varied somewhat across methods, they were consistent across trials.; CONCLUSIONS The psychometric analyses strongly support the reliability, validity, and responsiveness of the CC symptom PROs, with the Phase 3 trials confirming the results as initially documented with the Phase 2b data. Threshold estimates of the CC symptom PROs for classifying responders in future clinical trials were similar using different methods recommended in the FDA guidance.
(1069) Psychometric evaluation of patient-reported outcome measures assessing IBS-C symptom severity and change: results from two randomized, double-blind, placebo-controlled phase 3 trials of linaclotide Robyn Carson, MPH, Forest Research Institute; Mollie J. Baird, MPH, Ironwood Pharmaceuticals; Cheryl D. Coon, PhD, RTI Health Solutions; Sheri Fehnel, PhD, RTI Health Solutions; James MacDougall, Ironwood Pharmaceuticals; Lauren Nelson, RTI Health Solutions; Stavros Tourkodimitris, Forest Research Institute; Valerie Williams, PhD, RTI Health Solutions; Jeffrey M. Johnston, Ironwood Pharmaceuticals AIMS Patient-reported outcomes (PROs) are necessary to assess symptom severity in irritable bowel syndrome with constipation (IBS-C). In accordance with the FDA guidance on PROs, psychometric analyses were conducted to establish the reliability, validity, responsiveness, and threshold for meaningful change in a set of symptombased PRO measures.; METHODS 1,602 adult patients with IBS-C participated in two Phase 3 multicenter, randomized, double-blind, parallel-group clinical trials of linaclotide (290 lg) or placebo administered orally once daily. Nine daily PRO measures of abdominal and bowel symptoms and four weekly global PRO items were administered using interactive voice response system technology. Test–retest reliability (intraclass correlations), construct validity (Pearson correlations and factor analyses), discriminating ability (analyses of variance), and responsiveness (Guyatt’s statistics) were evaluated. Anchor- and distribution-based methods, and cumulative distribution plots, were used to establish thresholds of meaningful change.; RESULTS The IBS-C PRO items demonstrated satisfactory test–retest reliability (all intraclass correlations [0.75). Correlations among the IBS-C PRO items and with other measures provided evidence of the convergent and divergent validity of the PRO items. Factor analyses resulted in two
65 highly correlated (r = 0.63) factors—one containing abdominal symptoms, the other containing bowel symptoms. Hypothesis tests, based on a variety of responder definitions, were all statistically significant and in the predicted direction, substantiating the discriminating ability of the IBS-C PRO items. With few exceptions, responsiveness statistics were highly satisfactory (range 0.46–3.07), demonstrating that the items were easily capable of detecting change. Threshold estimates for defining responders were similar across the different trials.; CONCLUSIONS The psychometric analyses based on the two Phase 3 trials of linaclotide strongly support the reliability, validity, and responsiveness of the IBS-C symptom PRO measures. Threshold estimates for defining responders were consistent with the FDA-recommended interim endpoint for IBS-C trials.
(1070) Self-efficacy impacts quality of life: a methodological study in subject with multiple sclerosis Rosalba Rosato, PhD, University of Turin, Italy; Giorgia Molinengo, University of Turin, Italy; Silvia Testa, University of Turin; Alessandra Oggero, Referral Regional Multiple Sclerosis Center, Italy; Antonio Bertolotto, Referral Regional Multiple Sclerosis Center, AOU San Luigi, Orbassano, Turin AIMS Previous studies have underline that psychological interventions with the goal of improving health-specific aspects of selfefficacy have a great impact on quality of life (QoL) of multiple sclerosis (MS) patients. Self-efficacy is commonly proposed as being domain and task specific but Schwarzer argues that it can also be conceptualized as a generalized sense of self-efficacy that refers to a global ability to cope with a wide range of situations. This study aims at evaluating the relationship between general self-efficacy and QoL in MS patients taking into account socio-demographic and clinical information’s. This is, in our knowledge, the first attempt to study general self-efficacy in MS patients.; METHODS One hundred-fifty five patients suffering of MS were invited to fill in the MSQoL-54 and the general self-efficacy (GSE) questionnaires. Socio demographic and clinical characteristics were also obtained. Two multiple regression models were performed with respectively mental and physical composite score as dependent variables adjusting for clinical and demographic information. To better understand in which manner GSE clinical and demographic variables were linked together and with the two composite QoL scores two structural equation models were applied using only covariates with a statistically significant effect on regression analysis; RESULTS The multiple regression analysis underlined that socio-demographic variables (gender, marital status, education level and age) didn’t influence QoL. While clinical variables (EDSS, disease duration, changes on job or social habits) affect QoL. The strength of the influence of GSE on QoL is high for both physical and mental composite. Structural equation models showed that disease duration has great positive impact on EDSS, a positive moderate impact on QoL and changes on social habits, while it has no effect onjob modification. The two models differ as regard the impact of GSE on QoL: on mental QoL the GSE variable has the greatest influence.; CONCLUSIONS The finding of this study have practical implications for professions working with patients with MS confirming that cognitive remediation programs aimed at improving cognitive skills may also improve quality of life for patients with MS.
(1071) Stories of life with life-threatening illness Kara L. Schick Makaroff, PhD, RN, University of Alberta, Canada; Anita Molzahn, PhD RN, University of Alberta, Canada; Laurene Sheilds, University of Victoria; Anne Bruce, University of Victoria; Kelli Stajduhar, University of Victoria; Rosanne Beuthin, Vancouver Island Health Authority; Sheryl Shermak, University of Victoria
123
66 AIMS Advances in science and healthcare have enabled people with life-threatening illness to live longer and healthier lives. However, treatment outcomes are not certain and the experience of living in-between the promise of prolonged life and the possibility of recurrence or progression of the illness is not well understood. Traditional evidence-based approaches to life-threatening illness have not necessarily improved patient outcomes. However, if we are truly committed to using patient reported outcomes in clinical practice, we need to understand the stories of people living with the uncertainty of life-threatening illness. The importance of patient reported outcomes is growing, and there is evidence that stories can provide insight into how to provide better care. The aims of this research were to explore people’s stories of facing life-threatening illness and how experiences of uncertainty affect their understandings of health and living within the context of cancer, chronic kidney disease (CKD) and HIV.; METHODS Thirty-two participants took part in four in-depth interviews over 3 years. Participants included ten people with cancer, fourteen people with CKD, and eight people with HIV. The participants were diverse with respect to age (mean = 62), gender (18 male, 14 female), time from diagnosis (1–50 years), and prognosis (from remission to palliative). The research team engaged in narrative thematic analysis to elucidate the multitude of tensions permeating participants’ narratives.; RESULTS The researcher team found that trajectories of illness were continually changing with some people living in a state of relative wellbeing, others living with increasing burdens of illness, and still others experiencing life-threatening events and sometimes progressing to palliative care. Moreover, participants made sense of the uncertainty of life-threatening illness through narratives to communicate information/experience, manage the responses of others (including health care providers), and foreground what was important to them. These narratives highlight the complexity and at times ineffable nature of experiences of life-threatening illness.; CONCLUSIONS The significance of meta-narratives (e.g., medical, survivorship, stigma of illness, and death) that shape and define patient reports and experiences was evident in our research. These findings challenge assumptions that health narratives are complete, unified, consistent, and enduring.
(1072) Streamlining aged care assessment Janet E. Sansoni, MSc, University of Wollongong, Australia; Peter Samsa, University of Wollongong; Alan Owen, University of Wollongong; Kathy Eagar, University of Wollongong AIMS In Australia elderly people can be assessed and receive basic Home and Community Care services to assist them remaining in their community residence. If more frail they may be referred for a comprehensive assessment, currently undertaken by aged care assessment teams (ACATs). This enables access to more substantive community care and extended care at home packages, and to be assessed for entry to high or low level residential care, and respite care. Although all ACATs completed a standard aged care client record (ACCR), the assessment tools used to populate this summary record varied widely. This paper describes the processes used to standardise ACAT assessment and more recent initiatives to link initial intake assessment with comprehensive assessment.; METHODS A literature review of Medline and PsychInfo databases was undertaken using terms associated with aged care assessment and assessment tools. Based on selection criteria which included the availability and quality of information on instrument psychometric properties a number of leading and currently used tools were identified and systematically reviewed. The comprehensive review included the instrument’s applicability, availability, administration requirements, psychometric properties (reliability, validity, responsiveness) and the availability of normative and clinical reference data. For each area of assessment
123
Qual Life Res (2012) 21:1–132 (e.g. cognition, function etc.) comparative scores for the instruments, based on these elements, were derived. This material was submitted to an Expert Clinical Reference Group and a standardised assessment toolkit was designed.; RESULTS A number of instruments were recommended by assessment domain for the Standardized Assessment Toolkit. Alternative tools were recommended for rural and remote Aboriginal and Torres Strait Islander people and for people from culturally and linguistically diverse backgrounds. The assessment processes and instrument recommendations are presented.; CONCLUSIONS Standardizing comprehensive assessment processes is seen as a first step to developing information systems that are based on actual assessment data rather than summary information. This is the start of a continuous electronic record and has the potential to lead to the refinement of assessment processes over time. In the context of recent aged care reforms in Australia there is also now the opportunity to align and link assessment processes from initial intake assessment through to comprehensive assessment.
(1073) Threats to the Internal Validity of Spinal Surgery Outcome Assessment: Recalibration Response Shift or Implicit Theories of Change Joel A. Finkelstein, MSc, MD, FRCS(C), Sunnybrook Health Sciences Center, Sunnybrook Center for Spinal Trauma, University of Toronto, Toronto, Canada; Brian R. Quaranto, B.S., DeltaQuest Foundation, Inc., Concord, MA, USA; Gary Naglie, MD, FRCP(C), FGSA, Department of Medicine and Kunin-Lunenfeld Applied Research Unit, Baycrest Geriatric Health Care System; Research Departments, Toronto Rehabilitation Institute and Toronto General Research Institute; University of Toronto, Toronto, Canada. Geriatric Medicine Research; Departments of Medicine and Health Policy, Management and Evaluation, University of Toronto, Toronto, Canada; Alex Kiss, Ph.D., Department of Research Design and Biostatistics; Clinical Epidemiology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Canada; Aileen M. Davis, Ph.D., Division of Health Care and Outcomes Research, Arthritis and Community Research and Evaluation Unit, Toronto Western Research Institute, University of Toronto, Toronto, Canada; Carolyn E. Schwartz, Sc.D., DeltaQuest Foundation, Inc., Concord, MA; Departments of Medicine and Orthopaedic Surgery, Tufts University Medical School, Boston, MA, USA. AIMS A recalibration response shift will cause the patient to think about a self-report measure’s response options differently after a health state change. Commonly assessed using the retrospectivepretest design (‘‘then-test’’), recent guidelines suggest adjusting thentest estimates for competing explanations. This prospective longitudinal study investigated recalibration response shift after adjusting for implicit theories of change in patients undergoing spinal surgery. METHODS The Oswestry Disability Index (ODI) and Short Form-36 (SF-36) were collected before surgery (n = 169), and at 6 weeks (n = 102), 3 (n = 106), and 6 months (n = 68) after spinal decompression surgery. Then-tests of the measures were also collected at all post-tests. Recalibration response shift was operationalized as the then-minus-pre difference score on the evaluative SF-36. Implicit theories of change were operationalized as the then-minus-pre difference score on the perception-based ODI. ‘‘Cure vs. No Effect’’ patient groups were compared were using the Minimally Important Difference (±15 points) as a cut-off on the Visual Analogue Scale (VAS) items for back and leg pain. Logistic regression analyses investigated whether recalibration response shift had an independent effect distinguishing patient groups, after adjusting for implicit theories of change. RESULTS The sample (mean age 52, 39 % female) was well-educated, and 1/3 were working. All then-minus-pre difference scores were non-zero and stable over time. In the adjusted models distinguishing Cured versus No Effect groups, then-minus-pre
Qual Life Res (2012) 21:1–132 ODI difference scores were significant in the majority of the adjusted models at all timepoints, but only one them-minus-pre SF-36 difference score—for physical functioning recalibration—was significant and only at 6-weeks post-surgery. This suggests that implicit theories of change bias the estimation of post-surgical outcomes, but that recalibration response shift biased only the estimation of physical functioning and only at 6 weeks post-surgery. CONCLUSIONS Although both recalibration response shift and implicit theories of change can both be sources of bias in patient-reported outcome assessment, our findings suggest that implicit theories of change are the greater threat to validity in this patient sample. Future research using the then-test should control for implicit theories of change to minimize misspecification of effects.
(1074) Computer-based quality of life monitoring with the EORTC QLQ-C30 in daily oncological routine—much information in little time Monika J. Sztankay, Innsbruck Medical University, Austria; Lisa M. Wintner, Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Austria; Johannes M. Giesinger, PhD, Innsbruck Medical University, Austria; August Zabernigg, Department of Internal Medicine, Kufstein County Hospital, Kufstein, Austria, Austria; Georg Kemmler, PhD, Innsbruck Medical University, Austria; Eva-Maria Gamper, Department of Psychiatry and Psychotherapy, Innsbruck Medical University; Anne S. Oberguggenberger, Innsbruck Medical University, Austria; Bernhard Holzner, PhD BE, Univ. Hospital, Innsbruck, Austria AIMS Although the assessment of patient-reported outcomes (PROs), such as quality of life (QOL), is widely applied in clinical and policy research, the implementation into clinical routine remains more challenging, partly due to common misgivings concerning the procedure. Frequent concerns of both, health care professionals and patients relate to the time needed to complete PRO questionnaires. To meet these concerns, our study aims at measuring completion time of routine clinical QOL assessment and identifying predictors for prolonged completion.; METHODS We performed a retrospective analysis on data of cancer patients treated at the Medical University Innsbruck, the Kufstein County Hospital and the Natters County Hospital, Austria. Data collection was done within ongoing computerized patient-reported outcome monitoring (ePROM) as part of clinical routine. The electronic form of the EORTC QLQ-C30 was presented to patients by means of the software CHES (Computerbased Health Evaluation System) at diagnosis and at clinically relevant time points.; RESULTS A total of 5,343 assessments derived from 1,968 patients were included into the analysis. On average, patients were 51.2 years old (SD 17.1), treated mostly for testicular (52 %) and brain cancer (12 %) with chemotherapy (59 %) or in after-care (34 %). Mean questionnaire completion time amounted to 5.1 min (SD 0.9 min). Type of current treatment (p [ 0.001) as well as age (p = 0.001) had a significant impact on completion time as did the current health condition as measured with EORTC QLQ-C30 (p \ 0.000). Data indicated a tendency of faster completion with increasing amount of assessments, while consistency of responses remained stable.; CONCLUSIONS Our results show that the EORTC QLQ-C30 is generally completed in a short amount of time, supporting the notion that ePROM is possible without disruption of the clinical time schedule. The waiting time can be used productively for collecting valuable additional information about the patient’s wellbeing to be integrated in the subsequent medical appointment. Further, implementation of routine computer-based monitoring might contribute to mollifying patients’ concerns about application of computerized assessment procedures and prepare the ground for home-monitoring QOL via Internet. Based on these results,
67 possibilities of enhancing acceptance and integration of ePROM into clinical routine are presented.
(1075) Assessment of the influence of explanatory and confounding variables on health-related quality of life—does response shift bias the results? Pranav Gandhi, PhD, South College, United States; L. Douglas Ried, University of South Florida; I-Chan Huang, PhD, University of Florida; Teresa Kauf, University of Florida; Carole Kimberlin, University of Florida AIMS To examine the influence of explanatory and confounding variables on health-related quality of life (HRQoL), after accounting for the response shift.; METHODS Hypertensive patients with coronary artery disease (CAD) completed the SF-36 surveys at both baseline and 1 year after random assignment to hypertension treatment (atenolol-led or verapamil-led) (n = 788). Changes in systolic and diastolic blood pressure, occurrence of myocardial infarction or stroke during follow-up, and self-reported prior history of depression were included as explanatory variables. Hypertension treatment strategy, baseline comorbid conditions, sex, age, race, living status, and educational level were included as confounding variables. The influence of explanatory and confounding variables on HRQoL, after accounting for response shift, were evaluated using structural equation modeling (SEM). The type of response shift (reconceptualization, recalibration, and reprioritization) linked to changes in various parameters of the SEM measurement model was defined based on Oort’s approach (Oort et al., Qual Life Res 2005).; RESULTS We found that older patients reported worse PF (direct effects -0.18 (p \ 0.001) at baseline and -0.20 (p \ 0.001) 1 year later) and RP (-0.14, p \ 0.001 at baseline and -0.12, p \ 0.001 1 year later) scores than younger patients, conditioning on the latent trait of general physical health. We also found that gender had a direct effect on PF at both time points (0.10 (p \ 0.001) at baseline and 0.13 (p \ 0.001) 1 year later) indicating that males reported better PF than females, conditioning on the latent trait of general physical health. Before controlling for response shift, patients’ PF was not statistically different between baseline and 1 year after treatment (p [ 0.05). However, after controlling for recalibration response shift, PF scores significantly improved after the treatment (p \ 0.01).; CONCLUSIONS The SF-36 PF scale was vulnerable to response shift. After accounting for response shift, a positive true change contribution indicated that HRQoL effects may have been underestimated, leading to false negative results. An understanding of how different patient groups may undergo a shift in their HRQoL perception due to health status change or allocation to treatment strategies should be investigated, especially in longitudinal studies.
(1076) Assessment of effects of response shift in palliative care patients using the structural equation modeling approach Yoshimi Suzukamo, PhD, Tohoku Univ, Dept of Physical Med & Rehab, Japan; Kikuko Miyazaki, PhD, Schl of Pblc Hlth—Kyoto University, Japan; Kojiro Shimozuma, Dept. of Biomedical Sciences, College of Life Sciences, Ritsumeikan University, Japan; Takeo Nakayama, MD PhD, Kyoto University School of Public Health, Japan AIMS To detect response shift in palliative care patients using the structural equation modeling (SEM) approach.; METHODS Secondary analysis from a previous study, which determined the minimally important difference for the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire Core 15
123
68 for Palliative (QLQ-C15-PAL), was performed to assess the effects of response shift on the physical functioning (PF) and the emotional functioning (EF) of QLQ-C15-PAL. One hundred eighty cancer patients undergoing palliative care responded to the questionnaire at baseline and after 2 weeks between September 2009 and December 2011. Response shift was identified using the Oort SEM approach.; RESULTS Within 2 weeks, the PF score decreased from 59.1 to 57.8 and the EF score increased from 65.2 to 68.4. The SEM approach identified three cases of response shift between the no response shift model (baseline model) and the model without invariance constraints (final model). Difference of intercept of overtime was observed, which indicated uniform recalibration for the PF1 items that questioned the participants whether they needed help with eating, dressing, bathing, or using the toilet (intercept: 1.839 at baseline and 2.060 at follow-up). Difference of factor loading appeared as a direct effect from EF to PF1 (0.182) at follow-up but not at baseline. In addition, covariance between PF and EF in the final model was significantly different from that in the baseline model (covariance: 0.181 at baseline and 0.311 at follow-up). The fit of final model was acceptable (RMSEA = 0.050, CFI = 0.984, 92 = 24.722 (17), p = 0.101).; CONCLUSIONS This study showed that palliative care patients with cancer experienced response shift over a 2-week period. The difference of intercept indicates recalibration and the difference of factor loadings and covariance indicates reprioritization and reconceptualization of response shift. These results will help medical experts gain a better understanding of the requirements of palliative care patients.
(1077) The responsiveness of disability in multiple sclerosis patients: comparison of a disease-specific (MusiQoL) and a generic (SF36) instruments Mohamed Boucekine, Self-perceived Health Assessment Research Unit, France; Karine Baumstarck, France; Julie Berbis, Department of Public Health, Medicine University, France; Oscar Ferna´ndez, Institute of Clinical Neurosciences,Ma´laga, Spain; Peter Flachenecker, Neurological Rehabilitation Center Quellenhof, Bad Wildbad, Germany; Helmut Butzkueven, MS Unit, Melbourne, Victoria, Australia; Jean Pelletier, Neurology, Timone University Hospital, Marseille, France; Sergio Stecchi, Multiple Sclerosis Unit, Department of Neuroscience AUSL, Bologna, Italy; Egemen Idiman, Department of Neurology, Dokuz Eylu¨l University, Izmir, Turkey; Pascal Auquier, PhD, Laboratoire de Sante Publique, France AIMS To assess the responsiveness of the Multiple Sclerosis International Quality of Life (MusiQoL) questionnaire to change in disability over 24 months, defined by change in the Expanded Disability Status Scale (EDSS) score, in multiple sclerosis (MS) patients. To compare the responsiveness of MusiQoL and SF36 questionnaires.; METHODS Patients with MS were enrolled into a multicenter, multiregional, longitudinal observational study. Quality of life (QoL) was assessed using both the MusiQoL and the 36-Item Short-Form (SF-36) instruments at baseline and every 6 months thereafter up to month 24; neurological assessments, including EDSS score, were performed at each evaluation. Patients were defined as worsened or non-worsened based on change in EDSS score from baseline to month 24.; RESULTS Of the 600 patients enrolled, 580 had evaluable baseline EDSS and baseline QoL, and at least one EDSS score and QoL score at a same subsequent assessment. The 24-month EDSS was available for 524 patients. In the 107 worsened individuals, the sentimental and sexual life and the relationships with health care system dimensions of MusiQoL showed responsiveness at month 24 whereas only ‘physical’ scores of SF-36 detected EDSS changes. MusiQoL scores detected EDSS changes earlier compared to SF-36 scores.; CONCLUSIONS
123
Qual Life Res (2012) 21:1–132 The MusiQoL index detected moderately disability changes in worsened MS patients in a 24-month observational study. Specific dimensions, such as the relationships with health care system and sentimental and sexual life, identified interesting EDSS change and detected these changes earlier than the SF-36 scores. Future studies should provide responsiveness with longer follow-up time evaluation.
(1078) Patient recovery after surgery for bladder cancer: a comparison of open versus robotic approaches John M. Hollingsworth, University of Michigan, Department of Urology, United States; Jeffrey S. Montgomery, University of Michigan; Scott Gilbert, University of Florida, Department of Urology; John T. Wei, MD, University of Michigan, United States; Rodney Dunn, MS, University of Michigan Dept of Urology; Brent Hollenbeck, MD, MS, University of Michigan AIMS One of the touted advantages of robotic surgery for bladder cancer is quicker patient recovery. To date, however, no study has compared health-related quality-of-life outcomes between the traditional open approach and robot-assisted laparoscopic radical cystectomy.; METHODS We conducted a prospective study of 87 consecutive patients, who underwent radical cystectomy at our institution between 2008 and 2012. We assessed early recovery using the Convalescence and Recovery Evaluation (CARE) questionnaire. CARE consists of 27 items, which are divided among 4 domains: activity, cognitive, gastrointestinal symptoms, and pain. Domain scores range from 0 to 100, with higher scores corresponding to a better health state; these scores can be combined to generate a composite score. We evaluated for differences by surgical approach in domain scores using the nonparametric Wilcoxon rank-sum test.; RESULTS Baseline CARE scores were comparable between the 2 groups (Fig. 1). Relative to the robotic surgery group, the cognitive domain demonstrated an increase among patients undergoing open surgery at 4 weeks (28.5 vs. 21.7, P = .06). The open surgery group also had increased pain scores at 4 weeks (29.1 vs. 20.0, P = .02). However, both of these differences abated by 6 weeks. No differences in activity or gastrointestinal scores were seen between the 2 groups at any point.; CONCLUSIONS Short-term recovery following open and robot-assisted surgery for bladder cancer is similar.
Fig. 1
Qual Life Res (2012) 21:1–132 (1079) Does change in health-related quality of life (HRQoL) score predict survival? Sensitivity analysis of a lung cancer RCT Divine Ewane Ediebah, MSc., EORTC HQ, Belgium; Corneel Coens, MSc, EORTC HQ, Belgium; Efstathios Zikos, MSc, MA, EORTC HQ, Belgium; Chantal Quinten, MSc, EORTC, Belgium; Jolie Ringash, MD, The Princess Margaret Hospital, Canada; Carolyn C. Gotay, PhD, UBC School Pop & Pub Health, Canada; Eva Elfriede Greimel, PhD, Medical University Graz, Austria; David Osoba, FRCPC, QOL Consulting, Canada; Madeleine T. King, PhD, Psycho-oncology Cooperative Research Group (PoCoG), Australia; Henning H. Flechtner, PhD, Otto-von-Guericke University (KKJP), Germany; Martin Taphoorn, PhD, Medical Centre Haaglanden, Neurology, The Hague, The Netherlands; J. Schmucker von Koch, PhD, University of Regensburg, Medical Ethics, Regensburg, Germany; Joachim Weis, PhD, University of Freiburg, Psychooncology, Freiburg, Germany;Charles S. Cleeland, PhD, The University of Texas MD Anderson Cancer Center, Houston, TX, United States; Bryce B. Reeve, PhD, University of North Carolina at Chapel Hill, Unites States; E.F. Smit, PhD, Free University Hospital, Dept. of Pulmonology, Amsterdam, the Netherlands; Andrew Bottomley, PhD, EORTC, Quality of Life, Brussels, Belgium on behalf of the Lung Cancer Cooperative Group, Quality of Life Department and Patient Reported Outcome and Behavioral Evidence (PROBE) AIMS Over 60 cancer clinical trials have shown that baseline health-related quality of life (HRQoL) scores are prognostic for patient survival. Few studies have investigated the added value of change in HRQoL scores. Analysis in change from baseline considering a 10 point difference clinically important has been shown to be associated with survival. Our aim was to investigate whether a smaller change in HRQoL score from baseline (6 point difference) was also associated with survival.; METHODS We analyzed data from an EORTC 3-arm randomized clinical trial (RCT) in advanced non-small-cell lung cancer (NSCLC) patients, comparing gemcitabine + cisplatin, paclitaxel + gemcitabine, versus standard arm paclitaxel + cisplatin. HRQoL was measured in 391 patients using the EORTC QLQ-C30 at baseline and after each chemotherapy cycle. The prognostic significance of sex, age and WHO performance status (PS) (0–1 vs. 2) and the 15 QLQ-C30 subscales was assessed with Cox proportional hazard models stratified for treatment (significance level 0.05). Changes in HRQoL scores from baseline to each cycle were categorized as ‘‘improved’’, ‘‘stable’’ and ‘‘worsened’’ using a threshold of 6 points difference. Due to expected attrition, the analysis was limited to changes from baseline up to cycle 3.; RESULTS There were 248 patients in cycle 1, 213 in cycle 2 and 196 in cycle 3. We performed analyses separately using data at cycle 1, 2 and 3. In all analyses, survival was predicted as follows: at cycle 1: physical functioning (hazard ratio [HR] 0.94, 95 % CI 0.89–0.99; p = 0.0067), pain (1.07, 1.02–1.11; p = 0.0003), age (0.98, 0.97–1.00, p = 0.0388) and WHO PS (1.68, 0.94–3.02; p = 0.0374); at cycle 2, pain (1.06, 1.02–1.11; p = 0.0014), age (0.98, 0.96–1.00; p = 0.0309) and sex (0.64, 0.42–0.96; p = 0.0111); and at cycle 3, role functioning (0.96, 0.92–0.99; p = 0.0061) and age (0.98, 0.96–1.00; p = 0.0035). However, change in HRQoL was only an independent predictor for improvement at cycle 1.; CONCLUSIONS Our findings suggest that a change of 6 points from baseline over time in patients’ HRQoL scores, as measured on subscales of the EORTC QLQ-C30, is similarly predictive of survival as a 10 points difference. Further work is needed to assess the sensitivity of these findings in other cancer sites.
69 (1080) The new instrument for comprehensive symptom assessment in patients with chronic myeloid leukemia Tatiana Nikitina, PhD, Multinational Ctr for QoL Research, Russia; Denis A. Fedorenko, PhD, Nat’l Pirogov Med Surgical Ctr, Russia; Kira A. Kurbatova, Multinational Center for Quality of Life Research, Russian Federation; Elena I. Usacheva, Hematological Center 1, Saint-Petersburg, Russia; Tatiana I. Ionova, Multinational Center for Quality of Life Research, Russian Federation; Gary I. Gorodokin, New Jersey Center for Quality of Life and Health Outcomes Research, NJ, USA; Andrey A. Novik, National Pirogov Medical Surgical Center, Moscow, Russia AIMS The value of patient-reported outcomes in treatment efficacy assessment of chronic myeloid leukemia (CML) is of great importance. Recently a new tool, Comprehensive Symptom Profile in Chronic Myeloid Leukemia (CSP Leuk-CML), has been developed to assess symptoms specific to CML patients. We aimed to test clinical applicability of CSP Leuk-CML.; METHODS 35 patients with chronic phase CML resistant or intolerant to imatinib were enrolled in the study (mean age–48.7 years old, range–22–79 years; male/ female–18/17). The patients filled out CSP Leuk-CML twice: before second-line therapy with dasatinib and 1 month after treatment start. Analysis of symptom severity in different patient groups was conducted using t test. To assess symptom severity changes, the effect size (ES) was measured. Pearson coefficient was calculated to examine the inter-item correlations.; RESULTS Practicability of CSP Leuk-CML was shown: patients needed 7–10 min to answer it; the proportion of missing values was 1.3 % for all questions; all items were easy for the patients to read and understand; the data produced by the tool were clear for interpretation by physicians. Several groups of symptoms, both disease and treatment specific, have been identified on the basis of correlation analysis (r = 0.7–0.93), which were clinically relevant and increased the practicality of the tool. Usefulness of CSP Leuk-CML to distinguish patients in terms of severity and number of disease and treatment specific symptoms was demonstrated. All the patients had at least 3 symptoms. 77 % of patients experienced at least one moderate-to-severe symptom; 37 % patients—more than 7 moderate-to-severe symptoms. The construct validity of CSP Leuk-CML was proven by ‘‘known-group’’ comparison: several CML-specific symptoms were more severe in patients with minor/no cytogenetic response than with major cytogenetic response (excessive sweat at rest and during mild physical activity, feeling of worry/anxiety, edema, headache; ES, 0.40–0.86). The ability to detect changes was demonstrated by comparing symptom severity before and 1 month after treatment start (ES, 0.36–0.46).; CONCLUSIONS The CSP Leuk-CML is a practical and sensitive tool to assess symptom profile and severity in CML patients. Comprehensive symptom monitoring using this tool is recommended to clearly determine treatment outcomes in this patient cohort. (1081) Understanding the relationships between health-related quality of life domains during recovery from joint replacement Aileen M. Davis, PhD, Toronto Western Research Institute and University of Toronto, Toronto, Canada, Canada; Selahadin Ibrahim, Institute for Work and Health; Sheilah Hogg-Johnson, Institute for Work and Health and University of Toronto, Toronto, Canada; Anthony V. Perruccio, University Health Network and University of Toronto, Toronto, Canada; Rose Wong, Toronto Western Research Institute, Canada; Dorcas Beaton, PhD, St. Michael’s Hospital, Canada; Monique A. Gignac, Toronto Western Research Institute and University of Toronto, Toronto, Canada; Elizabeth M. Badley, Toronto Western Research Institute, Canada
123
70 AIMS The purpose of this work was to understand how health-related quality of life (HRQOL) domains were related to each other over the course of recovery in the year following total hip or knee replacement (TJR) for osteoarthritis. Evaluating domains based on the International Classification of Functioning, Disability and Health framework which describes human functioning through impairment of body structure and function, activity limitation and participation restriction in the context of a person’s social and physical environment, this work investigated the temporal relationships among HRQOL domains. Hypotheses: (1) physical impairment (manifested by pain), activity limitations and participation restrictions would be associated within time and longitudinally; (2) across time, prior status of a domain would be associated with future status; (3) prior status of one domain would influence status of a second domain (e.g. prior activity limitations would influence current participation restrictions); and, (4) magnitude of the within time relationships of the domains would vary over time.; METHODS Participants (n = 931) with primary TJR completed standardized outcome measures (physical impairment (pain), activity limitations and participation restrictions) pre-surgery and five times in the year post-surgery. Domain relationships were evaluated cross-sectionally and longitudinally using path analysis adjusting for age, sex, body mass index, hip versus knee, low back pain and mood.; RESULTS All domain scores improved significantly over time. The findings (path coefficients) supported the hypotheses in that both cross-sectionally and longitudinally, physical impairment was associated with activity limitation and activity limitation was associated with participation restriction; prior status and change in a domain were associated with current status in another component. The magnitude of the path coefficients varied over time with stronger associations among domains to 3 months post-surgery than later in recovery with the exception of the association between impairment and participation restrictions which was of similar magnitude at all times.; CONCLUSIONS This work enhances understanding of the complexities of domain relationships in evaluating HRQOL over time, particularly the influence of the status in one domain on change of a second domain. Recovery time and the impact of one outcome on another outcome need to be considered in interpreting outcome.
Qual Life Res (2012) 21:1–132 AIMS The aim of this study was to list and describe practices in PRO primary psychometric validation studies, using a systematic literature review.; METHODS We conducted this review by following Centre for Review and Dissemination guidance for undertaking reviews in health care. The first stage consisted in search strategy and identification of articles by specifying keywords, search string in the PubMed database, inclusion and exclusion criteria. The second consisted in articles pre-selection by reading titles, then selection by reading abstracts. Finally, data were collected from articles by filling in a reading grid and a synthesis was done. To pre-select articles, one author (EA) reviewed the title of every records retrieved from the search and three experts (LM, JBH and VS) each read 1/3 of the titles independently from EA. Disagreements were resolved through a third reviewer. The same procedure was used with abstracts for articles selection and disagreements resolution. EA extracted data to describe the study, the scale and more specifically whether sample size requirement were estimated, whether missing values were managed, whether psychometric properties were evaluated and whether confirmatory factor analysis (CFA) were evaluated. The reading grid consisted of 127 variables. LM, JBH and VS completed it independently from EA on a random selection of articles.; RESULTS We identified 4,541 potentially relevant articles published from January 2009 to September 2011. After the pre-selection and selection steps (kappa agreement coefficients between 0.76 and 0.86), 422 articles were included. Data were extracted on 100 randomly select articles and 32 were excluded according to exclusion criteria. Initial results indicated that sample size requirement was defined in 12 % of articles. Construct validity was evaluated in 95 % of articles and among them, 75 % performed an exploratory factor analysis and 15 % a CFA. Internal consistency was measured in 91 % of articles.; CONCLUSIONS Preliminary results revealed a wide range of practices. It will lead to suggest recommendations for authors and keys to make a critical reading of published papers. A focus will be made on sample size requirement.
(1084) Patient-specific measures for goal setting. Research plan and first results (1082) The importance of the domain in self-reported quality of life of elderly Chileans Alfonso Urzua, PhD, Escuela de Psicologia-Universidad Catolica del Norte, Chile AIMS Analyze the effect of weighting the evaluation by the grade of importance that each participant assigns to the domains of quality of life; METHODS With a transversal design, was evaluated 398 elderly with the specific scale of quality of life WHOQOLD-OLD; RESULTS The worries about death would be linked to a worse evaluation of quality of life, meanwhile, be satisfied with the activities made and to be made, would be linked to a better perception of quality of life. It also found than having more or less years of live, being a man or woman, and the presence or absence of chronic disease influence the perception of this.; CONCLUSIONS Differences are found in the evaluations when weighted by the grade of importance, especially in presence of chronic disease.
(1083) Practices in PRO psychometric validation studies: a systematic literature review Emmanuelle Anthoine, University of Nantes, France; Leila Moret, University Hospital of Nantes; Ve´ronique Se´bille, University of Nantes, France; Jean-Benoit Hardouin, PhD, University of Nantes—EA4275, France
123
Anita Stevens, Zuyd University of applied science; Anna Beurskens, PhD, Zuyd University of applied sciences, Netherlands; Albe`re Ko¨ke, Adelante Centre of Expertise in Rehabilitation; Trudy Van Der Weijden, Maastricht University AIMS The aim of this research project is to improve the use of patient-specific measures in the goal setting process for patients with complex care demands, in community practices for physical-, occupational and speech and language therapy. The following research questions are leading in this study: (1) Which patient-specific measures can be used in the goal setting process and what is known about their feasibility and measurement properties. (2) How do professionals use patient-specific measures in the goal setting process in daily practice? (3). How can patient-specific measures be shaped and adapted to a better use and how can these improvements be implemented in daily practice?; METHODS Literature searches were carried out to reveal patient-specific measures that can be used in the goal setting process (from goal negotiation, goal setting until appraisal and feedback) and to review their feasibility and measurement properties. In the second phase of the study several qualitative studies will reveal the use of the instruments in the process of goal setting in daily practice in physical and occupational therapy practices in the community. Up on these results an expert group of patients, professionals and researchers, will develop propositions for improvement and adaption of the goal setting measures and procedures. Next, implementation strategies will be developed, tested and evaluated. RESULTS In total eleven patient-specific instruments were
Qual Life Res (2012) 21:1–132 selected and assessed. Most frequently researched were the Canadian Occupational Performance Measure (COPM); Patient Specific Functional Scale (PSFS) and the Goal Attainment Scaling (GAS). In most of the eleven instruments the process of goal negotiation and setting was not well described. The scoring methods of some measures are very much similar and there is some evidence for reproducibility, validity and responsiveness.; CONCLUSIONS The review showed the clinimetric value of the scoring methods, but the feasibility and implementation of the instruments in the whole process of goal setting in daily practice needs further research.
(1085) Development and validation of the system of quality of life instruments for cancer patients (QLICP) in China Chonghua Wan, PhD, Guangdong Medical College, Peoples Republic of China; Zheng Yang, Guangdong Medical College, Peoples Republic of China; Jiahong Luo, Kunming Medical University; Qiong Meng, Kunming Medical University; Xiaoqing Zhang, Kunming Medical University; Xueliang Tang, Yunnan Tumor Hospital; Gaofeng Li, Yunnan Tumor Hospital; Yingli Cun, Yunnan Tumor Hospital AIMS To develop a Chinese QOL measure system called QLICP (Quality of Life Instruments for Cancer patients), expecting to include more than 20 cancer-specific scales.; METHODS Based on WHO’s definition of QOL and programmed decision procedures including nominal group, focus group discussion, in-depth interview to clinicians and patients, pilot test and field test, the QLICP was developed by the modular approach, with general module (QLICP-GM) being used to all types of cancers and specific modules developed only for specific cancers. Psychometric properties of the scale were evaluated by the field tests of longitudinal design of three time measurements with some indicators of Cronbach alpha, Pearson r, SRM (standardized response mean) and statistical methods of correlational analysis, cluster analysis, multi-trait scaling analysis, t tests, factor analysis/ structural equation modeling etc.; RESULTS The 32 items QLICPGM (4 domains and 9 facets) and 12 specific modules with the numbers of items ranging from 8 to 20 have been developed. All QLICP instruments have a very clear hierarchical structure (items, facets, domains, overall) and can present mean scores not only at the domain and the overall level but also at the facet levels. The six scales of the QLICP have been validated by the data from 692 patients (85 lung cancer for QLICP-LU, 186 breast cancer for QLICP-BR,133 head & neck cancer for QLICP-HN,86 stomach cancer for QLICP-ST, 110 colorectal cancer for QLICP-CR and 92 cervical cancer for QLICPCE) and the others are validating. Correlational and factor analysis/ structural equation model indicated good construct validity for all six available instruments. Good criterion-related validity coefficients were confirmed given relevant FACTs/EORTC QLQs as the criterions. Test–retest reliability coefficients for most domains of most instruments were higher than 0.80, and the internal consistency alpha were higher than 0.70. Most domains and overall scores of most instruments had a statistically significant change after treatments with moderate or higher effect size, SRM.; CONCLUSIONS The all scales of the system of QLICP have good validity, reliability and responsiveness and can be used as QOL instruments for cancer patients in China.
(1086) Health-related quality of life, marital satisfaction, and social support with a special focus on gender differences in medical staff in Tehran—Iran Arian Rostami, Umea˚ University; Mehdi Ghazinour, Umea University—Department of social work; Lennart Nygren, Umea˚ University—Department of Social Work, Sweden; Jo¨rg Richter, Centre for Child and Adolescent Mental Health, Oslo, Norway
71 AIMS To investigate the relationships between quality of life, marital satisfaction, and social support with a special focus on gender differences in medical staff in Tehran—Iran; METHODS This was a crosssectional study in medical staff in Tehran. Data collection was performed in 653 medical staff using sociodemographic questions, the SF-36 questionnaire, the ENRICH Marital Satisfaction Inventory, and the Social Support Questionnaire.; RESULTS The results revealed that men reported significantly higher quality of life and marital satisfaction than women, but there was no gender difference in social support. The quality of life and marital satisfaction domains were found to be differentially correlated among male and female participants. All domains of marital satisfaction were found to be significantly associated with the ‘‘mental health’’ domain among females but with only four marital satisfaction domains in the male. Also, social support was associated with almost all quality of life domains in both sexes. The multiple regression analyses indicated that all the study variables combined with sociodemographic factors can explain between 12 and 28 % of the variance in quality of life domains.; CONCLUSIONS Therefore, focusing on factors such as marital satisfaction and social support could be an effective approach to improve quality of life in medical staff as a group with a stressful job.
(1087) The association between oral health-related quality of life and self-rated presenteeism Hiroya Takiuchi; Aya Kimura-Ono, Okayama Univ. Grad. Sch. Med., Dent. and Pharmaceu. Sci., Okayama, Japan.; Hikaru Arakawa, Okayama Univ. Grad. Sch. Med., Dent. and Pharmaceu. Sci., Okayama, Japan., Japan; Takuya Mino, Okayama Univ. Grad. Sch. Med., Dent. and Pharmaceu. Sci., Okayama, Japan.; Asuka Masaki, Okayama University, Japan; Emilio Satoshi Hara, Okayama Univ. Grad. Sch. Med., Dent. and Pharmaceu. Sci., Okayama, Japan.; Manabu Kanyama, Okayama Univ. Grad. Sch. Med., Dent. and Pharmaceu. Sci., Okayama, Japan.; Yoshizo Matsuka, Okayama University, Japan; Takuo Kuboki, Okayama Univ. Grad. Sch. Med., Dent. and Pharmaceu. Sci., Okayama, Japan AIMS Presenteeism has been defined as a decrease in work productivity due to health problems. Presenteeism has been associated with low health-related quality of life (HRQOL); however, no report has showed its association with oral-HRQOL (OHRQOL). Therefore, the purpose of this cross-sectional study was to investigate the association between presenteeism and OHRQOL in the context of the generalHRQOL.; METHODS Data were collected from a total of 89 employees (29 males, 60 females; mean age 46.6 ± 9.4 years) of a health food company in Japan in June, 2011. Presenteeism was evaluated by a Japanese version of the Stanford presenteeism scale (J-SPS). OHRQOL were measured by the general-oral-healthassessment-index (GOHAI). General-HRQOL was assessed by the SF8 Health Survey (SF-8), which summarizes two scores: the physical (PCS) and the mental component (MCS). Intra-oral clinical examination was performed to evaluate the number of remaining teeth and the periodontal condition assessed by the community periodontal index (CPI). Health status (systemic disease, occupational stress level, smoking and fitness habits) was assessed by medical examination and questionnaire survey. Depression was evaluated by the self-rating questionnaire for depression (SRQ-D). A cluster analysis and a multiple linear regression analysis with stepwise procedure were used to analyze the association between OHRQOL and presenteeism, taking in consideration the general-HRQOL factors as confounding variables.; RESULTS Complete data were available from 74 employees (27 males, 47 females; 46.7 ± 9.8 years). The remaining 15 subjects were excluded due to incomplete data. Reduction in work productivity was reported to be in a range of 0–90 % (average 31.2 ± 19.7 %). The average GOHAI score was 52.0 ± 8.8. The
123
72 average SF-8 scores were 48.3 ± 6.0 for PCS and 44.4 ± 7.3 for MCS. Twenty-five employees (34 %) had systemic diseases, and 13 employees (18 %) were diagnosed as having depression. The average number of remaining teeth was 26.2 ± 2.7. Thirty-two employees (43 %) were diagnosed as having periodontal disease. The multivariate analysis adjusted for age, SF-8 PCS score, depression, number of remaining teeth, periodontal disease and fitness habit, showed that OHRQOL was significantly associated with presenteeism (p = 0.04, b = 0.24).; CONCLUSIONS The results of this study indicated that low OHRQOL was associated with presenteeism.
(1089) QOL and the disability among the elderly in Japan Miyako Tazaki, PhD, Toho University, Japan; Tetsuo Yamaguchi, Toho University; Mitsutoshi Yatsunami, Tokyo University of Science, Japan; Yoshibumi Nakane, Nagasaki University AIMS To identify how QOL of the elderly is influenced by the degree of their disability measured by the Japanese version of the WHO Disability Assessment Schedule (WHODAS 2.0), which was created to evaluate the levels of disabilities based on the International Classification of Functioning (ICF).; METHODS Two hundred seventy-one elderly patients (males, 126; females, 135) with an average age of 72.5 years were provided WHOQOL26 and WHODAS 2.0. Among them, 18 required social welfare care.; RESULTS The collection rate was 92 %. The average WHOQOL26 and WHODAS 2.0 scores were 3.63 ± 0.54 and 1.25 ± 0.46, respectively. The results indicate that there is a significant (p \ 0.01) but negative (r = -0.56) correlation between WHOQOL26 and WHODAS 2.0 scores, and a significant correlation between WHODAS 2.0 and the degree of disability(r = 0.69, P \ 0.01). The Q12 (financial resources) and Q13 (information acquisition) scores of WHOQOL 26 were found to be related to almost all items of other QOL questions. Moreover, Q11 (body image acceptance) was found to be related to the social participation domain of WHODAS 2.0. The scores of Cronbach’s alph of six domain of WHODAS 2.0 were more than 0.8.; CONCLUSIONS Thus, the overall QOL of the elderly was influenced by the degree of their disability, ability of information acquisition, and financial resources; social participation was determined by body images. In addition, the Japanese version of WHODAS 2.0 was proved a good outcome measure of disability with high reliability and validity.
(1090) Necessary and sufficient contributors to participation post stroke Nancy E. Mayo, PhD, McGill University, Royal Victoria Hospital, Canada; Lois Finch, MSc, Canada; Susan C. Scott, MSc, McGill University, Canada; Sydney Miller, Concordia University AIMS Participation, a construct within the disability/functioning framework, is evaluated on a person’s involvement in life situations including family, community, work, social and civic life. In the context of recovering from a major health event, participation is a treatment goal and is known to correlate with quality of life (QOL). The purpose of this study was to track the dynamics of participation over the first year post-stroke in relationship to two important stroke outcomes considered necessary (but not sufficient) for participation: positive mood and walking capacity.; METHODS An inception cohort of 120 persons with stroke was followed with assessments at 1 week post-stroke and every 3 months subsequently. Participation and mood were self-reported, walking capacity was measured distance walked in 2 min (2MWT). Group-based trajectory analysis (GBTA), a form of latent class analysis, was used to identify distinctive groups of individuals with similar trajectories. Dual trajectories were used to estimate concordance between participation
123
Qual Life Res (2012) 21:1–132 trajectory and trajectories for each of the two necessary constructs.; RESULTS Four trajectories of participation were identified, 2 with positive slope and 2 flat. Among the persons assigned to the highest (and increasing) trajectory of participation, 78.5 % were also assigned to the best trajectory of generic mood (high and improving); 91 % had the best trajectory (increasingly positive) on a stroke-specific measure of mood. However, no one was from the depressed mood trajectories. In contrast, only 15 % were in the best trajectory of the 2MWT but no one had the worst 2MWT trajectory. The poorest participation group was 35 % concordant with the poorest generic mood group, 54 % concordant with the poorest stroke-specific group, and 95 % concordant with the poorest 2MWT group.; CONCLUSIONS These analyses suggest that non-depressed mood and walking capacity are necessary for participation but not sufficient. Walking capacity would be considered formative for participation, however, even with these longitudinal comparisons, it is not possible to identify whether positive mood is formative or reflective of participation. Input directly from stroke survivors would be needed to further explore this relationship which may differ from person to person. Given the close association between participation and QOL post-stroke, these relationships would informative for QOL.
(1092) QOL and pessimism in alcohol complicated and alcohol free couples Selwyn Stanley, PhD, University of Plymouth, United Kingdom AIMS Alcohol misuse is a major cause of morbidity and mortality and is globally an important health care burden. While there are several studies which have studied the alcoholic per se or his spouse or cohabiting partner, there are hardly any studies which have investigated both partners within an alcoholic relationship. The uniqueness of this study is that it seeks to do just this. It compares both spouses in a marital relationship involving an alcoholic husband (alcohol complicated couple; n = 30) with both spouses in a marital relationship where neither partner is an alcoholic (alcohol free relationship; n = 30) in terms of their quality of life and the extent of pessimism.; METHODS A correlational ex-post facto design was used to carry out this comparative analysis. 30 couples were identified from a de-addiction center in India where the husband was enlisted for in patient alcohol treatment. An equal number of couples were included as controls through the study group respondents after administering the AUDIT and including those below a score of seven (indicating nonalcoholic status). The Beck’s Hopelessness Scale and the WHO-QOL Bref were administered to all respondents (N = 120).; RESULTS Analysis of variance indicates that on both these domains, both spouses in the alcohol complicated relationship had poorer quality of life and more pessimism than their counterparts in alcohol free relationships. Further, in the alcohol complicated relationships, the wives manifested poorer QOL and greater pessimism when compared to their alcoholic spouse. For both the spouses in the alcohol complicated relationship, no significant correlations were obtained between the domains of the Hopelessness Scale and Quality of Life.; CONCLUSIONS The findings indicate the need for therapeutic intervention for both partners in alcohol complicated relationships, even if only one of them has alcohol misuse issues.
(1093) Psychometric studies of the European Portuguese version of the quality of life index EUROHIS-QOL-8 in HIV-infected patients Marco Pereira, Universidade de Coimbra, Portugal; Maria Cristina Canavarro, Universidade de Coimbra, Portugal
Qual Life Res (2012) 21:1–132
73
AIMS In the 90’s, the World Health Organization (WHO) has developed, cross-culturally, two generic instruments to assess the quality of life (QoL) (WHOQOL-100 and WHOQOL-Bref) and later has developed specific modules for HIV infection (WHOQOL-HIV and WHOQOL-HIV-Bref). In 2003, the EUROHIS-QOL 8-index was developed as an adaptation of the generic measures, because studies of large scale that include multiple indicators, or any situation where it is necessary an indicator of QoL of rapid assessment, sometimes require brief indices. The purpose of the current study was to test the psychometric properties of the EUROHIS-QOL 8-index in a sample of HIV-infected patients.; METHODS A cross-sectional study was conducted in the main services/departments of infectious diseases of Portuguese Hospitals. The sample consisted of 1,196 HIV-infected patients. Patients also completed the Portuguese versions of the Brief Symptom Inventory (BSI) and of the Beck Depression Inventory (BDI).; RESULTS The EUROHIS-QOL 8-index showed good internal consistency (Cronbach’s alpha = .85). Confirmatory factor analysis of the index items indicated acceptable fit between observed data and the one-dimensional model [CFI = .96; RMSEA = .09 (CI .08–.10)]. The correlations between the EUROHIS-QOL-8 index and the six domains of the WHOQOL-HIV-Bref ranged from 0.46 (Spirituality) to 0.81 (General facet and Environment), and the correlations with the measures of general psychopathology and depression were -0.66 (p \ .001) and -0.61 (p \ .001), respectively. The EUROHIS-QOL-8 index was also able to discriminate relatively well between CD4 T cell count subgroups, F(2, 1064) = 29.16, p \ .001.; CONCLUSIONS The psychometric properties of the EUROHIS-QOL 8-index validate its use in the HIV population of our country, particularly in large studies that require the assessment of multiple health indicators.
The mean absolute difference in overall scores constituted 11 % of the possible range of scores. However, the significance of this difference is difficult to interpret. The ICC for overall scores was 0.84 indicating excellent agreement between mother and child pairs, and the ICCs for the items indicated moderate to substantial agreement only.; CONCLUSIONS Although mothers may be used as proxies for their children in some circumstances and for some purposes, the views of both should be obtained in order to fully represent child OHRQoL.
Friday Poster Session
(2003) Impairment of quality of life in parents of children with autism spectrum disorder
(2001) Agreement between mothers and children aged 5–6 years in rating child oral health-related quality of life Jenny Abanto, Department of Pediatric Dentistry, University of Sa˜o Paulo, Brazil; Georgios Tsakos, PhD, University of College London, United Kingdom; Saul Martins Paiva, Federal University of Minas Gerais; Daniela Pro´cida Raggio, University of Sa˜o Paulo; Marcelo Bo¨necker, Department of Pediatric Dentistry, University of Sa˜o Paulo AIMS To assess the agreement between mothers and children concerning the child’s oral health-related quality of life (OHRQoL).; METHODS A total of 298 pairs of mothers and children aged 5–6 years with dental caries and traumatic dental injuries completed the recently validated Scale of Oral Health Outcomes for 5-year-old Children (SOHO-5). The Brazilian SOHO-5 (validation considered children of both 5 and 6 years old) is composed of 14 items, comprising 7 items in each version (child-self report and parental), with 6 of them being common in terms of content. Agreement between overall and items scores derived from the versions were assessed in comparison and in correlation analyses. The former used mean directional differences between mothers and children to assess bias and mean absolute differences to assess agreement at the group level. The latter used intraclass correlation coefficients (ICCs) to assess agreement at the level of individual mother–child pairs.; RESULTS At the group level, agreement between mothers and children was good. Children reported an overall score worse than their mothers, however the mean directional difference of 0.13 was not statistically significant. In spite of this, they were statistically significant for the items avoid smiling (due to pain and appearance), suggesting that mothers reports in this items were systematically different. In general, there was small evidence of bias in mothers reports compared to those of their children.
(2002) The importance of domain in the evaluation of quality of life in childrens Alfonso Urzua, PhD, Escuela de Psicologia-Universidad Catolica del Norte, Chile AIMS Explore the impact that it might have on an assessment, when this is based on the importance that the children give to each one of the domains that compose their QoL.; METHODS Was evaluated 600 children from educational schools, average age of 12 years old, distributed equally by gender, using a specific questionnaire KIDSCREEN-27. Each child weighting the importance of each domain in his/her life.; RESULTS Results show differences in the perception of the QoL between genders, types of schools, and differences in age. These results are maintained considering the raw scores as decided by importance. There are minor differences between the domain best evaluated and worst evaluated. When ranking by importance, the most evident aspects are age and type of school.; CONCLUSIONS The results contribute to the understanding of the cognitive processes involved in the QoL evaluation.
Virginie Villes, Public Health Laboratory, Marseilles, France; Anne-Marie Bartolini, Resource autism center, marseilles; Clarisse Chatel, Resource autism center, marseilles; Franc¸ois Poinso, Resource autism center, marseilles AIMS Little data are available on the quality of life (QOL) in parents of children with developmental diseases. The level of parental stress has been found to be related to the level of severity of the children’s diagnoses. Aims of the present study were: (1) to evaluate QOL in parents of children affected by autism spectrum disorder; (2) to search the type of child behavior problems that have the greatest impact on their parents; (3) to compare the level of impairment in QOL of mothers and fathers; (4) to compare QOL with French norms.; METHODS The sample consists of 75 children affected by autism spectrum disorder and their 135 parents. The degree of autistic disorder was assessed by the scale CARS. Evaluation of the severity of the various autistic features was conducted with ADOS. Each parent filled three questionnaires on their self-representations of the QOL (SF 36, PGWBI, PAR-EN-QoL).; RESULTS There is a very good correlation between the responses of fathers and mothers. The impact of light autism disorder appears to be higher on the QOL of mothers than on fathers. For severe autism, QOL of mothers and fathers has the same degree of degradation. Studies show a good correlation between the QOL of parents and severity of autistic disorder, socialization of children and the narrow interests of children. Parental physical health is associated with intellectual deficit of children and children’s motor skills development. There is good correlation between child behavior disorders and parental mental health, especially for depressed mood of parents, parents’ self-control and parents’ vitality. Important differences in QOL were observed with French norms.; CONCLUSIONS Results show a significant decrease
123
74 in QOL of parents when the severity of autistic disorder increases. Stereotyped behavior, social relationships and level of communication of AUT children have the greatest repercussions for QOL of parents. Finally intellectual impairment and motor retardation affect parental physical health.
(2004) Structural relations between socio-demographic factors, parenting stress, parental perceived quality of life and young children’s weight outcomes in Hong Kong Christine MS Chan, PhD, Institute of Education, HK, Hong Kong; John H.M. Lam, The Hong Kong Institute of Education AIMS The aims of this study were to explore the interrelationship between socio-demographic factors, parenting stress, parental perceived children’s quality of life and their weight outcomes.; METHODS 336 Chinese parents of young children aged 2–7 years from 27 kindergartens of different socio-economic backgrounds in Hong Kong were recruited for a cross sectional study. The Paediatric Quality of Life Inventory (PedsQL, 4.0) and the Parenting Stress Index (PSI-SF) were employed. One-way ANOVA, Pearson correlation, structural equation model (SEM) and path models were used for data analysis.; RESULTS The children’s body mass index (BMI) ranged from 11.3 to 28.0 kg/m2, of which four types of weight statuses had been identified by the IOTF. The parental reported quality of life (RQoL) was different among the four weight statuses (F[3, 332] = 4.41, p \ 0.01), while no significant differences on the parenting stress (F[3, 332] = 0.87, n.s.). The normal weight group had the highest RQoL and the obese group has the lowest. There were no linear relationships between weight status and RQoL (r = -0.08, n.s.), Psychosocial Health (r = -0.05, n.s.) and parenting stress (r = 0.04, n.s.). On the other hand, parenting stress was slightly correlated with the RQoL (r = -0.22, p \ 0.001) and Psychosocial Health (r = -0.27, p \ 0.001). The goodness-of-fit indices of the SEM were good (v2 (41) = 74.481, p \ 0.01; RMSEA = 0.049, (90 % CI = (0.031–0.067); SRMR = 0.054; CFI = 0.960; NNFI = 0.947). When controlling the demographic background the parenting stress negatively predicted the RQoL (Standardized b-coefficient = -0.289, p \ 0.001); from the first path model, only the subscale of Difficult Child negatively predicted Emotional (Standardized b-coefficient = -0.293, p \ 0.001) and School Functioning (Standardized b-coefficient = -0.149, p \ 0.05); from the second path model, the parenting stress negatively predicted the Psychosocial Health (Standardized b-coefficient = -0.276, p \ 0.001); from the final path model, the parenting stress negatively predicted the RQoL (Standardized b-coefficient = -0.220, p \ 0.001).; CONCLUSIONS These findings could suggest directions for childhood obesity interventions.
(2005) Relationship between protective health behaviors and health-related quality of life in German preschool children Angela Gosch, Dr. Phil, Munich University of Applied Sciences, Germany; Rainer Pankau, Finkelstein Pediatric Hospital, Heidekreis, Germany AIMS The study aimed to explore the association between the presence of several protective health behaviors and physical and mental wellbeing among German preschool children.; METHODS In this representative survey in the course of school enrolment a total of 1,116 children, aged between 5 and 6 years, were assessed physically and psychologically. Their parents were asked to fill in questionnaires to assess seven daily health behaviors, their HrQoL (KIDSCREEN parent-proxy questionnaire, Ravens-Sieberer et al. 2010), and their
123
Qual Life Res (2012) 21:1–132 strengths and difficulties (SDQ a brief behavioral screening questionnaire, Goodman, 1999).; RESULTS In more than 90 % of the families children carry out daily health behaviors (to brush one’s teeth twice a day, more than 1 h daily exercises, etc.). One exception is the daily intake of five portions of fruit and vegetables (take five), only for 10, 5 % of the children display such behavior. A significant positive correlation between the total score of health behaviors and the child’s HrQoL is found. In addition, significant correlations between HrQoL and other determinants (SES, migration status, SDQ) are shown.; CONCLUSIONS Results indicate a protective role of health behaviors in (physical and) mental wellbeing (HrQoL). The importance of the results and the need for more health education interventions will be addressed.
(2006) Dimensions of quality of life of Brazilian children and adolescents in hemodialysis Isabella S. Abreu, Universidade de Sa˜o Paulo, Brazil; Danielle Maria S S dos Santos, MD, University of Sao Paulo, Brazil; Keila C. Deon, MD, University of Sa˜o Paulo, Brazil; Regina Aparecida Garcia de Lima, University of Sa˜o Paulo, Brazil; Maria Fernanda Cabral Kourrouski, University of Sa˜o Paulo, Brazil; Lucila Castanheira Nascimento, University of Sa˜o Paulo, Brazil; Roberta A. Reis, PhD, Universidade Federal do Rio Grande do Sul, Brazil; Claudia Benedicta dos Santos, MISP-EERP-USP, Germany AIMS Describe the dimensions of quality of life of Brazilian children and adolescents with Chronic Renal Failure on hemodialysis.; METHODS Methodological study, developed according to the method used by the DISABKIDS Group. Attended children and adolescents between the ages of 8 and 18 years old and their parents or caregivers. Were carried out 10 focus groups, between August 2011 to February 2012, at two hospitals of Brazil. The resulting depositions of focus group were recorded and transcribed in their entirety and arranged according to the physical, mental and social domains, using the card sorting method.; RESULTS The focus groups were conducted in accordance with the instrument provided by the DISABKIDS group, translated to Portuguese of Brazil language and adapted to the present study. The empirical material has been organized according to defining dimensions of quality of life used by DISABKIDS group, being identified a total of 51 reports for the limitation dimension, 48 for the independence dimension, 21 to social exclusion dimension, 16 for emotional dimension, 11 for social inclusion dimension and 07 for treatment dimension.; CONCLUSIONS The focal group proved effective for the collection of data, promoting free expression, thoughts, positions and discussions from the perspective of participants, which favored the understanding of their experiences, perceptions and ways of coping and disease treatment and how this relates to the quality of life. These results complemented with expert reports and literature will result in producing items that will make up a specific instrument DISABKIDS for Brazilian children and adolescents with Chronic Renal Failure on hemodialysis. Such an instrument will support a clinical practice of health professionals in the care of these patients, providing subsidies to help them cope with their condition to minimise the negative impact arising from this process, causing it to adapt to changes and to create tools built in confronting the limitations imposed by the chronic condition.
(2007) Patient reported outcomes in pediatric clinical trials: a systematic literature review Rachel E. Teneralli, MS, RD, Children’s Hospital of Philadelphia, United States; Katherine Bevans, PhD, Children’s Hospital
Qual Life Res (2012) 21:1–132 of Philadelphia; Christopher B. Forrest, MD, PhD, University of Pennsylvania, School of Medicine, United States AIMS Over the past two decades, a significant body of research has accrued demonstrating children can provide precise and valid assessments of their health and quality of life. As this evidence base has grown, the number of patient reported outcome (PRO) measures has proliferated. The degree to which this new ‘‘technology’’ has been adopted is unclear. The purpose of this study was to assess the uptake of PROs in pediatric clinical trials over the past several decades.; METHODS A systematic literature search was conducted to identify pediatric clinical trials that included a PRO measure. A clinical trial was defined using NIH criteria: a prospective research study involving human subjects designed to answer specific questions about biomedical or behavioral interventions. A PRO was considered, according to ISOQOL criteria: a standardized method of measuring the patient perspective and any outcome based on data provided by patients or parent proxies. A comprehensive search strategy was conducted in Medline and CINAHL to identify clinical trials conducted through 2011 that included PROs for children (\18 years). Search terms reflected PRO measurement, clinical trials, and pediatric populations, and limited articles to those in English.; RESULTS The first year a PRO instrument was used in a pediatric clinical trial was 1991. During 2011, 66 trials included a PRO measure. Over the past 20 year period, a total of 48,267 pediatric clinical trials were conducted of which 536 (1.1 %) included a PRO endpoint. An increasing trend in PRO adoption was observed starting in 1998 where 0.5 % of trials included a PRO, increasing to 2.1 % by 2011. PROs were most commonly used in educational and psychological interventions followed by medication trials. PROs were least likely to be used in device or surgical trials.; CONCLUSIONS Although this review revealed that the use of PROs in pediatric clinical trials has significantly increased over the past 20 years, they remain underutilized in pediatric clinical trials. These findings emphasize the substantial amount of work that is needed to enhance their adoption.
(2008) Does the severity of childhood asthma affect caregiver’s quality of life? Karen H. Sousa, RN PhD, University of Colorado Denver; College of Nursing; Judy A. Harris, Phoenix Children’s Hospital; Susanne W. Cook, Retired AIMS Understanding the quality of life of caregivers is important to understand the quality of life of children living with asthma. Little is known about the impact of childhood asthma on the caregiver’s quality of life. The goal of this analysis was to evaluate the relationship between caregivers’ quality of life and the child’s asthma severity.; METHODS Phoenix Children’s Hospital, Phoenix Arizona USA, has been operating a mobile asthma clinic, Phoenix Breathmobile, for medically underserved, inner-city school age children. The program provides asthma case detection, education, and interventional treatment programs. The population consists of children 5–18 years old (m = 9.48 years) with asthma symptoms that have been identified through a case-detection process. Patients’ (N = 992) charts were abstracted for all the children seen by the Breathmobile since its inception. 54.4 % were boys and 64 % were Hispanic. The PACQLQ was administered to the primary caregiver at the first visit and a severity ranking based on the provider assessment using a threepoint scale, mild to severe, was used. Examples of questions on the PACQLQ are: Did you feel helpless or frightened when your child experienced cough, wheeze, or breathlessness?; Did your family need to change plans because of your child’s asthma? Analysis was conducted using SPSS. Pearson correlation coefficients were used to assess the relationship between the PACQLQ and child’s severity scores.; RESULTS 19.9 % of the caregivers once in awhile felt
75 helpless or frightened when their child experienced cough, wheeze, or breathlessness. 13.2 % of the caregivers had sleepless nights because of child’s asthma. Additional results to the PACQLQ will be discussed. There were statically significant correlation with the child’s severity score and the over all PACQLQ and with each subscale of the PACQLQ.; CONCLUSIONS This study included families of innercity children living with asthma. This analysis suggests that it is important to measure the quality of life for caretakers of children with asthma. It is important to understand the burden experienced by both the caregiver and the child. Further implications will be discussed.
(2009) Instruments for measure of the quality of life of children and adolescents with atopic dermatitis: a literature review Keila C. Deon, MD, University of Sa˜o Paulo, Brazil; Danielle Maria S S dos Santos, MD, University of Sao Paulo, Brazil; Isabella S. Abreu, Universidade de Sa˜o Paulo, Brazil; Monika Bullinger, PhD, Inst for Medical Psychology, Hamburg Univ, Germany; Claudia Benedicta dos Santos, MISP-EERP-USP, Germany AIMS A range of health-related quality of life instruments have been developed or adapted for use in the Atopic Dermatitis, but no headto-head comparisons of these instruments, their structure, domains and application have been reported. The aim of this study was to review available instruments for assessing children and adolescents’ Quality of Life specifically in Atopic Dermatitis.; METHODS The search was conducted in four electronic databases, Medline, PsycInfo, Embase and the Cochrane Library, including the site Patient-Reported Outcome and Quality of Life Instruments Database, using the descriptors: Atopic Dermatitis, Health-related Quality of Life, Instruments, Child, Adolescent, Parents and Caregiver. The language considered was English, and it was searched publications between 1990 and 2012.; RESULTS Five instruments were found, one for children; one for children and adolescents; two for parents and caregivers; one for both, children and parents. The instruments Infants’ Dermatitis Quality of Life Index, proxy-version, for children aged zero until 4 years, with 10 items and 2 domains; DISABKIDS—Atopic Dermatitis Module, self and proxy-versions, for children and adolescents aged eight until 18 years, which has 12 items and two domains; Childhood Atopic Dermatitis Impact Scale, with 45 items allocated into five domains, the items altogether measures the Health related Quality of Life of both parents or caregivers and their child, aged until 6 years, by a proxy-version; Parents’ Index of Quality of Life in Atopic Dermatitis, to access the parents or caregivers’ Health related Quality of Life, which has 28 items in a selfversion; Quality of Life in Primary Caregivers of Children with Atopic Dermatitis, to measure caregivers’ Health related Quality of Life, with 19 items, in four domains in a self-version.; CONCLUSIONS Parents and caregivers are the main reporters, by the number of instruments with proxy-versions. The unique instrument for children that has a self-version is the DISABKIDS—Atopic Dermatitis Module, developed to children and adolescents in school age. There is a lack of instruments with self-versions and which give voice to the patient properly. It shows the importance of construction of them and their application in clinical practice to improve the care to this population.
(2010) Factors related to the quality of life of carers of people with Parkinson’s disease Michele Peters, PhD, University of Oxford, United Kingdom; David Morley, University of Oxford; Sarah Dummett, University of Oxford; Laura Kelly, University of Oxford; Jill Dawson, University
123
76 of Oxford; Raymond Fitzpatrick, PhD, Univ of Oxford, Dept of Public Hlth, United Kingdom; Crispin Jenkinson, Health Services Research Unit, United Kingdom AIMS The quality of life (QoL) of informal caregivers can be adversely affected by a number of factors. This issue, however, has not been well explored for carers of people with Parkinson’s disease (PD). This study aims to determine the main influences on carer QoL in this population.; METHODS Caregivers and people with Parkinson’s (PWP) were recruited via local branches of Parkinson’s UK. Parkinson’s specific questionnaires were administered by post. Carers completed the PDQ-Carer, a 29 item instrument incorporating four dimensions (Social and Personal Activities, Anxiety and Depression, Self-Care and Stress). PWP completed the PDQ-39, a 39 item instrument containing eight dimensions (Mobility, Activities of Daily Living, Emotional Well-Being, Stigma, Social Support, Cognitions, Communication and Bodily Discomfort). Reminder letters were sent 4 weeks after the original mailing and a response rate of 61 % was achieved. Regression analyses were performed to identify factors related to carer QoL.; RESULTS The sample comprised 238 carers (mean age 68.20 years) and 238 PWP (mean age 71.64). The mean duration of caring was reported as 8.13 years. Carer Social and Personal Activities were significantly related to caregiver age (p = 0.00), PWP Mobility (p = 0.00) and Cognitions (p = 0.00). Carer Anxiety and Depression were significantly related to caregiver age (p = 0.00), length of caregiving (p = 0.03), PWP Emotional Well-Being (p = 0.03), Cognitions (p = 0.00) and Communication (p = 0.04). Caregiver Self-Care was significantly related to caregiver age (p = 0.02), length of caregiving (p = 0.04), PWP Mobility (p = 0.00), Emotional well-being (p = 0.02) and Cognitions (p = 0.02). Caregiver Stress was significantly related to caregiver age (p = 0.02), length of caregiving (p = 0.01) and PWP Cognitions (p = 0.00).; CONCLUSIONS Results suggest multiple influences on the QoL of caregivers of PWP. Caregiver age and length of time in the caregiving role appear to be of particular importance, as do PWP levels of mobility and cognitive impairment. Practitioners and service providers should be aware of the heightened impact of PD on carers over time and also as PWP symptoms deteriorate.
(2011) The association between a parents’ emotional distress and the health-related quality of life of youths with type 1 diabetes: is the parents’ perception of the social and familial impact of the illness a relevant mediator? Helena Moreira, Faculty of Psychology and Education Sciences, University of Coimbra, Portugal; Maria Cristina Canavarro, Universidade de Coimbra, Portugal; Monika Bullinger, PhD, Inst for Medical Psychology, Hamburg Univ, Germany AIMS Research, particularly among children with diabetes, on the influence of parental distress on children’s health-related quality of life (HRQoL), is scarce. Additionally, the mechanisms through which the parents’ distress may influence their children’s HRQoL, are unknown. This study aims to examine the influence of parents’ emotional distress on children’s HRQoL and examine the mediation role of the parents’ perception of the social/familial impact of the illness on their lives. Additionally, we intend to explore whether this process applies to both children and adolescents.; METHODS The sample included 68 dyads composed of a youth with diabetes and one parent (92.6 % mothers). Youths’ ages ranged from 8 to 18 years old with a mean age of 13.31 years (SD = 3.04; 36.8 % \13 years). Children completed the DISABKIDS chronic generic module (HRQoL) and parents completed the HADS (emotional distress) and IOF (perceived impact of illness). Two moderated mediation models were tested, with depression/anxiety as independent variables; youths’ general HRQoL as the dependent
123
Qual Life Res (2012) 21:1–132 variable; perceived impact as the mediator; and youths’ age as the moderator. The PROCESS computation tool (Hayes & Preacher, 2012) was used. Conditional indirect effects were generated and interactions were probed through percentile based bootstrap confidence intervals.; RESULTS The conditional indirect effects of emotional distress on youths’ HRQoL through the parents’ perceptions of the impact of the illness were significant for adolescents above 14 years old only, i.e., above the 50th age percentile (depression model: point estimate at 50th percentile = -0.40; BCa 95 % CI of -1.05 to -.07; anxiety model: point estimate at 50th percentile = -0.61; BCa 95 % CI of -1.42 to -.01).; CONCLUSIONS The results of this study showed that the parents’ emotional adjustment has a key influence on the adolescents’ HRQoL, through the parents’ perception of the impact of diabetes on their lives. More anxious and depressed parents may perceive the adolescents’ diabetes as more burdensome, which seems to have a negative influence on adolescents’ HRQoL. This study is innovative and has significant clinical and theoretical implications as it highlighted an important mechanism through which the influence of parental adjustment on youth’s HRQoL is exerted, as well as the age at which it begins to manifest.
(2012) Quality of life interaction between caregivers and their family patients with hemodialysis Hiromasa Hirai, Japanese Society of Quality of Life Research, Sumaku, Kobe, Hyogo, JAPAN; Takashi Nakano, Japanese Society of Quality of Life Research, Japan; Satio Kawanishi, Japanese Society of Quality of Life Research, Sumaku, Kobe, Hyogo, JAPAN; Kunio Nakayama, Japanese Society of Quality of Life Research, Japan; Yuri Kitamura, Japanese Society of Quality of Life Research, Sumaku, Kobe, Hyogo, JAPAN; Rika Hayashida, MS, Siebold Univ of Nagasaki, Japan; Michiko Kobayashi, MD PhD, Kwassui Women’s College, Japan; Haruyasu Fujita, Japanese Society of Quality of Life Research, Sumaku, Kobe, Hyogo, JAPAN; Kozaburo Adachi, PhD, Agape-Kabutoyama Hospital, Japan; Takashi Mandai, MD, Japanese Society of Quality of Life Research, Japan AIMS The purpose of this study was to investigate the quality of life (QOL) interaction between caregivers and their family patients with hemodialysis.; METHODS Forty-four couples of caregivers and their family patients with hemodialysis were objects of this study. Including 44 couples, 66 caregivers and 91 patients with hemodialysis participated in this study. Our new original self-administered QOL questionnaire including 40 questions divided into 13 categories with 16 caregivers who took care of family patients with hemodialysis specific questions was used for caregivers. On the other hand, our new original self-administered QOL questionnaire including 40 questions divided into 13 categories with 21 hemodialysis patient specific questions was used for patients, too.; RESULTS Cronbach’s alpha coefficients of our questionnaire were excellent enough to accept for clinical use: 0.87 in mental function, 0.85 in dietary problems, etc. before hemodialysis on caregivers, and 0.87 in dietary problems, 0.83 in economical condition, etc. after one on caregivers. And, 0.91 in mental function, 0.85 in dietary problems, etc. before hemodialysis on patients, and 0.88 in mental function, 0.85 in medical service, etc. after one on patients, respectively. Our QOL questionnaire contained 13 main factors and cumulative contribution was 0.87 after hemodialysis on caregivers. The one also contained 12 main factors and cumulative contribution was 0.85 before hemodialysis on caregivers. Our QOL contained 12 main factors and cumulative contribution was 0.79 after hemodialysis on patients. The one also contained 10 main factors and cumulative contribution was 0.78 before hemodialysis on patients. There was significantly positive correlation between the mean caregivers’ QOLs and the main patients’ QOLs before hemodialysis (r = 0.57, p \ 0.001). There was also significantly positive correlation between the mean caregivers’ QOLs and
Qual Life Res (2012) 21:1–132 the mean patients’ QOLs after hemodialysis (r = 0.42, p \ 0.01), too. Beyond our expectation, both about one tenth caregivers and patients with hemodialysis showed the improvement of QOL after hemodialysis compared with before one.; CONCLUSIONS These findings indicate that our QOL questionnaire has excellent enough reliability and potency of validity to investigate the QOL interaction between caregivers and their family patients with hemodialysis. There were significantly positive correlations between patients’ QOL and caregivers’ ones both before and after hemodialysis. There is significant interaction on QOL between caregivers and patients with hemodialysis. Both caregivers and patients with hemodialysis must burden various limitations for long time, so excellent QOLs are essential for both patients and caregivers to continue the excellent medical treatment and care. Therefore we must pay more attention to improve the QOLs not only in patients but also in caregivers who take care of family patients with hemodialysis.
(2013) Anxiety and depression in fathers and mothers of children with a chronic disease Hedy Van Oers, Emma Children’s Hospital/Academic Medical Center; Lotte Haverman, MSc, Emma Children’s Hospital— Academic Medical Centre, Netherlands; Perrine Limperg, Emma Children’s Hospital/Academic Medical Center, Netherlands; Alice Van Dijk-Lokkart, VU medical centre; Heleen Maurice-Stam, MSc, Emma’s Children’s Hosp, Academic Medical Ctr, Netherlands; Martha Grootenhuis, PhD, Academic Medical Center, Netherlands AIMS We aimed to determine the levels of anxiety and depression in parents of children with a chronic disease compared to a Dutch reference group and to study which parental and child characteristics are associated with parental anxiety and depression.; METHODS Parents of a child with a chronic disease (0–18 years) were eligible and informed with announcements or actively approached at the outpatient clinics of the Emma Children’s Hospital and VU medical centre. After providing their email-address, parents received a login to complete online questionnaires. Anxiety and depression was assessed with the HADS (Hospital Anxiety and Depression Scale). Scores were compared to Dutch norm with t-tests and CHI2 test. Linear regression analyses were performed to examine which child and parental variables were associated with parental anxiety and depression.; RESULTS 566 mothers and 123 fathers participated. Fathers’ anxiety scores were comparable to the norm. However, fathers of the chronically ill children showed significantly more feelings of depression than fathers in the reference group (p \ .05). Mothers had significantly higher scores (p \ .0001) on both anxiety and depression. The amount of mothers in the clinical range of anxiety (31.8 vs 20.7 %) and depression (23.0 vs 12.0 %) was also significantly higher. Highest scores were reported by parents of children with cancer and end-stage renal disease. Practical problems in daily life and parenting stress showed the strongest association with anxiety and depression.; CONCLUSIONS Mothers of children with a chronic disease reported seriously high levels of anxiety and depression. Disease related characteristics of the child did not predict parental anxiety and depression. Practical problems in daily life and parenting stress showed the strongest association with anxiety and depression. Structural attention and supportive care by health care specialists is necessary.
(2014) A clinic-based assessment of the relation of depression and other clinical parameters to clinically significant worsening in walking time in persons with multiple sclerosis Deborah Miller, PhD, The Cleveland Clinic Foundation; Nicolas Thompson, Cleveland Clinic; Jeffrey Cohen, Clevleand Clinic;
77 Robert Fox, Cleveland Clinic; Jennifer Hartman, Cleveland Clinic; Kathleen Schwetz, Cleveland Clinic; Richard Rudick, Cleveland Clinic AIMS Depression is estimated to be twice as prevalent in persons with multiple sclerosis (MS) compared to the general population and may affect disease severity. If so, then treating depression might affect disease progression and improve quality of life. We created a model to describe the association of baseline patient-reported depression and other clinical characteristics with time to a clinically significant worsening (CSW) of 20 % in the Timed-25-Foot Walk (T25FW), a standard measure of walking ability in MS.; METHODS We analyzed data from MS patients with at least two clinical appointments with T25FW measurements (January 2008-December 2009). Data included demographics, patient- and clinician-reported health and functional status measures (HSMs) collected during routine appointments, using an information technology platform developed to collect HSMs. Explanatory variables were PHQ-9 score, age, gender, race, marital status, T25FW at first visit, and MS clinical course. These data were collected at irregular intervals, per clinical indication during usual care. Time to CSW was estimated from a parametric survival model. Because it is not possible to determine when between appointments the CSW occurred, data were considered interval-censored. Therefore, we fit an accelerated failure time model with a Weibull distribution imposed on survival times.; RESULTS Of 1,568 patients available, 314 experienced CSW. Those with a primary or secondary progressive clinical course were estimated to experience 20 % CSW in about 40 % of the time it took the relapsing-remitting patients. Gender (p = 0.005), T25FW baseline time (p = 0.0001), baseline clinical course (p = 0.001), and PHQ-9 score (p = 0.002) significantly affected time to 20 % CSW. The estimated time ratio for the PHQ-9 score was 0.966. Holding all other covariates constant, for every 1-unit increase in PHQ-9score, the time to 20 % CSW decreased by 3.4 %, on average. Thus, for two patients with identical covariate values but PHQ-9 scores of 15 and 5, the patient with a PHQ-9 of 15 would be expected to experience 20 % progression in about 70.8 % (0.96615–5) of the time it would take the patient with a PHQ-9 of 5.; CONCLUSIONS Depression is significantly associated with worsening ambulation in MS. Appropriate depression management may extend the time to CSW in walking time.
(2015) Can psychosomatic symptoms explain gender and age differences in health related quality of life among Swedish schoolchildren? Eva Boman, PhD, University of Gavle, Sweden; Pia Svedberg, PhD, Karolinska Institutet, Sweden AIMS The aims were to study if there are differences in Health Related Quality of Life (HRQoL) between girls and boys in two different age groups and if there are any interactions between sex and age group on different dimensions of HRQoL. Finally, if health symptoms can explain differences in HRQoL between girls and boys and between two age groups.; METHODS A cross-sectional study including 253 children, 99 (n = 51 girls, n = 48 boys) from grade five (ages 11–12 years) and 154 (n = 82 girls, n = 72 boys) from grade nine (ages 15–16 years) in Swedish schools. The KIDSCREEN-52 instrument covering 10 dimensions of HRQoL and questions about health symptoms were analyzed. Sex, age group differences and interaction effects were assessed with analyses of variance (MANOVA) on the KIDSCEEN-52 dimensions. To explore if symptoms explain sex and age group differences in HRQoL a regression analysis was conducted.; RESULTS Boys rated higher on HRQoL dimensions physical well-being, moods and emotions, self-perception, and autonomy, compared to girls. Children in grade five rated their HRQoL as better compared to those in grade nine on the dimensions: physical well-being, psychological well-being,
123
78 moods and emotions, self-perception, autonomy, and school environment. Interaction effects between sex and age group on moods and emotions and physical well-being were found. Eight health symptoms explained 26–49 % of the variance in HRQoL and most important were sleeping difficulties, depression and concentration difficulties. Tension, depression and concentration difficulties were associated to HRQoL among girls whereas stomach ache was associated to HRQoL among boys.; CONCLUSIONS Girls and adolescents experience worse HRQoL as compared to boys and younger children and psychosomatic symptoms seem to explain a substantial part of the variation in HRQoL. Sleeping difficulties was a major obstacle to all children’s HRQoL but also depression and concentration difficulties had a major impact. Tensions, depression and concentration difficulties were particularly associated to HRQoL among girls whereas stomach ache among boys. Finding strategies to promote health among schoolchildren appears to be of great importance, both in general but also from a gender perspective.
(2016) Health-related quality of life, depression, and sense of coherence in hospitalized patients with chronic heart failure Viviane M. Pelegrino, University of Sa˜o Paulo; Rosana Spadoti Dantas, PhD, University of Sa˜o Paulo, Brazil; Ma´rcia A. Ciol, University of Washington; Lidia A. Rossi, PhD, University of Sao Paulo, Brazil; Suemara P. Silva, University of Sa˜o Paulo; Carina M. Dessotte, University of Sa˜o Paulo; Fabiana Bolela, University of Sa˜o Paulo AIMS To assess the correlation between the measures of healthrelated quality of life (HRQoL), sense of coherence (SOC), and depression in patients hospitalized with chronic heart failure (CHF).; METHODS This was a cross-sectional study of 90 patients diagnosed with CHF hospitalized in a university hospital in Ribeira˜o Preto, Sa˜o Paulo state, Brazil. Data collected included demographic characteristics, the Minnesota Living with Heart Failure Questionnaire (MLHFQ, with possible range 0–105 where lower scores indicate better HRQL), the 13-item Sense of Coherence scale (SOC-13, range 13–91, higher scores reflecting higher coping capacity), and the Hospital Anxiety and Depression Scale—subscale Depression (HADS-D, range 0–21, higher scores indicating more depressive symptoms). Validated translations of instruments were used. Data were analyzed by means of descriptive statistics and Pearson correlations.; RESULTS Mean age was 62 years (s.d = 14.3), 51 % were female, 57.8 % were married or living with a significant other, mean number of years of education was 3.8 years (SD = 3.1), and 75.9 % were classified as NYHA-class III-IV. Average scales scores were: 65.9 (s.d. = 19.7; range 3–97) for MLHFQ, 61.1 (s.d. = 15.3; range 31–91) for coping capacity, and 7.3 (s.d. = 5.1; range zero to 18) for depression. Associations were found between MLHFQ and HADS-D (r = .259; p = 0.01), and between SOC-13 and HADS-D (r = -.447, p \ 0.001).; CONCLUSIONS The results show that people who reported higher levels of depression also reported lower quality of life and lower coping capacity. Interventions to improve any one of those aspects of life (quality, coping, and depression) might have an effect on the others, though this needs to be assessed in a longitudinal study.
(2017) Biological signal for human flourishing: the necessity of adjusting cognitive misjudgement in the capability approach Tadashi Hirai, PhD, The University of Tokyo, Japan; Yukio Ikemoto, The University of Tokyo AIMS If it is the case that we are biologically inclined to flourish as a species, then why do we continue to strive towards greater wealth at
123
Qual Life Res (2012) 21:1–132 the expense of our well-being? Although some previous research has examined this implication (e.g. Easterlin 1974), nothing has been done to address the fundamental relationship between our biological structure and cognitive judgment. This research aims to investigate mechanism in terms of happiness with a special focus on the application to the Capability Approach where autonomy has been stressed (e.g. Sen 1999).; METHODS For analytical purposes, three types of happiness are introduced: ‘biological happiness’ represents the organismic side based on human nature; ‘cognitive happiness’ and ‘eudaimonic happiness’ represent the cognitive aspect based on individual reasoning and public reasoning respectively. Using an empirical study by Richard Wilkinson, the relevance of stress to cognitive judgment is examined.; RESULTS In the Easterlin Paradox, biological happiness decreases whilst cognitive happiness increases. This reverse correlation is caused by the fact that individual reasoning has malfunctioned in its adaptation and social comparison. Adaptation and social comparison are essential for successful living: adaptation helps us modify our fluctuating living conditions; whereas social comparison helps us find individual strength within our society. However, being applied to materialistic goods (e.g. money, status), something easier to adapt to and compare, above a certain threshold, these features no longer provide biological happiness because of increased stress caused by never-satisfied material desire and never-ending competition at the sacrifice of intrinsic goods (e.g. human relation, dignity). Give that this situation jeopardizes well-being not only for the loser but also the winner in a society (Wilkinson 2005), it cannot be justified even from the perspective of other-regarding agency let alone one’s own wellbeing. Biological sign becomes indicative here, since individual cognitive judgment cannot always offer an accurate mediator towards human flourishing.; CONCLUSIONS In today’s world, individual cognitive judgment is vulnerable to adaptation and social comparison and so needs to be adjusted towards our ultimate objective. For this purpose, biological signal should be considered more seriously within the Capability Approach, given that it offers some important insight for public reasoning towards human flourishing.
(2018) Measuring ‘instrumental activities in daily living’ in dementia: the diagnostic usefulness of the Amsterdam IADL QuestionnaireÒ Sietske Sikkes, PhD, VU University Medical Center, Netherlands; Yolande Pijnenburg, VU University Medical Center; Elly De Lange-de Klerk, VU University Medical Center; Philip Scheltens, VU University Medical Center; Bernard Uitdehaag, VU University Medical Center AIMS Dementia is characterized by a progressive loss of cognitive functions, causing interference with the person’s ability to function independently. Informant-based questionnaires aimed at ‘instrumental activities of daily living’ (IADL) are generally used to measure the level of this interference. However, the quality of these instruments is limited. We therefore developed the Amsterdam IADL Questionnaire , aimed at measuring IADL in early dementia. We previously found good psychometric properties and we now aim to investigate the diagnostic usefulness of this questionnaire.; METHODS Informants of patients who visited the Alzheimer Center of the VU University Medical Center for dementia screening completed the Amsterdam IADL Questionnaire. Diagnoses were made independently of the Amsterdam IADL scores. Latent trait estimates were calculated using item response theory modeling. Amsterdam IADL scores of patient with and without dementia were compared using an independent t-test. Receiver operating characteristic curves were used to evaluate the diagnostic ability of the Amsterdam IADL Questionnaire and the Disability Assessment for Dementia (DAD), a commonly used IADL instrument.; RESULTS A total of 196 patients were diagnosed and had informants complete the
Qual Life Res (2012) 21:1–132 questionnaire. Patients diagnosed with dementia (n = 97) had lower Amsterdam IADL scores than patients without dementia (n = 99), t(194) = 5.1, p \ .001. The AUC for the Amsterdam IADL Questionnaire (.70, 95 % CI: .63–.78) was higher than for the DAD questionnaire (.63, 95 % CI: .54–.73).; CONCLUSIONS Results suggest that the Amsterdam IADL questionnaire is helpful in diagnosing dementia and performs better than an existing IADL questionnaire.
79 positively associated with emotional well-being in cross-sectional analysis. In contrast, only engaging in a hobby was associated with emotional well-being in longitudinal analysis.; CONCLUSIONS These results indicate that older adults who have favorable lifestyle (i.e. frequently go out, engage in a hobby, and good dietary habit), are more likely to be high in emotional well-being. In addition, engaging in a hobby predicted maintenance of emotional well-being in longitudinal analysis, suggesting that engaging in a hobby is particularly important for emotional well-being community-dwelling older adults.
(2019) Basic genetics: what the QOL researcher needs to know Jeff Sloan, PhD, Mayo Clinic, United States; Mirjam AG Sprangers, PhD, Academic Medical Center, Netherlands AIMS Exploring the genetic underpinnings of QOL domains is an emerging scientific endeavour. The aim of this presentation is to provide QOL researchers with basic information about genetics in order to understand how they might incorporate genetic variables into their QOL research.; METHODS We will review findings of recent publications of the GENEQOL consortium, focusing on interpretation of results for the non-geneticist. We will provide a glossary of genetic terms and a ‘‘roadmap’’ for how to incorporate genetic variables into QOL research studies. RESULTS Recent research has indicated relationships between cytokines, folate genes, the glutathione metabolic pathway and fatigue/QOL domains. Results have been replicated across separate studies and different disease populations. CONCLUSIONS QOL researchers need basic knowledge of genetic terms and recent results to be able to incorporate genetic variables into their research.
(2020) Emotional well-being and lifestyle factors among community-dwelling older adults Yuko Yoshida, Japan; Hajime Iwasa, Tokyo Metropolitan Institute of Gerontology; Shu Kumagai, University of Human Arts and Sciences; Takao Suzuki, National Center for Geriatrics and Gerontology; Yuko Yoshida, Japan; Yuko Yoshida, Japan; Yuko Yoshida, Japan; Hideyo Yoshida, Tokyo Metropolitan Institute of Gerontology, Japan AIMS Well-being is one of the important factors for achieving successful ageing, and it is useful to examine the factors that are relevant to a status of well-being. The purpose of this study was to clarify the cross-sectional and longitudinal relationship between emotional wellbeing and lifestyle factors among community-dwelling older adults in Japan.; METHODS The participants were 475 individuals (196 men and 279 women) aged 73 years and older who participated in a baseline survey in August 2009 and were subsequently followed annually for 2 years. The baseline survey had gathered data on basic demographic factors (sex, age, activities of daily living (ADL) and living arrangement) and lifestyle factors (drinking habit, smoking habit, frequency of going out every week, physical activity, sleep duration, engaging in a hobby, and dietary habits (dietary variety score consisting of 10 food category items)). Emotional well-being was assessed using the five-item Japanese version of the World Health Organization Well-Being Index (WHO-5). Multiple linear regression analysis, adjusted for age, sex, ADL, living arrangement and the WHO-5 (only in longitudinal analysis) was used to examine the association between emotional well-being and lifestyle factors.; RESULTS The mean age of the study participants was 78.3 ± 4.3 years (range 73–96 years). Among the participants, 86.7 % were physically independent, 9.1 % lived alone, 30.7 % drank alcohol, 6.4 % smoked, 81.4 % went out every day of the week, 37.4 % were physically active, 33.1 % slept 7–8 h per night, 49.3 % engaged in a hobby, and the mean number of food items consumed daily was 4.4 ± 2.6. Multiple regression analysis showed that the frequency of going out, engaging in a hobby, and dietary variety score were
(2021) Busting the top myths about quality of life assessment in clinical practice Carolyn C. Gotay, PhD, UBC School Pop & Pub Health, Canada AIMS Despite advances in quality of life (QOL) research and the potential for QOL data to inform clinical care, QOL tools remain underutilized in clinical care. This presentation will identify and dispel leading myths that may limit utilization.; METHODS During the author’s multi-decade experience working in this field, she has heard many contentions from clinicians about why they don’t incorporate systematic QOL assessment in their practice. Each of these contentions is a myth, countered by research data. This presentation will identify leading myths and provide data from the literature that argue against each myth.; RESULTS The seven myths are: (1) Patients don’t want to fill out questionnaires; (2) If my patients had QOL concerns, they would raise them with me; (3) QOL is a philosophical construct that can’t be measured; (4) It takes a long time because you have to ask so many questions; (5) I don’t have a nurse who can give out these questionnaires; (6) These questions will trigger lengthy discussion and will increase the length of the clinic visit; (7) QOL data haven’t informed clinical practice so there’s no need to ask about it. Data from QOL research will be summarized in response to each point.; CONCLUSIONS QOL assessment may be under-utilized because of unfounded beliefs about how to do it and the impact of such assessment on clinical practice. Data are available to counter these misconceptions. This presentation will provide useful pointers to ‘‘bust the myths’’ and facilitate clinical use of QOL assessment.
(2022) Symptoms and quality of life indicators of persons with lower limb loss Dagmar Amtmann, PhD, University of Washington; Daniel C. Abrahamson, University of Washington; Sara Morgan, University of Washington; Rana Salem, University of Washington; Robert L. Askew, BA, UT—M.D. Anderson Cancer Center; Brian Hafner, University of Washington AIMS The purpose of the study was to examine the pattern of symptoms and quality of life (QoL) indicators in a large sample of amputees and to compare their levels of pain interference, fatigue, physical function, sleep disturbance, depression, and satisfaction with participation in social roles to the PROMIS US general population norms.; METHODS The Patient Reported Outcomes Measurement Information System (PROMIS)-29 was administered to lower limb prosthetic users as a part of a cross-sectional study of mobility. PROMIS-29 assesses respondents on seven health domains: physical function, anxiety, depression, fatigue, sleep disturbance, participation in social roles and pain interference. PROMIS T-scores of 50 represent the US general population mean in each domain. Respondents were recruited via prosthetic clinics, magazine advertisements, listservs, websites, and social networks. Inclusion criteria were 18 ? years of age, ability to read English, unilateral lower limb amputation
123
80 with traumatic or dysvascular etiology, and use of prosthesis to ambulate. Items describing demographic and disease characteristics were also included.; RESULTS The PROMIS-29 was administered to 650 persons with limb loss (mean age = 53 years, SD = 14 years; 69 % male; 36 % transfemoral amputation, 66 % traumatic amputation). Respondents reported lower physical function (M = 42.69, SD = 8.6; p \ 0.01) and fatigue (M = 48.10, SD = 9.8; p \ 0.01), and higher pain interference (M = 54.48, SD = 9.2; p \ 0.01) than the general US population. Mean anxiety (M = 49.32, SD = 9.8; p \ .08), depression (M = 49.11, SD = 9.5; p \ 0.02), satisfaction with participation in social roles (M = 48.70, SD = 10.3; p \ 0.01) and sleep disturbance (M = 49.30, SD = 9.0; p \ 0.05) levels were similar to general population norms even though some differences reached statistical significance.; CONCLUSIONS Worse physical function and more pain would be expected in a population of users of lower limb prosthetics. That persons with limb loss report lower fatigue than the general population was not anticipated and has not been previously described. We hypothesize that reduced capacity for physical activity and pain associated with limb loss lead to limited physical activity and, by extension, lower fatigue. PROMIS-29 is an efficient and informative way to describe the pattern of symptoms and QoL indicators in amputees. The comparisons with general US population norms provide valuable information about functioning of individuals with limb loss.
(2023) Translation and linguistic validation of two COPD symptom diaries for use in global clinical trials Sonya Eremenco, MA, United BioSource Corporation; Michelle Mocarski, MPH, Forest Laboratories; Benjamin J. Arnold, MA, FACITtrans, Brazil; Andrew Palsgrove, United BioSource Corporation; Asha Hareendran, PhD, United BioSource Corporation, United Kingdom AIMS Two patient-reported outcome (PRO) measures, the Nighttime Symptoms of COPD Instrument (NiSCI) and Early Morning Symptoms of COPD Instrument (EMSCI), were developed to support treatment benefit endpoints in global clinical trials. Translations that were conceptually equivalent to the English source version and easily understood by the target country populations were needed. The purpose of this study was to translate and assess conceptual equivalence of the NiSCI and EMSCI for use in 6 countries: United States (Spanish), Canada (French), Australia (English), New Zealand (English), Hong Kong (Traditional Chinese and English) and Singapore (Simplified Chinese and English).; METHODS The NiSCI and EMSCI were translated following ISPOR guidelines for translation and cultural adaptation of PRO measures (Wild et al., 2005) using the universal approach discussed in the second Task Force Report (Wild et al., 2009). Translation into Spanish occurred in parallel with usability testing which resulted in minor changes to the English version. The revised English source version was reviewed by native English speakers in Australia, New Zealand, Singapore and Hong Kong to ensure suitability. For each non-English language, 2 forward translations by native translators, reconciliation of the forward translations, 1 back-translation by an English-speaker fluent in the target language, and final reconciliation by a native speaking language coordinator were conducted for both measures. Harmonization was performed to ensure conceptual equivalence across languages. Interviews were conducted among 60 native-speaking COPD patients in 6 countries. Interview data were analyzed to assess linguistic and cultural validity of the diaries in each language and confirm conceptual equivalence.; RESULTS Mean age of the sample (N = 60) was 60.91 years, (range, 40–86) and 68 % were male. The translations were well understood and considered relevant, with patients raising only minor issues during interviews. Challenges arose in finding patient-friendly language to
123
Qual Life Res (2012) 21:1–132 describe rescue medication, but were resolved through consultation with linguists and pilot-testing with patients. Only the Spanish translation required post-testing revision following changes in the English source version.; CONCLUSIONS The universal English, Spanish, French, Simplified and Traditional Chinese versions of the NiSCI and EMSCI were found to be conceptually equivalent and acceptable for use in the 6 countries evaluated.
(2024) Implementation of quality of life evaluation in oncology daily practice: breast cancer in focus Ariana Santos, Portuguese Oncology Institute, Porto, Portugal; Augusta Silveira, University Fernando Pessoa, Portugal; Joaquim Gonc¸alves, Polytechnic Institute of Ca´vado & Ave, Portugal; Cla´udia C. Ribeiro, Universidade Cato´lica Portuguesa, Portugal; Joana Teixeira, Nurse, Portuguese Oncology Institute, Porto, Portugal, Portugal; Joaquim Abreu Sousa, Portuguese Oncology Institute, Porto, Portugal; Teresa Sequeira, Portugal AIMS Quality of life evaluation as a routine clinic practise is still not a common procedure. However, considering breast cancer, recent studies suggest that Quality of Life is an increasingly prominent health marker. The present study aims to validate the EORTC-QLQ C30 and the specific module BR23 Portuguese PC-software versions, in order to implement Quality of Life routine assessment at the Breast Cancer Clinic of the Portuguese Oncology Institute-Porto, Portugal.; METHODS Ninety-nine women from the Breast Clinic of the Portuguese Oncology Institute participated in this study after given their informed consent. Patients answered the questionnaires made in both formats (paper and computer), with an interval of 30 (thirty) minutes between the two evaluations. The computer assessment was made using the Quality of Life Informatics Platform (QoLIP) used at this institution. The Wilcoxon test (95 % confidence interval) and percentiles calculus were used to evaluate sample similarity. In order to understand if QoLIP had any influence in the patient’s responses, we admitted two statistical hypotheses to realize a bilateral test: hypothesis H0:F (90) = F (91) or hypothesis H1:F (90) - F (91).; RESULTS The usability of the platform was high, considering that 76,8 % of the patients preferred the electronic format of the questionnaires. The null hypothesis (H0) was accepted for all questions and so QoLIP revealed to had any influence in the patient’s responses.; CONCLUSIONS The Quality of Life Informatics Platform revealed to be a powerful and effective tool for routine Quality of Life assessment in breast cancer patients. This patient reported outcome may contribute to support clinical decisions, identify unmet needs and can be used in routine care appointments. The electronic Quality of Life assessment revealed to be technically feasible in breast cancer patients at the Breast Clinic of the Portuguese Oncology Institute-Porto.
(2025) Unidimensionality, validity and responsiveness of the ANMS Gastroparesis Cardinal Symptom Index-Daily Diary Dennis Revicki, PhD, United BioSource Corporation; Wen-Hung Chen, PhD, United BioSource Corporation; Michael Camilleri, Mayo Clinic; Henry Parkman, Temple University AIMS Patient-reported symptom scales are necessary for evaluating treatments for gastroparesis. The ANMS Gastroparesis Cardinal Symptom Index-Daily Diary (GCSI-DD) was developed to assess daily symptoms in patients with gastroparesis. To evaluate the dimensionality, reliability, validity and responsiveness of the GCSI-DD in patients with gastroparesis.; METHODS A 4-week observational study was
Qual Life Res (2012) 21:1–132
81
conducted in patients with gastroparesis who were starting a new treatment. Patients completed the GCSI-DD as a paper diary each evening, assessing severity of nausea, early satiety, excessive fullness, bloating, and upper abdominal pain on a 0 (none) to 5 (very severe) scale, and number of episodes of vomiting. The GCSI-DD core symptoms score was calculated averaging the 5 symptom severity items and excluding vomiting episodes. Patients completed global assessments of change in overall symptoms at 4 weeks. Unidimensionality was assessed using confirmatory factor analysis (CFA) and Rasch analysis using data from the current study and from a prior study (Revicki et al. Quality of Life Research 2004;13:833–844). Responders were defined based on patient global ratings of a small or greater improvement in gastroparesis for estimating effect sizes (ES).; RESULTS 69 patients (mean age 40.2 ± 14.9 years, 83 % women) were recruited in the current study and 169 (mean age 45.7 ± 13.1 years, 77 % women) from Revicki et al. (2004). For the CFA, the comparative fit index was 0.96 at baseline and 0.97 at follow-up visit. The 5 items fit the Rasch model and no item had fit residuals outside the range of -3 and 3. Internal consistency (0.89–0.94) and test–retest reliability (ICC = 0.98) were excellent. GCSI-DD item scores were correlated with overall gastroparesis severity (r = 0.48–0.66). Mean core symptom scores were significantly different among clinician rated severity groups (p \ 0.001). Responders reported improvements in nausea (ES = 0.42, p \ 0.001), excessive fullness (ES = 0.83, p \ 0.001), bloating (ES = 0.34, p = 0.006), early satiety (ES = 0.53, p \ 0.001), and abdominal pain (ES = 0.29, p = 0.03) and core symptom scores (ES = 0.55, p \ 0.001).; CONCLUSIONS GCSI-DD demonstrates excellent reliability and validity, and responsiveness to treatment for gastroparesis. A composite of five core symptoms capture the symptom complex of gastroparesis. The GCSI-DD represents a good patient reported outcome for evaluating the effectiveness of treatments in clinical trials for gastroparesis.
and 86,8 % received somatostatin analog treatment. On average, patients were included for 2 years in QOL-monitoring and each patient completed 5,6 assessments (599 assessments in total). Patients admitted for radionuclide therapy received lutetium-170 (55,3 %) or yttrium-90. Mixed-linear model analysis revealed several differences in patients’ QOL (p \ 0.05). No treatment with somatostatin analogs was associated with worse role, social, emotional and cognitive functioning, more fatigue, dyspnea and appetite loss. Female patients reported worse role functioning and more fatigue. Highest fatiguelevels are reported by patients participating one to 2 years in QOLassessment. Patients included longer than 4 years in QOL-monitoring reported better role, social, emotional and physical functioning, less fatigue, pain, appetite loss, sleep disturbances and diarrhea. Irrespective of patients’ age and computer literacy, acceptance of electronic data-collection was high and response rates equal to paper– pencil administration. Though QOL-impairments persisted over a long period of time, the scores slowly approximated the level of healthy controls.; CONCLUSIONS Besides data-collection, CHES provides patients’ data immediately to their medical caregivers. Our data illustrates that NET-patients suffer from several QOL-impairments, also in the long run of treatment. The GI-NET21 module takes the distinct needs of this particular patient group into account. To provide timely and adequate QOL-targeted intervention, routine QOL-monitoring using electronic devices for assessment and datapresentation should be the method of choice.
(2026) Long-term quality of life monitoring in patients with neuroendocrine tumors
AIMS To evaluate the articular movement of HRQL of JIA patients after an exercise program of the Pilates method.; METHODS Clinical trial of a consecutive sample of 50 JIA patients of the oligoarticular (46 %), polyarticular (28 %) and systemic (26) % subtypes, with an average 11,1 ± 4,0 years of age. The patients took part in Pilates sessions twice a week for 6 months. The range of articular movement was measured with the use of a goniometer according to the pediatric Escola Paulista de Medicina Range of Motion Scale (pEPM-ROM) and the HRQL was accessed by means of the comparison of the PedsQL4.0 scores. Both evaluations were performed in the baseline and post-intervention periods. The t-test was used for the comparison of mobility and HRQL in both periods.; RESULTS There has been a significant improvement in range of articular movement of patients at the end of the intervention (p \ 0.05). The averages of the PedsQl.40 scores after 6 months of intervention have been significantly greater (p \ 0.05) in the physical, (52.4 ± 20.3 vs. 88.4 ± 12.4), emotional (48.0 ± 10.4 vs. 86.5 ± 12.6), social (56.9 ± 18.5 vs. 87.2 ± 12.8), school-related (46.1 ± 22.0 vs. 64.9 ± 14.2), psycho-social (50.4 ± 17.1 vs. 79.5 ± 10.6) and total score (53.1 ± 16.5 vs. 80.8 ± 10.9) domains.; CONCLUSIONS The Pilates exercise program has promoted improvement in joint range of motion of JIA patients as well as positive impact on the HRQL in psychosocial and total domains and summaries of the PedsQL4.0 of these patients.
Lisa M. Wintner, Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Austria; Eva-Maria Gamper, Department of Psychiatry and Psychotherapy, Innsbruck Medical University; Johannes M. Giesinger, PhD, Innsbruck Medical University, Austria; Anne S. Oberguggenberger, Innsbruck Medical University, Austria; Sabine Buxbaum, University Clinic for Nuclear Medicine, Innsbruck Medical University; Dietmar Waitz, University Clinic for Nuclear Medicine, Innsbruck Medical University; Daniel Putzer, University Clinic for Nuclear Medicine, Innsbruck Medical University; Bernhard Nilica, Innsbruck Medical University, Austria; Irene Virgolini, University Clinic for Nuclear Medicine, Innsbruck Medical University; Bernhard Holzner, PhD BE, Univ. Hospital, Innsbruck, Austria AIMS As neuroendocrine tumors (NETs) generally are slow-growing tumors with promising treatment possibilities, patients have to manage their lives living with NETs. Since knowledge on long-term quality of life (QOL) in NET-patients is sparse, we implemented routine electronic QOL-monitoring and collected data on patients’ particular symptom and psychosocial burden.; METHODS Patients admitted for an inpatient stay (either for radionuclide therapy or aftercare) were asked to complete the EORTC Quality of Life Questionnaire (QLQ-C30) and its disease-specific module for gastrointestinal NETs (GI-NET21) Assessment was facilitated by the use of the specialized computer software Computer-based Health Evaluation System (CHES).; RESULTS Between 2005 and 2012, 106 patients suffering from NETs participated in routine PRO-monitoring (QLQ-C30). Of these patients, 41,5 % also completed the GI-NET21 module, 37,7 % were female, on average aged 64 years (±11,5) old,
(2027) The effects of Pilates exercises in articular movement and HRQL of JIA patients Tania Maria S. Mendonca, Sr., Universidade Federal de Sao Paulo, Brazil; Claudio A. Len, Sr., UNIFESP-EPM; Maria Teresa RA Terreri, Sr., UNIFESP-EPM; Maria Odete E. Hilario, Sr., UNIFESP-EPM; Carlos Henrique Martins da Silva, MD, Brazil; Roge´rio M C Pinto, Federal University of Uberlandia
(2028) Sense of coherence among people with psoriasis Eva Langeland, Bergen University College, Norway; Hilde S. Robinson, University of Oslo; Marie H. Larsen, University of Oslo, Norway; Anne-Lene Krogstad, Oslo University Hospital; Torbjørn
123
82 Moum, University of Oslo, Norway; Astrid Wahl, PhD, University of Oslo, Norway AIMS The aims for the present study is to investigate Sense of Coherence (SOC); the central concept in salutogenesis and its relation to people with psoriasis by exploring the effect on SOC as a result of climate therapy and investigate factors that predict change in SOC among people with psoriasis.; METHODS The present study has a prospective design, including a baseline assessment and one follow up in relation to a 3 week climate therapy program. A total of 254 adults with psoriasis participated. SOC was measured by the Sense of Coherence questionnaire (SOC-13), illness perception by using the Revised Illness Perception Questionnaire (IPQ-R) and positive and active engagement in life was measured by the positive and active engagement subscale from the Health Education Impact Questionnaire (HeiQ) (all Norwegian versions). To evaluate possible change in SOC score from baseline to follow-up paired samples T-test was performed. With regard to investigate the predictive value of illness perception versus positive and active engagement in life on SOC, multiple linear regression analysis was used.; RESULTS SOC improved significantly before to after climate therapy. Results show a significant statistical change (p \ 0.001) in SOC from baseline to follow-up with a difference of 2.6, (95 % C.I: 1.6–3.6). Furthermore, the attitude of positive and active engagement in life (standardized beta coefficient .22, P = .0.002) and illness perception—coherence dimension (standardized beta coefficient .15, P = .0.04) predict change in SOC.; CONCLUSIONS This is the first study that investigates the effect on SOC as a result of climate therapy. SOC significantly improved before to after climate therapy. Further the results indicate that improving positive and active engagement in life and a coherent understanding of the illness might provide important opportunities to improve SOC among people with psoriasis. We need more studies that investigate SOC among people with psoriasis undergoing climate therapy and what factors that predict positive change in SOC.
(2029) Qualitative interviews with psoriasis patients evaluating paper to electronic migration of the psoriasis symptom inventory (PSI) Mona L. Martin, RN MPA, Health Research Associates, Inc.; TeChieh Chen, Health Research Associates; Dina Chau, Amgen Inc; Hema Viswanathan, Amgen, Inc AIMS Advantages of data collection by electronic format are recognized within the clinical trial environment. The goal of this study was to evaluate and document the migration of a newly developed Patient Reported Outcome (PRO) instrument (the Psoriasis Symptom Inventory—PSI) from the existing paper format to an electronic daily diary format.; METHODS Adult patients (aged = 18 years) with moderate to severe psoriasis were enrolled from 2 sites in the United States. Two waves of cognitive interviews with six patients in each were conducted by trained interviewers to explore patients’ understanding of concepts presented within existing PSI items across the paper and electronic formats. Half of the patients received the electronic version first followed by the paper version; the other half received the paper version first followed by the electronic version. Patients were then interviewed with a series of semi-structured questions assessing their perceived meaning for each item, response option clarity, and whether or not the format change would affect the answer they would give for each item.; RESULTS The twelve patients completing these interviews had a mean age of 49.2 years and a varied range of education, income, and marital status. Patients reported no change in interpretation of the PSI items, response options, or instructions across the two administration modes. One patient found that viewing a single item at a time in the electronic
123
Qual Life Res (2012) 21:1–132 format, rather than seeing all items at once in the paper format, potentially resulted in more thought given to the question prior to responding. However, there were no differences between the two modes in the responses patients selected, nor were there differences between those completing the paper version first versus the electronic version first. All patients reported that the electronic version was acceptable and simple to use with easy readability of the screen and response options.; CONCLUSIONS Results of the interview process indicated successful migration from paper to electronic format of the PSI and patients confirmed satisfactory ease of use for the electronic device.
(2030) Feasibility and acceptability of patient-reported outcomes data collection for clinical care following breast reconstruction Andrea Pusic, MD MHS FRCSC, Memorial Sloan Kettering Cancer Ctr; Stefan Cano, BSc PhD, Peninsula College of Medicine & Dentristry, United Kingdom; Anne Klassen, D Phil, McMaster University, Canada; Amie M. Scott, MPH, Memorial Sloan-Kettering Cancer Center; Marwan Shouery, Memorial Sloan-Kettering Cancer Center; Ethan Basch, MD MSc, Memorial Sloan-Kettering Cancer Center; Peter Cordeiro, Memorial Sloan Kettering Cancer Center AIMS To date, systematic measurement of patient-reported outcomes (PROs) has played an important role in cancer research, but not in routine clinical care. Our objective was to evaluate the feasibility of developing and piloting an electronic PRO data collection in clinical care among breast reconstruction patients using the BREAST-Q, a previously developed condition-specific PRO measure for breast surgery patients that measures quality of life (e.g. psychosocial, physical and sexual well-being) as well as patient satisfaction (e.g. satisfaction with breasts, with information, with surgeon).; METHODS The BREAST-Q was loaded to the MSKCC WebCore, a generic electronic patient-reporting platform adhering to strict privacy and security standards. Patients attending visits at the MSKCC Breast Reconstruction Clinic were asked to complete the BREAST-Q electronically prior to scheduled visits. For patients with email addresses, a reminder with web-link to the questionnaire was emailed automatically prior to the visit.; RESULTS Over a 9 month start-up period, BREAST-Q surveys were completed by 1,442 patients. Patients completed the questionnaire at set time points before and after surgery. A total of 2,340 BREAST-Q surveys were completed overall. Mean completion time was 5:53 min. Acceptability was high with both patients and clinical staff contributing positive comments along with suggestions for improvement via email.; CONCLUSIONS This pilot experience suggests that ePRO data can be efficiently collected among outpatient breast surgery patients with high acceptability. In the next phase of this project, we will introduce real-time individual patient reports to the clinical team and evaluate the impact of this information on clinical care and quality improvement.
(2032) Effectiveness of social supports on quality of life in breast cancer patients: a meta-analysis of randomized controlled trials Ayako Matsuda, National Cancer Center, Japan; Kazue Yamaoka, PhD, Teikyo University, Graduate School of Public Health, Japan; Toshiro Tango, Center for Medical Statistics, Japan; Tomohiro Matsuda, National Cancer Center, Japan; Hiroshi Nishimoto, National Cancer Center, Japan AIMS Most of the breast cancer patients receive social supports, however effect of the social supports is still unclear. Quality of life
Qual Life Res (2012) 21:1–132 (QOL) is one of the important assessment tools. The aim of this study was to evaluate the effectiveness of social supports for breast cancer patients, taking into consideration QOL using a meta-analysis.; METHODS We conducted search on the Cochrane Database and MEDLINE up to April 2012 to identify randomized controlled trials (RCTs) of breast cancer treatments with QOL measurement as an outcome. Standardized mean differences of the global QOL at less than 5 months (\5 m) and 5–6 months (5–6 m) after intervention in breast cancer patients receiving social supports (intervention group) compared with the non-social support (control group) were defined as the effect sizes in this study. The social supports were then classified to three groups, psychoeducational support, psychological support and coping group. The effect sizes were analysed using a fixed-effects model and a random-effects model (DerSimonian-Laird Method).; RESULTS In total, 11 RCTs were eligible and used for the metaanalyses: Psychoeducational support group; 4 RCTs (\5 m) and 3 RCTs (5–6 m), psychosocial support group; 3 RCTs (\5 m) and 3 RCTs (5–6 m), coping group; 3 RCTs (\5 m). In less than 5 months, the only effectiveness of psychoeducational support group was observed with increasing global QOL (mean difference = 2.42, 95 % CI: 0.51–4.34; P = 0.013, the random effects model). In 5–6 months, none of the effectiveness of social support groups was observed.; CONCLUSIONS Following the intervention of psychoeducational support, the higher global QOL was reported after less than 5 month, but not in psychological support and coping group. In order to improve QOL of the breast cancer patients, further longterm intervention study to examine the effects of social supports is needed.
(2033) Patient-reported outcome is important in psychosocial intervention for dementia: a secondary analysis on RCT of group reminiscence approach data Kenichi Meguro, Dpt Geriatr Behav Neurol, Tohoku Univ Grad Sch Med, Japan; Kyoko Akanuma, Dpt Geriatr Behav Neurol, Tohoku Univ Grad Sch Med AIMS We previously (Dement Geriatr Cogn Disord 2007;24:48) evaluated the beneficial effect of the group reminiscence approach (GRA) in patients with vascular dementia (VaD). Sixty patients were randomly assigned to three arms for once a week for 3 months: GRA arm, social contact (SC) arm, and control arm. Improvement in cognitive function and behavioral activities were defined as the primary outcome. No significant improvements in the main outcome measures seemingly supported the hypothesis that GRA was beneficial. However, the dropout ratio was higher in the SC arm (45 %) compared with the GRA (15 %) and control (15 %) arms. The participants in the SC arm pointed the intervention very boring.; METHODS Since we had assessed their emotional ratings for the GRA and SC arms, we herein performed the secondary analysis. We had assessed 5 ratings: ‘‘much enjoyable,’’ ‘‘enjoyable,’’ ‘‘normal,’’ ‘‘boring,’’ and ‘‘much boring.’’ For the GRA arm, 13/17 (=76 %) answered much enjoyable or enjoyable, whereas the SC arm, 3/11 (=27 %) answered so. We thus operationally classified them into two groups, ‘‘Enjoyable’’ group (n = 16) and ‘‘Boring’’ group (n = 12), independent of ‘‘outer’’ approach by intervention of the GRA or the SC. The standard deviation (SD) of changes in MOSES over 21 days was 5.01 in the placebo arm. Decrement greater than the SD multiplied by 1.96, i.e., 10 or more of the MOSES was defined as improvement.; RESULTS We disclosed that the ‘‘Enjoyable’’ group had statistically significant improvement on the MOSES (p \ 0.05) compared with the ‘‘Boring’’ group.; CONCLUSIONS Contrary to the clinical trials of drugs, the psychosocial approaches may not be adequate for the RCT design, and should be cautious for ‘‘superficial’’ approach. Instead, patient-reported outcome (PRO) may be important for psychosocial intervention.
83 (2034) CONSORT PRO 2012: translating guidance to practice Melanie Calvert, PhD, The University of Birmingham, United Kingdom; Jane Blazeby, MD, University of Bristol, United Kingdom; Doug G. Altman, University of Oxford, United Kingdom; Dennis Revicki, PhD, United BioSource Corporation; Brenda Bass, Queen’s University, Canada; David Moher, Ottawa Hospital Research Institute; Michael Brundage, MD, Queen’s University, Canada AIMS The CONSORT PRO 2012 Statement provides evidence-based guidance for improving reporting of Patient Reported Outcome (PRO) data in randomised trials. Recognizing that translation of this knowledge needs to occur internationally within a complex system of engagement between researchers and knowledge users, the aim of this work was to facilitate success by structuring implementation activities within an appropriate knowledge translation (KT) framework.; METHODS A working group of the International Society of Quality of Life Research (ISOQOL) led the development of the CONSORT PRO guidance. Twenty-nine stakeholders attended the CONSORT PRO 2012 consensus meeting in London, January 2012 including: journal editors, methodologists, clinical trialists, policy makers, clinicians, representatives of UK and US funding agencies, pharmaceutical industry, and patients and knowledge transfer experts. A key agenda item was elicitation of stakeholder perspectives on KT strategies. The conceptual framework developed by Graham and colleagues and adopted by the Canadian Institutes of Health Research (CIHR) was used as a basis for discussion.; RESULTS Potential barriers to the use of the CONSORT PRO 2012 statement may include ‘checklist fatigue’, a lack of awareness or use by authors or peer reviewers, challenges for editors in ensuring compliance with the new checklist, potential space restrictions in journals and lack of endorsement by funding agencies. The group recommended that uptake of the guidance may be improved through the development of web based education packages and tools to support authors, editors and reviewers. Endorsement by journal editors and funding bodies was identified as an essential facilitator of widespread implementation. Evaluation of improvements in PRO reporting was recognised as an important component of the KT framework.; CONCLUSIONS Developing guidelines alone is unlikely to lead to improved PRO reporting in trials in the absence of a comprehensive, co-ordinated and evaluable approach to KT. Given the many contexts in which the PRO guidance must be implemented, and the need for multiple approaches to facilitate both its implementation and its evaluation, the CIHR KT model with its emphasis on ‘‘collaborative problem solving through linkage and exchange’’ is well-suited to guide this initiative.
(2035) Developing 5 item banks to measure patient-reported physical functioning as part of the Spinal Cord Injury Functional Index (SCI-FI) measurement system David Tulsky, PhD, University of Michigan; Pamela A. Kisala, University of Michigan; Alan Jette, PhD, Boston University; Mary D. Slavin, Boston University; Pengsheng Ni, Boston University AIMS Recent advances in Patient Reported Outcomes (including the PROMIS and Neuro-QOL measurement systems) have included item banks measuring physical functioning. PROMIS included a single, unidimensional scale while Neuro-QOL included two scales measuring mobility and upper extremity functioning separately. Individuals with Spinal Cord Injury (SCI) report that there are a significant number of issues related to mobility and activities of daily living care that cannot be categorized within general factors of physical functioning. Even broad mobility and upper extremity may be too broad to accurately measure physical functioning in this population. This study uses qualitative analyses on focus group feedback as well as a quantitative confirmatory factor analyses approach to examine the factor
123
84 structure of physical functioning for individuals with SCI.; METHODS 69 individuals with traumatic SCI participated in focus groups and 855 adults with traumatic spinal cord injury participated in these studies. All participants were from 6 SCI Model Systems sites. In the calibration data collection, each participant responded to approximately 300 items thought to assess different dimensions of physical functioning. Qualitative analyses on focus group data used NVivo software. Confirmatory Factor Analyses testing alternate models were conducted using MPlus software to determine optimal factor structure of physical function in an SCI population.; RESULTS Several a priori models of physical functioning were tested. Fit statistics (CFI, TLI, and RMSEA) support many of the models. However, a 5-factor model (including Basic Mobility, Ambulation, Wheelchair Mobility, Self Care, and Fine Motor) provided not only strong support that the model fits the data but was also strongly advanced as important structure in the qualitative analyses.; CONCLUSIONS The 5-factor SCI-FI measure of physical function will more accurately assess physical functioning than existing measures that rely on a one- or two-factor model of physical functioning for an SCI population. Through Item Response Theory (IRT) analyses, each factor has been developed into a standalone computerized adaptive test (CAT); therefore, any participant can take a small set (e.g. 4–8) of items in each topic area, and receive a score in each individual area.
(2036) Health states utilities in chronic immune diseases Cla´udia C. Ribeiro, Universidade Cato´lica Portuguesa, Portugal; Augusta Silveira, University Fernando Pessoa, Portugal; Teresa Sequeira, Portugal; Isabel Santos, Hospital Santa Maria da Feira; Augusta Marques, Universidade Santiago de Compostela; Carlos Vasconcelos, Hospital Santo Antonio Centro Hospitalar do Porto AIMS Utility scores are used to estimate Quality Adjusted Life Years (QALYs), applied in determining the cost-effectiveness of health care interventions. In studies where no preference based measures are collected, indirect methods have been developed to estimate utilities from clinical instruments. The aim of this study was to evaluate a published method of estimating the Short Form-6D (SF-6D) (preference based) in patients with chronic immune diseases and evaluate the impact of socio demographics economics and clinical characteristics on quality of life (QoL) and potential predictors for QoL improvements.; METHODS We studied 320 patients with chronic immune diseases (103 men and 226 women with a mean age: 45.21; 84 people living with chronic immune diseases. All responders to the Portuguese SF-36 version 2.0 questionnaire can be assigned an SF-6D score provided the 11 items used in the SF-6D have been completed using a computerized administration. To assess socioeconomic status, we use the Graffar’ scale, clinical and demographic variables were assessed by a questionnaire specifically designed for the present study.; RESULTS The mean SF-6D utility value was .595. Male, gender, young, single, individuals with high educational attainment level and Graffar’s scale Class high reported higher utility levels. As expected, those who takes therapeutics’ or have a higher length disease reported lower mean utility levels than those who were in a less severe stadium of the disease or without therapeutic.; CONCLUSIONS The authors conclude that SF-6D is an effective tool for measuring HRQL from a populations leave with chronic immune diseases. The preference-based measures used in this study distinguish patient groups with chronic immune diseases’ in terms of sociodemographics characteristics and clinical groups. The normative values can be used economic evaluation and clinical studies as they incorporate patient’s preferences and translate the value attribute to patients’ health state.
123
Qual Life Res (2012) 21:1–132 (2037) Predicting SF-6D preference-based score based on the European Organization for Treatment and Research of Cancer Quality of Life Questionnaire in colorectal cancer Carlos King Ho Wong, Mphil, BSc, Department of Family Medicine and Primary Care, The University of Hong Kong, Hong Kong; Cindy L.K. Lam, MD, The University of Hong Kong, Peoples Republic of China; Donna Rowen, PhD, University of Sheffield, United Kingdom; Eric Yuk Fai Wan, The University of Hong Kong AIMS To develop a mapping model for estimating SF-6D preferencebased scores from the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Core module (QLQ-C30) and Colorectal-specific module (QLQ-CR38/CR29) in patients with colorectal Cancers (CRC), with and without adjustment for clinical and demographic characteristics.; METHODS Ordinary least squares regression model was applied to the model development using a cross-sectional data of 216 patients with CRC collected from outpatient specialist clinic of a regional hospital in Hong Kong. The items responses or scale scores of QLQ-C30 in conjunction with QLQ-CR38/CR29 data, selected demographic (sex and age) and clinical (cancer stage, treatment completion and presence of stoma) characteristics of patients were used to predict the SF-6D scores. The goodness-of-fit of models were examined by exploratory power (R-square and adjusted R-square), Akaike information (AIC) and Bayesian information criterion (BIC) whereas the predictive performances were evaluated by root mean square error (RMSE), mean absolute error (MAE) and Spearman’s correlation coefficients between predicted and observed SF-6D scores. The models were validated against an independent set of 56 patients with CRC to assess the predictive performance.; RESULTS Both scale and item response models obtained the variation more than 67 % explained in the SF-6D scores. The model using the main and interaction effects of the QLQC30 item scores in supplemented with QLQ-CR29 item responses plus clinical variable of stoma was chosen as the best-performing mapping model because of having the best fit with the highest R-square of 75.9 %, and good predictive performances with Spearman’s correlation coefficient of 0.871 and low predictive errors (RMSE: 0.079; MAE: 0.062). There was over-prediction for poorer health states at the observed SF-6D scores lying below 0.7 but underprediction was observed for better health states when SF-6D scores were valued at 0.7 or above.; CONCLUSIONS SF-6D preferencebased scores can be predicted using QLQ-C30 and QLQ-CR38/CR29 scores with satisfactory precision in patients with CRC for appraising clinical interventions in economic evaluations.
(2038) Adding categories to EuroQol-5D instrument to remove the ceiling effect Jorge S. Cortes, Sr., Universidad de la Frontera, Chile; Juan F. Blanco, Sr., Trauma Unit, Facultad de Medicina, Universidad de Salamanca, Espan˜a; Claudio R. Castro, Sr., Facultad de Estadı´stica e Informa´tica, Universidad Veracruzana, Me´xico; Rosa A. Sepu´lveda, Departamento de Estadı´stica, Universidad de Salamanca, Espan˜a; Sergio R. Mun˜oz, Sr., Departamento de Salud Pu´blica, Universidad de la Frontera, Chile, Chile; Norman Moreno-Garcia, Sr., Universidad Cato´lica de Temuco, Chile AIMS Show that the ceiling effect that occurs in the EuroQol-5D instrument in general or chronic population is the product of a design flaw of the Instrument.; METHODS Compare the observed ‘‘ Selfreported Overall health state’’ (EVA) in patients that have been operated trough integrated instrumental arthrodesis in contrast to The hypothetical behavior of the same segment of the Spanish population, using unbalanced factorial design, in which the experimental unit was redefined as the response of each patient on each of the five
Qual Life Res (2012) 21:1–132 dimensions of observed ‘‘overall Health’’. Under this scheme, the variables are: ‘‘OVERALL HEALTH’’ of a patient who responds to the i-th dimension, STATE, is the used measure (1: EVA 2: RATE), SEX of the patient; LEVEL indicates the level of response of a patient in a given dimension, DIMENSION, a variable that indicates the dimension in which a patient responds, GRUPOE, variable that accounts of the age group a patient belongs to.; RESULTS It was determined that the ‘‘overall Health’’ of these patients, does not behave the same way it would if the patients were people with the same characteristics but belonging to the Spanish general population. The patient age is a determining factor in the perception of ‘‘overall Health’’ for those who respond with the response level ‘‘1’’. Regarding to patients who respond with the level of response ‘‘2’’ to any dimension of the descriptive part of the Euroqol-5D, men have a better perception of their ‘‘overall Health’’ than women. Finally, analysis of patient responses to any dimension of the descriptive part of the Euroqol-5D, shows that not all dimensions equally affect the perception of patients regarding their ‘‘overall Health’’.; CONCLUSIONS The source of the problem is the design of the instrument, and therefore we developed a modification that partitions the category ‘‘No problems’’ for each of the 5 dimensions, in three new categories that conceptually are: Easily, moderately easy, and no problems as such.
(2039) Quality of life and violence against women Rosa Aurea Q. Fernandes, PhD, Universidade Guarulhos, Brazil; Mari Anna T. Lima, Guarulhos University, Brazil AIMS To assess the Quality of Life (QL) of women who suffer violence and to verify the association between QL and social demographical variables, type of aggression, the aggressor’s family relationship and frequency of the aggressions suffered; METHODS Descriptive exploratory study performed from May–August 2010, with participation of 70 women followed-up at an institution which assists women who suffer violence. The instruments used to evaluate QL was WHOQOL-BREF, and in the statistical analysis the t-Student, ANOVA and Bonferroni tests with significance levels of 5 % (p \ 0.05).; RESULTS The mean age of the women was 36 ± 9.8 years, 30.0 % had consensual relationships with partners, 34.3 % had less than 8 years of elementary education and 34.3 % had studied for 11 years; 44.3 % were Caucasian; 95.7 % had children and 64.3 % had jobs. The most frequent types of violence were associated physical and psychological (45.8 %) and psychological (38.6 %); the main aggressor was their intimate partner (70.0 %) and 62.9 % suffered aggression more than 11 times. The aggressor’s profile showed mean age of 39.9 ± 10.6 years; 47.0 % were Caucasian; 40.0 % had less than 8 years of elementary education; 60.0 % used alcohol or drugs and 82.9 % had jobs. The body parts mostly hit by aggressions were upper limbs (91.4 %), front of the head and back (71.4 %) and lower limbs (67.1 %). The women’s general QL was low, with an average score of 42.1 %. The area with highest score average was social relationships (51.9 %). There was a statistically significant association with lower QL levels of married women compared to separated women in the psychological domain (p = 0.038) and to single women in the social relationship domain (p = 0.027). Also noticed was QL of women who had children in the psychological domain (p = 0.042) and in the environment (p = 0.011), also those who suffered repeated violence in the psychological domain (p = 0.030) and in the general QL (p = 0.005), where there was a statistical difference for those who suffered physical violence (p = 0.007).; CONCLUSIONS Women who suffer violence have lower QL, sspecially those who suffer physical violence, are married, have children and repeatedly suffer violence.
85 (2041) Association of perceived health status and mental health in patients after percutaneous coronary intervention following hospital discharge Rejane K. Furuya, University of Sao Paulo at Ribeirao Preto College of Nursing; Vitor C. Richter, University of Sa˜o Paulo at Ribeira˜o Preto College of Nursing; Mariana Coelho, University of Sa˜o Paulo at Ribeira˜o Preto College of Nursing; Eliana C.A. Costa, University of Sa˜o Paulo at Ribeira˜o Preto College of Nursing; Andre´ Schmidt, School of Medicine of Ribeira˜o Preto, University of Sa˜o Paulo; Rosana Spadoti Dantas, PhD, University of Sa˜o Paulo, Brazil; Ma´rcia A. Ciol, University of Washington; Lidia A. Rossi, PhD, University of Sao Paulo, Brazil AIMS To assess the association between perceived health status and mental health of patients after percutaneous coronary intervention (PCI).; METHODS Cross-sectional study with 79 patients submitted to PCI who were under outpatient care from 2 to 7 months following hospital discharge. The data collection was carried out from July 2011 to April 2012. The 36-Item Short Form Health Survey (SF-36) was used to assess perceived health status and the Hospital Anxiety and Depression Scale (HADS) to assess mental health, both validated for Brazilian Portuguese. The HADS has two subscales: anxiety and depression. In each subscale, a person is classified as having the condition (scores 0–7) and not to have the condition (scores 8–21). We further classified the participants into three groups: without anxiety and depression (G1), with anxiety or with depression (G2), and with both (G3). To compare the means of the perceived health status measures we used ANOVA, with significance level set at 5 %.; RESULTS Forty patients (50.6 %) were men and submitted to ICP with the use of stents (82.3 %). Considering mental health, 41.8 % did not have anxiety or depression symptoms, 31.6 % had one of the two conditions, and 26.6 % had both. There were statistically significant differences for all domains of the SF-36 measures among the three groups. Means and the standard deviation of G1, G2, and G3 groups, followed by the p value of the ANOVA, were as follows: Physical Functioning: 75.9 (23.1), 54.8 (27.1), 50.0 (27.5), p = 0.001; Role Limitations due to Physical Problems: 59.0 (41.3), 20.4 (30.5), 32.1 (41.1), p = 0.001; Bodily Pain: 74.5 (27.0), 44.4 (20.5), 53.5 (24.8), p \ 0.001; General Health Perceptions: 76.0 (14.8), 57.3 (22.3), 51.6 (19.9), p \ 0.001; Vitality: 75.1 (19.4), 51.8 (23.1), 41.1 (26.1), p \ 0.001; Social Functioning: 89.7 (21.0), 65.5 (28.0), 47.0 (31.1), p \ 0.001; Role Limitations due to Emotional Problems: 70.7 (37.9), 34.6 (37.8), 44.4 (39.9), p = 0.002; General Mental Health: 77.4 (18.1), 52.1 (21.5), 43.4 (18.8), p \ 0.001.; CONCLUSIONS As expected, perceived health status and mental health were associated, with the group with no symptoms of either anxiety or depression reporting higher scores of perceived health status in all domains of the SF-36. This study was sponsored by FAPESP.
(2042) The impact of self-stigmatization on functioning in bipolar disorder Omer Aydemir, Celal Bayar University, Turkey; Burak Uykur, Celal Bayar Uni. Dept. of Psychiatry Manisa Turkey AIMS It has been shown that self-stigmatization causes low self-esteem and social anxiety in bipolar patients. In this study it is aimed to demonstrate the impact of self-stigmatization on functioning in bipolar disorder.; METHODS The study was carried out with 70 patients diagnosed with bipolar disorder. All patients were in remission and the remission state was confirmed by a 17-item Hamilton Depression Rating Scale score less than 7 and a Young Mania Rating Scale score less than 4. For the assessment of self-stigmatization, sense of stigmatization subscale of Bipolar Disorder Functioning Questionnaire (BDFQ-Stigma) was used. In the assessment of functioning,
123
86 Functioning Assessment Short Test (FAST) covering the domains such as autonomy, occupational functioning, cognitive functioning, financial issues, interpersonal relations, and leisure time was applied. In the statistical analysis, regression analysis was performed.; RESULTS The mean age of the bipolar patients was 40.7 and 55.7 % were female. Duration of the illness was 11.9 years with a mean of 6.2 episodes. In the regression analysis, R square was found to be 0.205 and the model was statistically significant (F = 3.225, p = 0.007). The only domain which is affected by self-stigmatization was found to be interpersonal relations (Beta = 0.888, t = 2.000, p \ 0.05).; CONCLUSIONS It is concluded that self-stigmatization has negative impact on interpersonal relations by causing social anxiety as demonstrated previously. The other domains of FAST including occupational functioning were not affected by the self-stigmatization.
(2043) Psychosocial correlates in visually impaired people: a gender based analysis Selwyn Stanley, PhD, University of Plymouth, United Kingdom; Velliah Sethuramalingam, PhD MA, Bharathioasan University, India AIMS The experience of disability by itself is said to be accompanied by a range of psychosocial problems such as poor self-esteem and higher levels of depression, pessimism and feelings of inferiority. Proponents of the feminist theory hold that women in particular experience the twin disadvantage of gender as well as their disability and could hence be expected to undergo more adverse life experiences than their male counterparts. This study compared male and female visually impaired adult respondents from two non-governmental organisations in Tiruchirappalli, South India.; METHODS Thirty visually impaired women were compared to 55 visually impaired men with regard to their quality of life, pessimism, depression, anxiety and stress scores. The WHOQOL Bref scale, Hopelessness scale (Back & Wiseman, 1974) and the DAS scale (Lovibond & Lovibond, 1995) were the instruments administered.; RESULTS t tests reveal no significant statistical differences between the two groups on all domains studied except for the ‘feelings about the future’ component of the pessimism scale. Correlational analysis shows that quality of life scores were significantly correlated with depression, anxiety, stress and pessimism scores for the women but not for visually impaired male subjects.; CONCLUSIONS The findings of the study indicate that there are no significant differences in the life experience of visually impaired adults based on gender. Irrespective of gender, therapeutic intervention is merited for visually impaired people who experience psychosocial difficulties.
(2044) Completion time reflecting the quality of outcomes across multiple questionnaires Ma Yuan, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Chen Zhihui, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yu Changhe, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Lv Meijun, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS It might change the quality of outcomes while multiple questionnaires were filled out at the same time. The aim of this manuscript was to explore whether the completion time would reflect the changes of the quality of outcomes.; METHODS In this research, 3 questionnaires in the sequence of Short Form (SF-12), Traditional Chinese
123
Qual Life Res (2012) 21:1–132 Medicine Syndrome of Angina Questionnaire-Qi Deficiency and Blood Stasis Syndrome pattern (TCMSAQ-QBS) and Seattle Angina Questionnaire (SAQ) were filled out. There was one common dimension across the 3 questionnaires, it was called the physical component summary in SF-12, the physical limitation in TCMSAQQBS and SAQ. One item from this dimension was picked up, it was described as ‘‘climbing several flights of stairs’’ in SF-12, ‘‘daily physical activities, such as climbing one ladder or walking 500 miles’’ in TCMSAQ-QBS, ‘‘climbing a hill or a flight of stairs without stopping’’ in SAQ. The outcome was dealt so that it was comparable and then the Spearman correlation analysis was adopted to test the consistency of this item. The completion time of the questionnaires was assessed to test whether respondents had worse consistency with fewer time and the better consistency with more time.; RESULTS 211 angina patients (male 71, age 59.50 ± 9.70) were included, the mean and SD of the item score were 0.51 ± 0.29 (SF-12), 0.67 ± 0.16 (TCMSAQQBS), 0.59 ± 0.17 (SAQ). The Spearman correlation coefficient (CC) values of the item score were 0.436 (SF-12, TCMSAQ-QBS), 0.458 (SF-12,SAQ) and 0.464 (TCMSAQ-QBS,SAQ), the CC values of the item score those completion time of the questionnaires more than the mean time (0.46 min) were 0.560(SF-12, TCMSAQ-QBS), 0.541 (SF-12,SAQ) and 0.526 (TCMSAQ-QBS,SAQ), less than the mean time were 0.360 (SF-12, TCMSAQ-QBS), 0.412 (SF-12,SAQ), 0.457 (TCMSAQ-QBS,SAQ), revealing a tendency: the consistency among participants using much more time better than those using less time.; CONCLUSIONS While multiple questionnaires were filled out at the same time, the time spent on the questionnaires reflected the quality of the outcomes should be considered.
(2045) Integrating disability with sickness leave days for return-to-work assessment in traumatic limb injuries Wen-Hsuan Hou, E-Da Hospital & I-Shou University, Taiwan; Hung-Yi Chuang, Department of Public Health, Kaohsiung Medical University; Jing-Shiang Hwang, Institute of Statistical Science, Academia Sinica AIMS To purpose a new indicator named disability-adjusted leave days (DALDs) that integrates the sick leave days before return-towork (RTW) and the disability level.; METHODS Totally 1,167 limb injuries were consecutively recruited in a teaching hospital from January to December 2009. The number of leave days (LDs) was calculated between the date of injury and the first episode of RTW. We integrated the product of proportion of not-RTW function and daily disability index (DI) function over time to calculate the expected DALDs. We also compared the expected DALDs between two subgroups according to various characteristics.; RESULTS Women and elder workers tended to have longer LDs than their counterparts, with an increase in percentage change of 16.0 and 139.5 %, respectively. After adjusting for DI, the corresponding figures for DALDs were both increased at 28.7 and 186.58 %. The corresponding figures for the less educated workers and blue collar workers were 185.7 and 155.8 %. The expected DALDs all revealed significant differences in all subgroup analyses.; CONCLUSIONS We concluded that the proposed DALD can be of potentially great usefulness for the clinical decision and industrial policy-making with respect to the assessment of the importance of limb injury due to worker’s sick-leaves.
(2046) Cohen’s kappa assesses reliability, not agreement Henrica C.W. de Vet, Sr., PhD, EMGO Institute VU Medical Center, Netherlands; Caroline B. Terwee, PhD, VU University Medical Center, Department of Epidemiology and Biostatistics, Netherlands;
Qual Life Res (2012) 21:1–132 Lidwine B. Mokkink, VU University Medical Center, Netherlands; Dirk L. Knol, Department of Epidemiology and Biostatistics, VU University medical center AIMS Kappa was presented by Cohen (1960) as a parameter to assess agreement between repeated measurements. Kappa is commonly used by clinicians to assess inter- and intra-observer agreement between categorical outcomes, e.g. diagnostic tests. There are situations where the observed agreement is quite high, while Cohen’s kappa is low. This has led to many extensions of Cohen’s kappa. We aim to interpret Cohen’s kappa, considering the questions that clinicians want to answer.; METHODS We present examples of kappa calculations with different prevalences of abnormalities, and discuss the interpretation of the kappa values considering the distinction between agreement and reliability. Agreement is expressed in the degree to which cases are classified in the same category (i.e. observed agreement). Reliability parameters assess how well patients can be distinguished from each other, despite misclassifications. They relate the misclassification or measurement error to the variability between patients. In a heterogeneous population, i.e. large variability between patients, it is much easier to distinguish patients from each other. RESULTS We argue that kappa is a reliability parameter. The value of kappa equals an intraclass correlation coefficient (ICC), which is clearly a parameter of reliability. Our examples show that Cohen’s kappa is dependent on the prevalence of the abnormalities. A population with a high or low prevalence represents a homogeneous population. The value of kappa is then lower, given the same amount of observed agreement, than in a heterogeneous population. In the latter situation, patients are easier to distinguish. Clinicians are interested in agreement because they want to know whether another clinician would come to the same diagnosis in a particular patient. They are not interested in distinguishing patients from each other.; CONCLUSIONS Clinicians should not calculate kappa when interested in the agreement between repeated measurements, but rather base their conclusion solely on the observed agreement.
(2047) Changes in health-related quality of life in patients with heart failure Antonio Escobar, PhD MD, Hospital Universitario Basurto, Spain; Esther Rekalde, Hospital Mendaro; Maria L. Fernandez-Soto, Hospital Universitario Donostia; Marta Gonzalez-Saenz de Tejada, Hospital Universitario Basurto; Iratxe Lafuente, Hospital GAldakao-Usansolo AIMS In Spain, Health-Related Quality of Life (HRQoL) in patients with heart failure (HF) has not been measured regularly. The objective was to evaluate changes in HRQoL at 6 months post discharge, and the influence of hospital readmissions and emergency rooms visits (re-admitted patients); METHODS prospective study of patients hospitalized by HF. Questionnaires were filled out pre-discharge and 6 months post. HRQoL was measured by Minnesota Living with Heart Failure (MLWHF) and SF-12. MLWHF has 21 items, from 0 to 5, with a total score from 0 to 105 (best to worst). It has two dimensions, physical (score 0–40) and emotional (score 0–25). SF-12 has two dimensions, Physical Component Summary (PCS) and Mental Component Summary (MCS). It ranges 0–100(worst to best). Analysis was performed by mean of paired t test between baseline and 6 months. Differences in changes between readmitted and no readmitted were performed by Student’s t test.; RESULTS There were 421 patients. 247 (58.7 %) men, mean (SD) age was 75.9 (10.5). There were 175 patients readmitted (41.6 %). In the sample, at 6 months there was improvement in MLWHF global score (52.6–41.8; p \ 0.001), physical domain (25.1–19.4; p \ 0.001) and the emotional domain
87 (10.8–9.8; p = 0.012). In SF-12 there were no improvement at 6 months in PCS (31.2–32.0: p = 0.2) but yes in MCS (43.6–45.6: p = 0.01). Regarding patients readmitted and those no readmitted, there were differences (p \ 0.001) in HRQoL in all MLWHF dimensions. In global score, improvement in no readmitted patients was 18.0 versus 1.2. In physical domain there was improvement of 8.7 vs 1.7. In the emotional subscale patients readmitted had a loss of -0.8 vs a gain of 2.5 points. SF-12 showed differences in the improvements of MCS; 3.4 vs 0.02 (p = 0.04) but not in the PCS; 1.4 vs -0.1.; CONCLUSIONS patients experience an improvement at 6 months, measured by a specific and a generic questionnaire, except in the PCS. However, patients who have been readmitted have minor improvements including losses in their HRQL, again except in the PCS.
(2048) Development and preliminary validation of the system of quality of life instruments for chronic diseases (QLICD) in China Chonghua Wan, PhD, Guangdong Medical College, Peoples Republic of China; Zheng Yang, Guangdong Medical College, Peoples Republic of China; Xiaomei Li, Kunming Medical University; Xiaoqing Zhang, Kunming Medical University; Chuanzhi Xu, Kunming Medical University; Zhihuan Zhao, The first affiliated hospital of Kunming Medical University; Wu Li, The first affiliated hospital of Kunming Medical University; Zhaoping Lu, The first affiliated hospital of Kunming Medical University AIMS To develop a Chinese QOL measure system called QLICD (Quality of Life Instruments for Chronic Diseases) based on the modular approach, considering that the Chinese cultures are not fully captured in existed scales. It is expected to include more than 30 disease-specific instruments.; METHODS The QLICD was developed based on WHO’s definition of QOL and programmed decision procedures including nominal group, focus group discussion, in-depth interview, pilot test and field test. The psychometric properties of the scale were evaluated with respect to validity, reliability and responsiveness employing correlation analyses, multi-trait scaling analysis, t tests and structural equation modeling etc. based on field tests data from 1,093 patients of hypertension (157), coronary heart disease (133), chronic gastritis (124), peptic ulcer (120), irritable bowel syndrome (99), COPD (114), chronic pulmonary heart disease (120), bronchial asthma (67) and diabetes mellitus (159).; RESULTS The QLICD (V1.0) includes a 30-items general module QLICD-GM (3 domains and 10 facets) and 9 specific modules with the numbers of items ranging from 14 to 21, and thus forming 9 specific scales of hypertension (QLICD-HY), coronary heart disease (QLICD-CHD), Chronic gastritis(QLICD-CG), peptic ulcer (QLICD-PU), irritable bowel syndrome (QLICD-IBS), chronic obstructive lung disease (QLICD-COPD), Chronic pulmonary heart diseases (QLICD-CPHD), bronchial asthma (QLICD-BA) and diabetes mellitus (QLICD-DM). All QLICD scales have a very clear hierarchical structure (items, facets, domains, overall) and can present mean scores at different levels (e.g. domains, facets). Correlational and factor analysis/structural equation model indicated good construct validity for all available instruments. Good criterion-related validity coefficients were confirmed given SF-36 as the criterions. Test–retest reliability coefficients for most domains of most instruments were higher than 0.80, and the internal consistency alpha were higher than 0.70. Most domains and overall scores of most instruments had statistically significant changes after treatments with moderate or higher effect size, SRM. Additionally, other 20 scales including rheumatoid arthritis (QLICD-RA), stroke (QLICD-ST), HIV/AIDS (QLICD-HIV), drug addiction (QLICD-DA), etc. are developing or validating.; CONCLUSIONS The all scales of the system of QLICD have good validity,
123
88 reliability and responsiveness and can be used as QOL instruments for patients with chronic diseases in China.
(2049) Items analysis on general module of the system of quality of life instruments for chronic diseases (QLICD-GM) by item response theory and classical test theory Zheng Yang, Guangdong Medical College, Peoples Republic of China; Chonghua Wan, PhD, Guangdong Medical College, Peoples Republic of China; Xiaomei Li, Kunming Medical University; Chuanzhi Xu, Kunming Medical University; Zhihuan Zhao, The first affiliated hospital of Kunming Medical University; Liping Duan, The first affiliated hospital of Kunming Medical University; Hong Li, The first affiliated hospital of Kunming Medical University AIMS Quality of life is now concerned worldwide, but few specific QOL instruments are available in China. By combining a general module and disease-specific modules, we have developed the Chinese QOL measure system called QLICD (Quality of Life Instruments for Chronic Diseases), which includes a general module (QLICD-GM) and more than 30 specific modules for different diseases such as hypertension (QLICD-HY). This research is aimed to analyze items of the QLICD-GM by item response theory and classical test theory considering QLICD-GM is of greatest importance since it serves as the core of the system.; METHODS 1,024 inpatients from 8 different chronic diseases (hypertension 155, coronary heart disease 133, chronic gastritis 124, peptic ulcer 120, irritable bowel syndrome 99, chronic obstructive lung disease 114, chronic pulmonary heart disease 120, diabetes mellitus 159) were measured by QLICD-GM. With the software MULTILOG, Samejima’s Graded Response Model was used to calculate parameters of discrimination and difficulty, information function, and to present item characteristic curve. Also the parameters of discrimination and difficulty of items were analyzed based on classical test theory.; RESULTS Based on item response theory, the discrimination parameters of all 30 items were between 1.17 and 1.88, difficulty parameters of each item were increasing monotonically for all items (ranging from -2.88 to 2.27). 24 items were selected as good ones by average information (ranging from 0.37 to 0.99) and items characteristics. The analysis of classical test theory indicated that difficulty coefficients of the QLICD-GM were moderate with ranging from 0.42 to 0.67 for physical domain, from 0.60 to 0.83 for psychological domain and from 0.52 to 0.80 for social domain. Indices of discrimination of the 30 items were at good levels with average being 0.46, 0.72 and 0.47 for three domains respectively, except of three items PH2 ‘‘‘Have you felt fatigue easily? ‘‘(0.20), PH6 ‘‘Have you had a good appetite?’’ (0.26) and PH7 ‘‘Were you satisfied with your sleep?’’ (0.28).; CONCLUSIONS Items analysis based on item response theory can select good items with higher information and refined characteristics to overcome weakness of classical test theory CTT.
(2050) Religiosity and health related quality of life in elderly people Gina A. Abdala, MPH, Brazil Adventist University, Brazil; Miako Kimura, PhD, University of Sao Paulo, School of Nursing, Brazil; Yeda Aparecida De Oliveira Duarte, PhD, University of Sao Paulo-Brazil, Brazil; Maria Lu´cia Lebra˜o, University of Sao Paulo, Brazil; Lucas P. Damiani, University of Sao Paulo AIMS To analyze association between religiosity and health-related quality of life (HRQoL) of elderly people in Sao Paulo/Brazil.; METHODS Cross sectional study, with quantitative approach of the Health, Well Being and Aging (SABE) Study. Those who had
123
Qual Life Res (2012) 21:1–132 cognitive impairment were excluded, and, thus, the final probabilistic sample was 844 respondents (=60 years). The Physical Component (PC) and Mental Component (MC) were scored by the Short Form Health Survey (SF12) using the median score. For religious survey, participants answered seven questions about religiosity (organizational, non-organizational and intrinsic religiosity questions). The Cronbach’s Alpha for the standard questionnaries were: 0.85 (Physical component), 0.80 (Mental component) and 0.91 (Intrisic religiosity), and, it was used Logistic Regression for analysis according to the gender (p = 0.05 and CI = 95 %).; RESULTS Controling the sociodemographic and health conditions variables, there is a significant association (p = 0.05, CI = 95 %) between intrinsic religiosity and quality of life for elderly people. When analyzed by sex, elderly (women) who considered religion important in their lives were 9.55 more likely to be among the ones with good physical health (CI = 1.5–60.3) compared to those who do not consider religion important. Elderly (women) who thought that religion gives strength to cope life’s difficulties were 5.5 more likely to be among the group with good mental health. Elderly (men) who considered themselves ‘‘religious’’ had 2.47 more chance of being of the group with good physical health, compared with those who did not consider themselves religious. Elderly (men) who thought that religion helps them understand the difficulties of life were 3.98 more likely to be among the group with good mental health, compared to the elderly that religion did not help understand the difficulties of life (CI = 1.13–13.96).; CONCLUSIONS Intrinsic religiosity is associated with Physical and Mental components of HRQoL, for both male and female elderly of Sao Paulo, what means, giving importance to religion, believing that religion gives strength and understanding to cope with difficulties of life help the elderly to have a better HRQoL.
(2051) Quality of life is a process not an outcome Leah M. McClimans, PhD, University of South Carolina, United States; John Browne, PhD, University College Cork, Ireland AIMS Healthcare providers are increasingly evaluated using outcome measures. This is a radical departure from the traditional focus on process measures. More radical still is the proposal to move from relatively simple and proximal measures of outcome such as mortality, to complex outcomes such as quality of life. Examples include the ‘Patient-Reported Outcome Measures’ (PROMS) program in the English NHS and the proposal to include patient-reported measures in the ‘Pay for Performance’ program for Medicare in the United States. While the practical, scientific and ethical issues associated with the use of mortality and morbidity to compare healthcare providers have been well rehearsed, the specific concerns associated with the use of quality of life measures in quality of care research have received little attention to date. As with much research on quality of life there is a tendency to assume that the disadvantages are outweighed by the general virtue of ‘listening’ to patients. In this paper we disagree with this assumption and argue that quality of life is a process, not an outcome.; METHODS Our argument is threepronged. First, we suggest that the use of outcome measures presupposes a consequentialist outlook in the evaluation of quality health care. Second, the methodological challenges of measuring quality of life beg the question as to whether or not they can capture patients’ perspectives. Third, changes in quality of life are such that they are usually distal to any associated healthcare intervention rendering quality of life an unreliable indication of quality of care.; RESULTS Our arguments suggest that many of these problems come from the attempt to generate outcomes measures from constructs of ‘quality of life’.; CONCLUSIONS Following up on the results from our arguments we consider the methodological and political consequences of construing quality of life as a process instead of an outcome.
Qual Life Res (2012) 21:1–132 (2052) Pre-testing the EORTC QLQ-ELD15 in patients aged 70 years and above with a haematological malignancy Colin Johnson, Mchir, Dept of Surgery, United Kingdom; Sally J. Wheelwright, University of Southampton, United Kingdom; Deborah Fitzsimmons, Swansea University, United Kingdom; Teresa Young, BSc, Mount Vernon Hospital, United Kingdom AIMS The EORTC QLQ-ELD15 has been developed to supplement the EORTC QLQ-C30 in order to measure health-related quality of life (QOL) in patients aged 70 years and above across a wide range of solid cancer sites and treatments. The aim of this study was to assess the content and acceptability of the QLQ-ELD15 in older patients with haematological malignancies.; METHODS The EORTC guidelines for module development were used to undertake a modified phase 3 (pre-testing) phase. Thirty-one patients with a range of haematological malignancies across 3 centres in the UK and the Netherlands were recruited. Patients completed the EORTC QLQ-C30 and QLQELD15. A debriefing interview was used to check for any redundant items or significant omissions. After scoring each item in the ELD15, a sub-set of 14 patients rated the importance and relevance of each item. Responses were compared with the previously collected data from patients with solid cancers.; RESULTS The median age of patients was 76 (range 70–86) years. 65 % of patients were aged 70–79 years; 35 % over 80 years. All 15 items met accepted decisions for inclusion criteria (mean score of [1.5, no obvious floor ceiling effect with dispersion of responses over at least 2 scores, [50 % prevalence) and were rated as quite or very important by at least 79 % of patients. Patients scored 9 items as meeting the criteria for relevance ([60 %), with 5 items slightly below (57 %) and 1 item (worries about end of life) scoring 50 %. Two patients raised additional issues across the main sample. One patient identified 1 item (joint stiffness) as irrelevant. The pattern of responses in patients with haematological malignancies were very similar to those obtained in the solid tumour population.; CONCLUSIONS Responses to the QLQ-ELD15 were similar in patients with haematological and solid malignancies. Although some questions were considered only moderately relevant by the haematological patients, these items met all other inclusion criteria. After discussion with the module team, these items were retained for investigation in a larger sample. The large-scale international validation study of the EORTC QLQ-C30 and QLQ-ELD15 can now be undertaken in patients with haematological malignancies.
89 symptoms (intrusion: p = 0.803, avoidance: p = 0.161, hyperarousal: p = 0.193) and physical quality of life (p = 0.775).; CONCLUSIONS The findings indicate that less resilient women need more psychological support and interventions to cope with their pregnancy complication. A screening for resilience, depression and quality of life may be appropriate to identify these women.
(2054) Scale-specific minimal important differences of the thyroid-specific patient-reported outcome ThyPRO Torquil Watt, MD PhD, Copenhagen University Hospital, Denmark; Jakob Bjorner, PhD, National Research Centre for the Working Environment, Denmark; Steen J. Bonnema, Odense University ˚ se K. Hospital; Laszlo Hegedu¨s, Odense University Hospital; A Rasmussen, Copenhagen University Hospital Rigshospitalet, Denmark; Ulla Feldt-Rasmussen, Copenhagen University Hospital; Mogens Groenvold, MD, PhD, University of Copenhagen, Denmark AIMS To establish scale specific minimal important difference (MID) of the newly developed thyroid-specific patient-reported outcome ThyPRO, using scale-specific change items.; METHODS ThyPRO measures quality of life (QoL) in patients with thyroid diseases in 13 multi-item scales ranging 0–100 (higher score indicating worse QoL). A sample of 386 patients with benign thyroid disease (non-toxic goitre (n = 127), toxic nodular goitre (n = 75), Graves’ hyperthyroidism (n = 86), Graves’ orbitopathy (n = 16) and autoimmune hypothyroidism (n = 82) completed the ThyPRO at baseline and 6 weeks after institution of clinically relevant treatment of thyroid disease (e.g. medical, surgical or radioactive treatment of goitre, hyper- or hypofunctioning or eye affection). Clinically it was expected, that after 6 weeks of treatment, sufficient number of patients would have experienced a small change in their disease level.. Thus, at 6 weeks follow-up, 13 scale-specific change items were administered immediately after each scale, evaluating change on a 7-point scale
Table 1 Mean change (with 95 % confidence interval (CI) and number of patients in each category) in scale scores (0–100) for each ThyPRO scale for the groups reporting being ‘‘A little better’’ and ‘‘A little worse’’, adjusted for the change found in the group reporting ‘‘No change’’ Scale
Adjusted mean change in ‘‘A little better’’ group mean (95 % CI)/n
Adjusted mean change in ‘‘A little worse’’ group mean (95 % CI)/n
(2053) Pregnancy complications, resilience and quality of life
Goitre symptoms
-6.3 (-11.2 to -1.3)/52*
Eva Mautner, Medical University Graz, Austria; Christina Stern, Medical University Graz; Maria Deutsch, Medical University Graz; Eva Elfriede Greimel, PhD, Medical University Graz, Austria; Uwe Lang, Medical University Graz; Mila Cervar-Zivkovic, Medical University Graz
Hyperthyroid symptoms
-2.8 (-7.6 to 2.0)/55
5.3 (-3.3–14.0)/14
Hypothyroid symptoms
-2.7 (-8.5 to 3.2)/36
10.1 (3.3–16.9)/26**
AIMS Resilience is defined as the ability to successfully withstand a threatening or challenging situation. The aim of this study was to examine differences in resilience on post-traumatic stress symptoms, depression and quality of life in a population of women with previous hypertensive pregnancy disorders. We hypothesized that resilience might be a protective variable, which shields against psychological distress.; METHODS The study was cross-sectional. We used international validated questionnaires to measure quality of life (SF-12), depression (EPDS), post-traumatic-stress symptoms (IES-R) and resilience (RS-13). 100 women were asked to participate. 67 returned the questionnaires. Statistical analyses were performed using ANOVA and Chi-square test.; RESULTS Women with high resilience showed significantly less depression (p = 0.001) and better mental quality of life (p = 0.002) compared to women with low resilience. There were no differences in post-traumatic stress
Eye symptoms Tiredness
12.2 (2.8–21.6)/12*
-9.6 (-13.9 to -5.2)/32***
7.3 (2.8–11.8)/30**
-15.1 (-20.6 to -9.5)/67***
7.7 (0.5–14.8)/34*
Cognitive impairment
-6.2 (-11.8 to -0.6)/48*
4.4 (-2.1–10.9)/34
Anxiety
-7.3 (-13.7 to -1.0)/42*
10.5 (3.5–17.4)/34**
Depressivity
-13.8 (-18.8 to -8.8)/64***
12.6 (6.7–18.5)/42***
Emotional susceptibility
-14.8 (-20.1 to -9.5)/55***
7.3 (1.6–12.9)/47*
Impaired daily life
-19.6 (-25.4 to -13.7)/48*** 10.9 (4.4–17.5)/36**
Impaired social life
-6.4 (-11.8 to -1.1)/34*
Impaired sex life
-8.4 (-17.8 to 1.0)/26
19.7 (8.6–30.8)/18***
Cosmetic complaints
-1.9 (-7.8 to 4.0)/25
5.9 (0.6–11.2)/32*
8.9 (2.3–15.6)/21**
These adjusted change scores represent the scale specific MIDs for improvement and worsening, respectively * p \ 0.05; ** p \ 0.01; *** p \ 0.001
123
90 (no change, small, moderate and large worsening or improvement, respectively). For each scale, the mean change in the group reporting a small improvement or worsening in the construct measured by that scale was estimated, controlled for the change in the group reporting no change, using linear regression. The resulting adjusted mean changes were interpreted as the minimally important difference for improvement and worsening respectively, for the particular scale.; RESULTS As seen in Table 1, these anchor-based, scale-specific MID range from 2.7 to 19.6 for improvements and 5.3–19.7 for worsening on these 0–100 scales, with a mean MID of 8.8 (improvement) and 9.4 (worsening).; CONCLUSIONS Minimally important differences for the ThyPRO scales were established using anchor-based methods. These can be applied in power calculations and for interpretation of clinical data. The size of the MIDs varied moderately among scales, but overall the MID levels found here are similar to those found for other questionnaires and the mean is about ‘half a standard deviation’, as often found. Future analyses could apply a combination of anchorand distribution-based methods, to evaluate particularly some of the very small and statistically non-significant scale-specific MIDs.
(2055) Comprehensive symptom profile in patients with breast cancer: practicability and sensitivity of the new symptom assessment tool CSP-Br Tatiana I. Ionova, Multinational Center for Quality of Life Research, Russian Federation; Denis A. Fedorenko, PhD, Nat’l Pirogov Med Surgical Ctr, Russia; Ivan V. Rykov, Regional Clinical Hospital, Saint-Petersburg, Russia; Tatiana Nikitina, PhD, Multinational Ctr for QoL Research, Russia; Kira A. Kurbatova, Multinational Center for Quality of Life Research, Russian Federation; Gary I. Gorodokin, New Jersey Center for Quality of Life and Health Outcomes Research, NJ, USA; Andrei Novik, MD PhD, National Medical Surgical Center, Russia AIMS Comprehensive symptom assessment and monitoring before and during treatment as well as at follow-up is worthwhile in breast cancer patients. Recently a new tool, Comprehensive Symptom Profile in Breast Cancer patients (CSP-Br), has been developed to assess symptoms specific to breast cancer patients. We aimed to test practicability and sensitivity of CSP-Br in breast cancer patients.; METHODS A total of 106 breast cancer patients (Stage—I-IV) were included in the study. Mean age was 52.8 years old (range 28–74). Patients filled out the CSP-Br before treatment and during chemotherapy (CT). The CSP-Br is a self-reported tool which allows the assessment of the severity of 57 symptoms in breast cancer patients: 47 core symptoms, 10 additional questions for patients who have/have not performed surgery. It consists of numerical rating scales, scored from ‘‘0’’ (no symptom) to ‘‘10’’ (most expressed symptom). Wilcoxon’s matched pairs test was used to compare symptom severity at different time-points. The prevalence of moderate-to-severe symptoms was studied using Chi-square test or Fisher’s exact test.; RESULTS Practicability of the CSP-Br was demonstrated by means of interviewing both patients and physicians. The patients acknowledged the comprehensiveness of the tool. All of the items were easy for the patients to read and understand. Completion of the CSP-Br in paper and pencil format took 5–7 min; the proportion of missing values was less than 2 % for all questions; the questionnaire found high acceptance reflected by no refusals. Usefulness of the CSP-Br to distinguish patients in terms of severity and number of symptoms experienced was demonstrated. Clear changes in symptom profile and severity were observed 1 week after CT cycle completion. The severity of 21 out of 47 symptoms increased significantly (p \ 0.05). Changes in symptom profile and severity during and after CT were used by the oncologists for the decision-making.; CONCLUSIONS The CSP-Br is a sensitive and practical tool to assess symptom profile
123
Qual Life Res (2012) 21:1–132 and severity in breast cancer patients. Comprehensive symptom monitoring is recommended to clearly determine treatment outcomes in breast cancer patients. Further studies are needed before the widespread use of the CSP-Br in clinical practice and clinical trials.
(2056) Diminished quality of life among U.S. student pharmacists: how self-perception toward pharmacy study plays role? Nalin Payakachat, BPharm, MS, PhD, Univ of Arkansas for Med Sci; Sarah E. Norman, Univ of Arkansas for Medical Sciences, United States; Anne Pace, Univ of Arkansas for Medical Sciences College of Pharmacy; Schwanda K. Flowers, Univ of Arkansas for Medical Sciences College of Pharmacy; Denise Ragland, Univ of Arkansas for Medical Sciences College of Pharmacy; Paul O. Gubbins, Univ of Arkansas for Medical Sciences College of Pharmacy AIMS This study assessed quality of life (QOL) of student pharmacists in a U.S. public college of pharmacy. We also explored factors that might contribute to lower QOL outcomes among the students.; METHODS This was a qualitative and quantitative study. Semistructured focus group interviews were conducted to explore in-depth perceptions of QOL in student pharmacists who had received a D or F grade in any professional year. Content analysis was performed. Subsequently, a survey was developed and administered to all student pharmacists (first to fourth year students: P1-P4) in the college. The survey included attitudes and academic-related self-perception (40 questions with 12 domains), health profile (SF-12 version 2.0), and personal information. We evaluated differences of the SF-6D scores among each of the four classes of a 4 year professional program using analysis of covariance, controlled for age, gender, and receipt of D or F grade. Stepwise backward regression was used to explore factors related to QOL outcomes. SF-12 physical and mental component scores (PCS and MCS) were also reported.; RESULTS In the qualitative phase, 3 focus groups (n = 6) were conducted and revealed that students felt their QOL outcomes were dramatically diminished during their professional education. Themes extracted from content analysis included sleep deprivation, stress, depression, relational, and social limitation (students described how their lives became restricted from families and friends). In the quantitative phase, 299 usable surveys analyzed (68 % response rate). Average age of the students was 24.9 ± 3.4 years. 65 % were female and the majority was Caucasian. The average SF-6D was 0.75 ± 0.12. The average SF-6D scores were significantly higher for P4 students (0.81 ± 0.13) than for P1-P3 (0.72 ± 0.11, 0.76 ± 0.12, and 0.71 ± 0.11; p \ 0.05). P1 students had the lowest SF-12 MCS, followed by P3, P2, and P4, while there was no difference of SF-12 PCS. Vitality, social functioning, role limitation due to emotional, and mental health for P1-P3 students were significantly lower than P4 students (p \ 0.05). Self-perceptions toward their academic effort and performance including scale lack of regulation (level of confusion about how he/she should study), selfenhancing ego-orientation (higher scores present higher levels of desire to be judged able), and self-defeating ego orientation (higher scores present higher levels of concern not to be negatively perceived by others) negatively impacted the students’ SF12-MCS, while selfperceived academic competence (higher scores present higher levels of confidence in his/her ability to cope with course load) increased their SF12-MCS.; CONCLUSIONS P1-P3 student pharmacists have lower QOL when compared to general population and their QOL became normal when nearing graduation (P4). However, the aspects of lack of sleep, relationship with families and friends were not captured by the SF-12. Students who have higher self-perceived academic competence are more likely to cope with stress and emotional related to pharmacy study. Self-perception toward academic effort and performance plays an important role on QOL of student pharmacists.
Qual Life Res (2012) 21:1–132 (2057) Construct validity of two commonly used functional medicine instruments: Five Facet Mindfulness Questionnaire (FFMQ) and Medical Symptoms/Toxicity Questionnaire (MSQ) Duska M. Franic, PharmD, PhD, Univ of Georgia, College of Pharmacy; Cathy Snapp, Tallahassee Memorial Healthcare Foundation; Ruth DeBusk, Tallahassee Memorial Healthcare Foundation; Narayanan Krishnamoorthy, Tallahassee Memorial Healthcare Foundation AIMS Increasing evidence shows that greater mindfulness results in improved patient outcomes including greater physical and emotional quality of life. However, no peer reviewed publications have reported if and to what extent the Five Facet Mindfulness Questionnaire (FFMQ) and Medical Symptoms/Toxicity Questionnaire (MSQ), two widely used scales in functional medicine, are associated with two validated and commonly used quality of life instruments, Short Form12 (SF-12) and EuroQol’s EQ5D (Index and Visual Analogue Scale, VAS). This study aims to fill this gap.; METHODS As part of a larger study, adult clinic patients with chronic conditions (asthma, diabetes, congestive ehart failure or chronic obstructive pulmonary disease) were eligible to participate at Tallahassee Memorial HealthCare, USA. Patients completed a battery of patient reported outcome measures including FFMQ, MSQ, EQ5D Index, EQ5D VAS, and SF-12, prior to their physician visit on site. Pearson r correlations were used to assess all associations between all functional medicine scale variables and quality of life measures.; RESULTS Useful survey responses were available from 39 clinic patients (mean age 62 years, SEM = 1.89, 73 % female) of low socioeconomic status with numerous comorbidities. All of FFMQ dimensions correlated significantly to relevant SF-12 dimension scales; e.g., FFMQ’s describing correlated with 6 of 8 SF-12 dimension scores including physical, role, social functioning, mental health and bodily pain; FFMQ’s nonreactivity correlated with 5 of 8 SF-12 dimensions (p \ .05). While MSQ was notable for the correlation of joints and digestive tract dimensions with 7 of 8 and 6 of 8 SF-12 dimensions (p \ .05). MSQ’s grand total correlated with all SF-12 dimension and summary scores and EQ5D-VAS and EQ5D-Index (p \ .05).; CONCLUSIONS There is significant overlap in the constructs or dimensions measured by the two frequently used MSQ and FFMQ instruments and the widely used health related quality of life scales SF-12 and EuroQol’s EQ5D. MSQ and FFMQ scale scores and MSQ total scores tap into the dimensions of health related quality of life.
(2058) Identifying symptom clusters in the endocrine scale of the functional assessment of cancer therapy measurement system (The FACT-ES) Jacobien Kieffer, Ph.D, The Netherlands Cancer Institute, Department of Psychosocial Research, Netherlands; Neil K. Aaronson, PhD, The Netherlands Cancer Institute, Netherlands AIMS The Functional Assessment of Cancer Therapy Endocrine Symptom Scale (the FACT-ES) is an 18 item self-report questionnaire developed to assess treatment-induced endocrine symptoms (e.g., among women with breast cancer undergoing chemotherapy or hormonal therapy; women with ovarian cancer with an oophorectomy). The scoring algorithm for the FACT-ES yields a single, overall score. The aim of this study was to determine whether the FACT-ES also compromises a number of distinct symptom clusters that could yield subscale scores.; METHODS FACT-ES data were obtained from two samples: (1) 330 women who had experienced chemotherapy-induced menopause; (2) 372 women who had undergone prophylactic oophorectomy due to a heightened risk of familial breast/ovarian cancer. In the breast cancer sample, we used principal component analysis (PCA) with oblimin rotation to determine if a
91 psychometrically robust and clinically interpretable factor structure could be identified. We subsequently tested the tentative factor structure in the oophorectomy sample using confirmatory factor analyses (CFA) with WLSMV estimator.; RESULTS The PCA identified six symptom clusters, reflecting: (1) vasomotor symptoms (3 items; e.g., hot flushes); (2) gynaecological symptoms (4 items; e.g., vaginal dryness); (3) sexual symptoms (3 items; e.g., loss of interest in sex); (4) mood (2 items; e.g., irritability); (5) weight (2 items, e.g., weight gain); and (6) other endocrine symptoms (4 items; e.g., feeling light-headed). These 6 factors explained 64 % of the variance in the data (KMO = 0.68; Bartlett’s test of sphericity (2 = 1,681, p \ 0.001). The CFA confirmed the six cluster solution found in the PCA and demonstrated acceptable to good fit to the data (CFI = 0.99; TLI = 0.99; RMSEA = 0.06). Correlations between clusters were significant and ranged between 0.15 and 0.78. A second order factor was introduced into the model to explain the correlations between the clusters. The model demonstrated an acceptable to good fit (CFI = 0.99; TLI = 0.99; RMSEA = 0.07).; CONCLUSIONS Results suggest the tenability of a subscale structure as well as an overall score for the FACT-ES. Including subscale scores can provide us with more insight into the co-occurrence of endocrine symptoms and allow for a more tailored management of these symptoms.
(2059) Who in the general population may experience poor health-related quality of life? Therese Djarv, MD PhD, Karolinska Institutet, Sweden; Anna Wikman, Karolinska Institutet, Sweden; Mohammed A. Johar, B.Sc.(Hons), M.Sc., Karolinska Institutet, Sweden; Pernilla Lagergren, PhD RN, Karolinska Institute, Dept. of Molecular Med, Sweden AIMS To clarify the risk factors for poor HRQL in an unselected population.; METHODS A Swedish population-based cross-sectional survey was conducted. Predefined individual characteristics were; sex, age, educational level, marital status, Body Mass Index, physical activity and tobacco smoking. Five aspects of HRQL assessed with the EORTC QLQ-C30 included in the analyses were the following functions: physical, role, emotional, social and cognitive. Individuals were categorised into: ‘‘poor HRQL’’ if they scored at least 10 points (scale 0–100) poorer than the mean score of the total sample. Multivariate analyses of characteristics were performed using classification and regression tree (CART) analysis.; RESULTS Of 6,969 randomly selected people, 4,910 (70.5 %) participated. The first split in all five CARTs was number of comorbidities, as well as the second split often was physical activity. However, Body Mass Index and tobacco smoking status had minor impact.; CONCLUSIONS This large population-based study indicates that number of comorbidities is the best predictor and physical activity is the next best predictor of poor HRQL.
(2060) Psychosocial problems, family functioning and general perception of health Juan J. Dapueto, MD, PhD, Facultad de Medicina, Universidad de la Repu´blica, Uruguay; Laura De Souza, Institute of Clinical Psychology, Psychology, Universidad de la Repu´blica; Nicola´s Fiori, Faculty of Social Sciences; Claudia Antiga, University Health Division, Universidad de la Repu´blica; Iris Dutra, Faculty of Nursing, Universidad de la Repu´blica; Darı´o Padula, Faculty of Psychology. Universidad de la Repu´blica; Rau´l Ramı´rez, Faculty of Economics AIMS Family dysfunction has been associated with health problems in primary care. We included the Family APGAR as a measure of social support and family functioning in a survey of health
123
92 perceptions and quality of life in a sample of university workers in Uruguay to depict the associations between psychosocial problems, family dysfunction and general perception of health.; METHODS The Family APGAR was scored on a 5-point Likert-type scale and dichotomized in two categories, dysfunctional and functional family. General health perception was assessed with the Spanish translation of the PROMIS Global Health Scale version 1.0. Psychological distress was evaluated with the PROMIS Emotional Distress Anxiety and Depression Scales. Statistical analyses consisted of Chi square and T test statistics.; RESULTS Of a randomized sample of 1,200 subjects, 744 (62 %) completed the questionnaires. Mean age was 42.4 years (SD: 12.08), 57 % were women, 64 % had a university degree, with 162 subjects (21.8 %) reporting family dysfunction. Family APGAR Cronbach alpha coefficient was .91. Variables associated with good family functioning were living with other, being married, and having at least one child. Having a chronic disease, high risk alcohol intake, history of mental disorder, suicidal thoughts, being victim of domestic violence, and feeling discriminated at work were associated with family dysfunction (chi2 p \ .05). Family features related to family dysfunction were chronic disease, disability or history of suicidal attempt in family member (chi2 p \ .05). Age, sex, ethnicity, working status were not significantly related to family functioning. Significant differences in mean scores of the PROMIS Global Health Physical and Mental subscales, as well as the PROMIS Anxiety and Depression Subscales (all p values \.000) were observed, with subjects with family dysfunction reporting lower physical and emotional well-being, and higher anxiety and depression.; CONCLUSIONS The Family APGAR was accurate in discriminating between subjects according to marital status, physical and mental conditions, and psychosocial issues at home and work. Family dysfunction was associated with living alone, health problems, risk behaviors and psychosocial problems. These results support the assumption that family support can modulate the effect of stressful events reducing the impact on QOL and emotional well-being.
(2061) Evaluation of fatigue levels among workers affected by the East Japan Great Earthquake in order to determine health care needs and prevent disaster-related disease Toshiko Yoshida, PhD, Miyagi Univ School of Nursing, Japan; Daisuke Sato, Miyagi University; Toshie Sawaguchi, Miyagi University; Yukie Yamada, Miyagi University, Japan; Makoto Shimoyama, Miyagi University; Masayasu Inoue, Miyagi University; Hirohiko Kuratsune, Kansai University of Welfare Sciences AIMS The East Japan Great Earthquake caused serious damage to Japan. People in the stricken areas still face many health problems as well as increased risk for chronic and mental diseases. In our study, we evaluated the fatigue level of people working in the stricken areas in order to prevent disaster-related disease, particularly due to overwork.; METHODS We conducted a survey to investigate fatigue levels using a fatigue diagnosis system which evaluate autonomic function, oxidative stress (D-ROMs, BAP), sleep-wake rhythms, and the questionnaire of perceived fatigue. We investigated 368 public servants working in the stricken areas from October 2011 to March 2012. The approval of the Ethics Committee of Miyagi University was obtained for the evaluation of fatigue levels.; RESULTS Measurements of autonomic function showed that 27 % of people had lost the balance of LF (Low-frequency) and HF (High-Frequency). The results of a questionnaire of perceived fatigue showed that 62 % of people had subjective symptoms of fatigue. Evaluation of sleep-wake rhythms showed that the average number of halfway awakenings (1 min or more) during sleep was 6.1 ± 3.8 times, five times or more of the halfway awakening was 54 %, and less than 5 times of halfway awakening was 46 %. This suggested that sleep quality was
123
Qual Life Res (2012) 21:1–132 deteriorating. The result of oxidative stress showed that the average value of d-ROMs was higher than value of healthy people under normal stress revels (p \ 0.001). Moreover, 35 people high d-ROMs value with a low BAP value, results which suggest that those people are at high risk for diseases.; CONCLUSIONS The fatigue levels of people working in the stricken areas placed them at elevated risk of disease. Long-term support from health care professionals and the opportunity for health consultation and mental support are important for the improvement of QOL for a people working in the stricken areas. This study was supported by a Grant-in-Aid for research from the Japan Science and Technology Agency (JST)/Research Institute of Science and Technology for Society (RISTEX), of the Ministry of Education, Cultural, Sports, Science and Technology, Japan.
(2062) The DysDD: a new patient assessment for dysmenorrhea Josephine M. Norquist, MS, Merck Sharp & Dohme Corp.; Louise Humphrey, Adelphi Values; Helen Kitchen, Adelphi Values, United Kingdom; Ellen Mommers, Merck Sharp & Dohme, Corp.; Tayyaba Rehman, Merck Sharp & Dohme, Corp.; Allison Martin, MHA, Merck & Co. Inc. Pharmaceuticals AIMS To develop an electronic diary (eDiary) designed to assess dysmenorrhea, a syndrome of painful menstruation. Prevalence is estimated at 23–90 % of post-menarcheal females. Although not lifethreatening, dysmenorrhea can be debilitating and psychologically taxing for many women.; METHODS A literature review was conducted to develop a draft conceptual framework which was then revised following concept elicitation (CE) interviews with US subjects (n = 52). Diary items were generated for the Dysmenorrhea Daily Diary (DysDD) based on the CE findings and then tested in an electronic format as part of a usability and cognitive debriefing (CD) study with US subjects (n = 24). Subjects completed the diary for at least one menstrual cycle and then participated in a CD interview in which they were queried on the relevance and comprehensibility of the items and usability of the electronic device. Diary data collected was used to perform preliminary quantitative analyses to support the content validity of the DysDD.; RESULTS The DysDD is a 9-item, 24-h recall eDiary assessing severity of pelvic pain/cramps, severity of menstrual bleeding, rescue medication use, and impact on work/school, social/ leisure, physical functioning, and sleep. For CE, 52 subjects with dysmenorrhea were recruited including 24 adolescents (aged 14–17) and 28 adults. Conceptual saturation was reached in both groups. For CD, 24 subjects (12 adolescents, 12 adults) were recruited to test the DysDD at home and take part in CD. Subjects across both age groups confirmed that all items were relevant to their dysmenorrhea experiences and did not report any problems using the electronic device. Analysis of the response distribution frequencies from days with bleeding and/or pain revealed no ceiling or floor effects and all response options on each item were selected throughout the testing, suggesting that the response scales were appropriate and captured different severity levels.; CONCLUSIONS The DysDD instructions, items, and response scales were well understood by subjects. The DysDD is a comprehensive assessment of dysmenorrhea; concepts measured were relevant to subjects’ experience with no identified missing concepts. The DysDD demonstrated content validity in both adults and adolescents experiencing dysmenorrhea. Psychometric properties will be assessed in an upcoming validation study.
(2063) Cognition and functional recovery after hip fracture Gian Jhangri, University of Alberta; Allyson Jones, PhD, University of Alberta, Dept. of Physical Therapy, Canada; David Feeny, PhD, University of Alberta, United States
Qual Life Res (2012) 21:1–132 AIMS Clinicians have long suspected that baseline cognition affects recovery after hip fracture. In a prospective longitudinal cohort study, we examined whether cognitive status affects functional recovery during the first 6 months after hip fracture.; METHODS A population-based prospective inception cohort of 383 patients was recruited. Patients were 65 years or older with a primary diagnosis of hip fracture. Patients were interviewed in hospital within 3–5 days after surgery and re-interviewed at 1, 3 and 6 months. At baseline patients completed the Mini-Mental State Exam (MMSE). At each of the interviews patients completed an assessment of function using the Functional Independence Measure (FIM). The Health Utilities Index Mark 2 and Mark 3 questionnaire was also completed. Family caregivers were recruited as proxy respondents and were the only respondents for those patients with MMSE \18. Linear mixed models were developed for functional recovery to evaluate changes over time while adjusting for covariates including age, gender, baseline FIM, comorbidities, type of hip fixation, baseline hip pain, living accommodation, and length of stay (LOS).; RESULTS The mean age was 81.3 (SD 7.5) years, and 268 (70 %) were female. 104 (27 %) patients had MMSE scores less than 18. The median LOS in an acute care setting was 10 days. 6 month mortality rate was 12.2 % (n = 47). The mean FIM scores at baseline was 59.4 (SD 20.0) which improved to 80.6 (SD 31.5) at 1 month; 87.9 (SD 33.1) at 3 months; and 88.7 (SD 32.4) at 6 months (p \ 0.05). Baseline cognition was statistically significant and quantitatively important in explaining functional recovery over the 6 month period; patients with lower baseline cognition showed less recovery. Other factors which were statistically significant included age, LOS, living in facility versus home, type of hip fixation received, number of comorbidities, and hip pain at baseline.; CONCLUSIONS Although functional status is very poor within the first week of injury, large improvements are seen within 3–6 months. The pattern of recovery, however, is slower for those patients with cognitive impairment. These patients recover less successfully than patients without cognitive impairment.
(2064) Inside the black box: Patients’ attitudes to PRO assessment Rebecca L. Mercieca-Bebber, The University of Sydney, Australia; Madeleine T. King, PhD, Psycho-oncology Cooperative Research Group (PoCoG), Australia; Phyllis Butow, PhD, University of Sydney, School of Psychology, Australia AIMS The subjectivity of patient-reported outcomes (PROs) leads some critics to question their reliability and validity. This qualitative study, conducted as part of a larger, longitudinal study of perceived symptom benefit in women with resistant/refractory ovarian cancer (Stage 1 of the Gynaecologic Cancer Intergroup Symptom Benefit Study (GCIG-SBS)), aimed to examine the attitudes of patients to completing PRO forms. The aim of Stage 1 GCIG-SBS was to identify the best from a suite of PRO measures to take forward into Stage 2 GCIG-SBS.; METHODS Twenty patients, having completed a battery of well-validated PRO questionnaires over a course of chemotherapy, were interviewed about their attitudes to the questionnaires and response scales. Interviews were transcribed and reviewed by three researchers to identify common themes. Thematic analysis was conducted and managed through NVivo Software.; RESULTS Thematic analysis uncovered two major themes: 1) attitudes to questionnaires and the measurement process; 2) difficulty reflecting on own treatment, symptoms and mortality. The first theme included four subthemes: frustration, confusion, scale preferences and altruism. Patients were frustrated by the repetition of issues across the battery of questionnaires and confused about the purpose of the study. Consequently, patients admitted they did not always complete
93 questionnaires accurately and honestly. Patients preferred shorter questionnaires with fewer response options and many reported difficulty relating to the numeric scales. Patients took different approaches to the response scales: while some considered the visual space between response options, some used the words attached to the scale and a smaller group used the scale numbers. Some patients reported difficulty responding due to uncertainty about the cause of their symptoms. Patients felt their responses were inaccurate as a result. Despite this, participation was motivated by altruism. Patients wanted to increase knowledge about their disease and help future patients. The second major theme included subthemes of dishonesty and distress. Some patients deliberately understated their symptoms to avoid clinical repercussions of their side-effects, while others did so to avoid distress caused by reflecting on their health and QOL.; CONCLUSIONS This analysis revealed the different attitudes and approaches of patients in responding to questionnaires and serves to caution researchers when interpreting quantitative data.
(2065) Subjective well-being and six-year survival among very old people living in a Japanese community Hajime Iwasa, PhD, Tokyo Metropolitan Institute of Gerontology, Japan; Yasuyuki Gondo, Osaka University; Yukie Masui, Tokyo Metropolitan Institute of Gerontology, Japan; Hiroki Inagaki, Tokyo Metropolitan Institute of Gerontology AIMS Empirical studies have reported a longitudinal relationship between psychological health status and longevity in later life. However, few studies have investigated the relationship among very old people. This study sought to examine the relationship between subjective well-being and all-cause mortality among oldest-old elderly people (85 years and over), using a 6-year prospective cohort study design with mortality surveillance.; METHODS The participants were 66 men and 109 women aged 85 years and over at the baseline, living in an urban Japanese community. The dependent variable in the analysis was survival time, calculated as the number of days between the baseline and the date of death or censoring (including survivors and dropouts due to migration from the target region). The 17-item Philadelphia Geriatric Center Morale Scale was conducted at baseline to assess subjective well-being, and used as an independent variable. The baseline data on age, education level, chronic disease (including stroke, heart disease, and diabetes mellitus) and activities of daily living (measured by the Barthel Index) were used as covariates. Multivariate Cox proportional hazards model, adjusted for the potential confounders, was performed to clarify the longitudinal relationship between subjective well-being and all-cause mortality, stratified by genders.; RESULTS During the 6-year follow-up period, 90 subjects (42 men and 48 women) died, and 6 subjects (3 men and 3 women) moved to a different region of Japan and were lost to followup. The mean subjective well-being scores for women and men were 11.2 ± 3.1 and 10.9 ± 3.7, respectively. Use of the multivariate Cox proportional hazards model showed that a low level of subjective wellbeing was independently associated with the risk for all-cause mortality in women (1-standard-deviation decrement; hazard ratio [HR] = 1.47, 95 % confidence interval [CI]: 1.11–1.95, p \ 0.01), and not in men (HR = 0.99, 95 % CI: 0.71–1.39, p = 0.98).; CONCLUSIONS This study found that subjective well-being is a reliable predictor for all-cause mortality among the oldest-old women, suggesting that subjective well-being is essential for surviving in very old age. Therefore not only physical health but subjective well-being should be also focused on for development of longevity-promoting strategies.
123
94 (2066) Equating scores on two dermatology-specific patient-reported outcome measures James Twiss, BSc, Galen Research Ltd, United Kingdom; Stephen P. McKenna, PhD, Galen Research Ltd., United Kingdom AIMS Recent research has shown that it may not be valid to compare scores across diseases using generic outcome scales. An alternative approach is now available: Scores from two different disease-specific measures can be co-calibrated onto the same scale using Rasch analysis as long as the measures are based on the same measurement model. The aim of the study was to investigate whether it is feasible to co-calibrate scores for patients with different skin diseases using measures specific to psoriasis and atopic dermatitis.; METHODS The Psoriasis Quality of Life (PsoriQoL) scale and the Quality of Life in Atopic Dermatitis (QoLIAD) scale were both based on the needsbased model of quality of life (QoL). Each was developed using Rasch analysis and provides unidimensional measurement. The scales share 5 common items allowing co-calibration of scores using a common item design. Initially, the scales were analysed separately for fit to the Rasch model. Problematic items were removed. The scales were then combined and assessed again for fit to the model. Common items were assessed for Differential Item Functioning (DIF) by disease.; RESULTS The sample consisted of 146 psoriasis (male 50 %, mean age = 44.2 range = 17–83 years) and 146 atopic dermatitis patients (male 50 %, mean age = 45.5, range = 20–82 years). Both scales showed adequate overall fit to the Rasch model (PsoriQoL: Chi2 = 56.45, df = 50, p = 0.25; QoLIAD: Chi2 = 55.34, df = 50, p = 0.28). Evidence of multidimensionality was present in the QoLIAD. One item was removed due to high residual correlations with 2 others, improving model fit statistics. One item in each scale showed misfit and these were removed. After combing the items one common item showed DIF by disease and was also removed. The final set of items showed excellent fit to the model (Chi2 = 99.20, df = 84, p = 0.12).; CONCLUSIONS It is now possible to provide valid comparative QoL scores for patients with different diseases. Such comparisons cannot be made meaningfully using generic PROs that also tend to have poor measurement properties.
(2067) Outcomes of a secondary preventive self-management program in occupational skin disease Katja Spanier, M.A., Hannover Medical School, Coordination Centre for Applied Rehabilitation Research, Germany; Matthias Lukasczik, University of Wuerzburg, Germany; Sandra Nolte, PhD, Deakin University, Australia, Germany; Ute Pohrt, (4) Inst. for Statutory Accident Insurance and Prevention; Michael Schuler, University of Wuerzburg; Monika Schwarze, Medical School Hannover, Germany AIMS In Germany, secondary individual preventive (SIP)-programs for insures with occupational skin diseases are regularly conducted by the Institution for Statutory Accident Insurance and Prevention of the Health and Welfare Services (BGW). Courses take place at training and advisory centers. To date, program effectiveness has only been evaluated via user satisfaction surveys. Objective of this study is to examine whether the Health Education Impact Questionnaire (heiQ), a questionnaire that has been developed to measure proximal outcomes of patient education interventions, is an appropriate instrument to evaluate a self-management program aimed at prevention. In addition the Skindex-29, an instrument measuring health-related quality of life in skin diseases, was applied.; METHODS A total 502 participants of the SIP-programs filled out the heiQ and the Skindex-29 at the beginning (T1), at the end (T2), and 6 months after courses (T3). Ninety-three percent were female, aged 39.9 (12.2) years. Factorial validity (confirmatory factor analysis), concurrent validity (Pearson
123
Qual Life Res (2012) 21:1–132 correlations), and reliability (CRC-values) were tested. Further, Standardized Effect Sizes (SES) and Standardized Response Means (SRM) were calculated.; RESULTS The heiQ shows good factorial validity, while reliability was satisfying to very good. When tested with the Skindex-29, only heiQ scale Emotional distress showed a clear correlation. heiQ baseline scores were generally high, i.e. above 3 on the 4-point Likert scale. Further, between T1 and T2, and T1 and T3 heiQ scales Skill and technique acquisition (STA) and Self-monitoring and insight (SMI) showed small effects (T1 to T2: SES = 0.24 (STA), SES = 0.13 (SMI); T1 to T3: SES = 0.22 (STA), SES = 0.37 (SMI)). All other heiQ scales showed significant effects but none were clinically meaningful. In contrast, small to medium significant effects were found in the Skindex-29.; CONCLUSIONS The findings prove that the program seems to be successful. However, the usefulness of the heiQ in the context of preventive self-management programs in skin disease seems limited as most heiQ scales already show high values at the beginning of courses, i.e. potential ceiling effects cannot be ruled out. We therefore recommend using heiQ scales STA and SMI, showing reasonable sensitivity to change, in conjunction with diseasespecific questionnaires such as the Skindex-29.
(2068) Development and evaluation of a syndrome diagnostic instrument for angina: Syndrome Differentiation Questionnaire of Qi Deficiency and Blood Stasis (SDQ-QDBS) Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Ma Yuan, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Chen Zhihui, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yu Changhe, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Lv Meijun, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS The present study was conducted to develop and evaluate a syndrome diagnostic questionnaire for the angina patients with Qi Deficiency and Blood Stasis syndrome (QDBS).; METHODS 13 items of the draft scale were formed by the patient focus group, the expert panel, physician survey and literature review. A total of 200 healthy subjects and 991 angina inpatients were asked to complete Syndrome Differentiation Questionnaire of Qi Deficiency and Blood Stasis (SDQ-QDBS). All the inpatients were diagnosed QDBS or non-QDBS by Chinese medicine clinicians. The survey data was used to estimate the reliability, validity, and to define the optimum cut-off point of the questionnaire.; RESULTS The final scale had 2 dimensions, 2 facets and 11 items: angina symptom (4 items), QDBS syndrome which included Qi deficiency syndrome facet (5 items) and Blood stasis syndrome facet (2 items). It had satisfactory reliability with Cronbach’s alpha coefficient equal to 0.791, the Spearman correlation coefficient (CC) between 0.888 and 0.967 and the item-total correlation coefficient between 0.431 and 0.722. The final scale could distinguish the QDBS group from the healthy group in both angina symptom and QDBS syndrome dimension (p \ 0.001), and it could distinguish the QDBS group from the nonQDBS group in syndrome dimension (p \ 0.001) too. In the receiver operating characteristic curve (ROC) analysis, the sensitivity, specificity, the area under the curve and the Youden index were 85.2 %, 86.0 %, 0.910 and 0.712 respectively. The optimum cut-off level was defined as 20.5 points, which meant the questionnaire could be used as a diagnostic tool for angina with QDBS.; CONCLUSIONS The questionnaire was a validated and disease specific questionnaire to diagnose the QDBS syndrome of patients with angina in TCM.
Qual Life Res (2012) 21:1–132 (2069) The application of the Health Education Impact Questionnaire (heiQ) in Japan—adaptation and validation of a measure to evaluate chronic disease self-management interventions Sandra Nolte, PhD, Deakin University, Australia, Germany; Rie Morita, Division of Health Sciences, Osaka University; Mikako Arakida, International University of Health and Welfare; Hiroshi Mikami, Osaka University, Japan; Gerald R. Elsworth, Deakin University; Richard H. Osborne, Deakin University AIMS Most quality of life and health status measures do not appropriately capture expected outcomes from chronic disease selfmanagement and health education programs. To ensure a comprehensive evaluation of these interventions, the Health Education Impact Questionnaire (heiQ) was developed in Australia in 2007. It is currently applied in many countries worldwide and translated into over 20 languages. The heiQ comprises eight separate domains. It has been found to have robust psychometric properties across several English speaking countries. The aim of this paper is to describe the translation of the heiQ into Japanese and its adaptation to the cultural specificities. In particular, the aim was to adapt the heiQ for application in the evaluation of the Japanese Specific Health Consultation (SHC) targeted at people with Metabolic Syndrome (MetS), i.e., people at risk of diabetes.; METHODS A standardized translation and adaptation protocol, with forward and back translation and consensus meetings, was applied. A draft version was piloted and included cognitive interviews (n = 12). To explore the psychometric properties, 250 participants of the SHC completed the heiQ-J and the following comparator scales: the Japanese version of the SF-36, the Japanese Health Locus of Control (JHLC) scale, the Sense of Coherence (SOC) scale, and the Social Support Measurement scale.; RESULTS The cognitive interviews revealed that the heiQ-J was understood by participants for most items and scales. Internal consistency (a) was good to very good for all scales (0.70–0.88); the test–retest intraclass correlation coefficients were high (=0.83). Concurrent validity was supported by high correlations with like scales and weak correlations with dissimilar scales.; CONCLUSIONS The heiQ-J has good validity and reliability. It is likely to be a useful measure in the Japanese context and deliver important data for clinicians, health professionals, and commissioners of patient education in Japan.
(2070) The outcomes and experiences questionnaire: development and content validation Elizabeth J. Gibbons, MSc, University of Oxford, United Kingdom; Carolina Casanas I Comabella, PROM Group, University of Oxford; Monica Hadi, PhD, University of Oxford, United Kingdom; Anne Mackintosh, BA, PROM Group, Univ of Oxford, United Kingdom; Raymond Fitzpatrick, PhD, Univ of Oxford, Dept of Public Hlth, United Kingdom AIMS Measuring outcomes which are important to patients, healthcare professionals and commissioners of services poses several challenges, specifically, whether patient-sensitive outcomes are relevant indicators of the quality of care. In England there are several national mandatory initiatives using Patient-reported Outcome Measures (PROMS), for example: PROMs in elective surgical procedures, the National patient experience survey programme, and the Commissioning for Quality and Innovation (CQUIN) payment framework. Although these methodologies are now enshrined in the National Health Service (NHS) infrastructure, there are some gaps and opportunities for further utilisation of PROMs. What appears to
95 be lacking is a measure which captures not only the patient’s perceptions of the quality of care but also their perceived benefit from healthcare. The aim of this study was to develop a patient-focused questionnaire which measured: (1) Outcomes and benefits of care from the patient’s perspective; (2) patients’ experiences of care. This work was commissioned by the UK government Department of Health.; METHODS The underpinning policy and literature which informed the development of the Outcomes and Experiences Questionnaire are described. The methods of further development, item reduction and consultation with patients are presented. Thirty patients were recruited as advisors who had some experiences with hospital care during the last year. Three rounds of cognitive interviews were conducted with these advisors. The questionnaire was subjected to close evaluation by an expert panel, patient and public experience and engagement policy leads from the Department of Health, and other researchers with survey development expertise.; RESULTS The final questionnaire has two main constructs: Patientperceived benefit of outcomes and Patient experiences of care. A global retrospective judgement of outcome item is also included. Its brevity and cross-sectional administration reduces administrative and completion burden. It has high face and content validity. The final questionnaire is currently undergoing secondary psychometric evaluation.; CONCLUSIONS Following psychometric evaluation, it is anticipated that this questionnaire will be used in the NHS in the future and will complement the existing PROMs programme in England. 1.http://www.ic.nhs.uk/proms. 2. http://www.nhssurveys. org/. 3. http://www.institute.nhs.uk/world_class_commissioning/pct_ portal/cquin.html
(2071) Effect of recall period on item responses to activity limitation items in multiple sclerosis Peter Hagell, Kristianstad University, Sweden AIMS Patient-reported rating scales typically include instructions regarding recall periods. However, it is unclear whether such instructions affect how items are responded to. This study addressed the impact of specifying a recall period on how respondents with multiple sclerosis (MS) treat activity related items.; METHODS The Activity section of the Patient-Reported Outcome Indices for Multiple Sclerosis (PRIMUS-A) was mailed to 285 people randomly selected from a local Swedish MS registry, of whom 199 (70 %) responded. Respondents had been randomized to receive either the original PRIMUS-A (instruction: ‘‘please describe your ability to do each of the 15 activities listed below during the last week, without the use of aids or assistance’’) or an experimental PRIMUS-A version (instruction: ‘‘please describe your ability to do each of the 15 activities listed below without the use of aids or assistance’’). Response categories were ‘‘Able to do on own without difficulty’’, ‘‘Able to do on own with difficulty’’ and ‘‘Unable to do on own’’, scored 0–2, respectively. Complete data from 195 respondents were analyzed regarding Differential Item Functioning (DIF) by questionnaire version (original vs experimental PRIMUS-A) using the Rasch measurement model. RESULTS Floor/ceiling effects were 30/5 %, respectively. Reliability (Person Separation Index) was 0.93. Data demonstrated overall fit to the Rasch model (item-trait interaction chi-2, 19.02; P = 0.94) with standardized item residuals ranging between -1.79 and 0.12 (chi2, \4.71; unadjusted P, = 0.09). Very similar results were obtained when analysing the two versions separately. There was no DIF by questionnaire version (F, \3.99; unadjusted P, = 0.05). Nor were there any DIF detected by gender, age, symptom severity or ability to walk outdoors. CONCLUSIONS Lack of DIF between PRIMUS-A versions suggests that recall period specification does not affect item functioning among respondents. This argues against the use of
123
96 specified recall periods and suggests that questionnaires can be simplified by omitting such instructions withoutaffecting outcomes. Furthermore, rating scale outcomes should not be over interpreted in relation to specified recall periods. Additional studies, including cognitive debriefing interviews using additional scales and samples are needed for a fuller understanding of response processes in relation to recall periods and other factors.
(2072) Which scales in the thyroid-specific patient-reported outcome ThyPRO are most responsive to treatment of Graves’ disease? Per Cramon, Copenhagen University Hospital, Denmark; Jakob Bjorner, PhD, National Research Centre for the Working Environment, Denmark; Steen J. Bonnema, Odense University Hospital; Laszlo Hegedu¨s, Odense University Hospital; Ulla Feldt-Rasmussen, Copenhagen University Hospital; ˚ se K. Rasmussen, Copenhagen University Hospital Rigshospitalet, A Denmark; Mogens Groenvold, MD, PhD, University of Copenhagen, Denmark; Torquil Watt, MD PhD, Copenhagen University Hospital, Denmark AIMS To examine how scales in the thyroid-specific patient-reported outcome ThyPRO respond to treatment of Graves’ disease.; METHODS The study included 59 newly referred patients with active Graves’ disease (hyperthyroidism), who were subsequently treated with antithyroid drugs, surgery or radioiodine. Baseline ThyPRO scores were compared with ThyPRO scores 6 months after treatment (paired t test). Changes in patient scores were evaluated by effect sizes (ES; mean difference/SD baseline). ES between 0.2 and 0.5 was defined as a small effect, while ES between 0.5 and 0.8 and above 0.8 was defined as moderate and large effects respectively.; RESULTS ThyPRO scores improved significantly after treatment of Graves’ disease on 9 of 13 scales. Large treatment effects were found for the Hyperthyroid Symptoms (ES = 1.13) and Tiredness (ES = 0.83) scales, while moderate treatment effects were found on the Anxiety (ES = 0.77), Emotional Susceptibility (ES = 0.56) and Impaired Daily Life (ES = 0.53) scales. Small effects were found on the Hypothyroid Symptoms (ES = 0.37), Cognitive Complaints (ES = 0.35), Depressivity (ES = 0.33) and Goitre Symptoms (ES = 0.21) scales, whereas no significant effects were found on the Impaired Social Life, Impaired Sex Life, Cosmetic Complaints and Eye Symptoms scales.; CONCLUSIONS Treatment of Graves’ disease improved QoL scores significantly on 9 of 13 ThyPRO scales. The scales most responsive to treatment were the Hyperthyroid Symptoms and Tiredness scales. The relative sizes of the changes within the various scales are clinically meaningful. Hyperthyroid symptoms and tiredness are core elements of hyperthyroidism, as are anxiety and emotional lability. It is possible, that the aspects not improving with therapy require longer follow-up in order for improvement to occur. Another possibility is that the aspects not improving were not affected by Graves’ disease at baseline.
(2073) Using the Personal Wellbeing Index (PWI) in a population living with HIV Vicki E. Hutton, Monash University; Roseanne Misajon, Monash University, Australia AIMS The Personal Wellbeing Index—Adult (PWI-A) measures life satisfaction across 8 domains (standard of living, personal health, achievement in life, personal relationships, personal safety, community-connectedness, future security, and spirituality-religion). The aim of this study was to use Rasch analysis to assess the internal validity of the PWI and determine the suitability of the use of PWI in a
123
Qual Life Res (2012) 21:1–132 population living with HIV.; METHODS The sample included 274 participants recruited through Australian AIDS councils and HIVspecific online support groups. The questionnaire included the PWIA, the EQ-5D, as well as the 10-item Emotional wellbeing/ Living with HIV subscale from the Functional Assessment of Human Immunodeficiency (FAHI) questionnaire. The internal construct validity of the PWI was assessed by undertaking Rasch analysis using the RUMM2030 software. Concurrent validity was assessed through comparator measures.; RESULTS Analysis of the PWI domains indicated that there was good fit to the Rasch model, provided that the eighth domain, spirituality-religion, be removed, reducing the PWI to a 7-item measure. There were strong correlations between the PWI and the Emotional wellbeing/living with HIV subscale, as well as with the EQ-5D.; CONCLUSIONS The findings in this study demonstrate that the PWI has good psychometric properties in an HIV population, living in Australia or USA. The PWI (7-item) provides a brief, valid measure to quickly and reliably assess subjective wellbeing across broad life domains, and may be useful as a screening tool to assess support needs for individuals with HIV.
(2074) Improving PRO content analysis using World Health Organization definitions and the ICF Nora Fayed, University Health Network, Canada; Alarcos Cieza, Ludwig-Maximilians-Universita¨t, Munich, Germany AIMS The International Classification of Functioning, Disability and Health (ICF) has been useful for content analysis of patient-reported outcomes (PRO). Mapping to the ICF on its own can merely demonstrate the health and health-related content of a PRO but not the extent to which quality of life (QOL) perspectives are found in a PRO. The aim of this study is to develop and test a method for distinguishing whether PROs contain health content according to the ICF lens of functioning, or QOL with the World Health OrganizationWHOQOL definition.; METHODS The definitions of functioning and disability from the ICF manual and QOL from the WHOQOL manual were reviewed by two content analysts for similarities and gaps in order to determine what distinguishes the concept of QOL from health. Criterion was established for making the conceptual distinction in adult and child PROs. Two additional analysts independently applied the definitions to 10 existing child PROs on an item by item basis to code whether the health or QOL content perspectives were represented in the items. The reliability between analysts’ interpretation and application was calculated with percentage agreement and Cohen’s kappa.; RESULTS The WHO definitions demonstrated that there was overlap in the domains assessed by the ICF definition of health and QOL. QOL was distinct from ICF-health because it emphasized an individual’s perception of their life or health situation, whereas the ICF perspective represented the actual presence, absence severity of a health or health-related state. Agreement between analysts ranged from 74 to 100 % with percentage agreement and 0.59–1.00 with Cohen’s kappa.; CONCLUSIONS Linking to the ICF is insufficient to distinguish a functioning or health-status PRO from a QOL PRO. Analyzing whether a health or QOL perspective is used to measure health and health related areas of life can be completed by applying interpreted WHO definitions. We suggest that comparing results of PROs that link to the same ICF code is conceptually incorrect unless the perspective is also taken into account.
(2075) The effect of neurological function on dyssomnia of patients with post stroke depression (PSD) Hai Ying, Department of Neurology, Affiliated Hospital of Liaoning University of TCM, Peoples Republic of China; Lv Meijun, Liaoning
Qual Life Res (2012) 21:1–132 University of Traditional Chinese Medicine, Peoples Republic of China; Yu Li, Department of Neurology, Affiliated Hospital of Liaoning University of TCM, Peoples Republic of China; Qi Wencheng, department of neurology, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yuan Dongchao, department of neurology, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Chen Zhihui, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yu Changhe, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS Dyssomnia is one of main symptoms of PSD patients. This study is aimed to investigate the effects of neurological functions on dyssomnia of PSD patients.; METHODS Methods: Restoration (3 weeks to 6 months) stroke patients were selected from an ongoing study of quality of life on ischemic stroke patients. The NIHSS was used to assess the neurological function and quality of life. Patients were firstly divided into 2 groups (PSD and non-PSD). These two groups were then separated based on 3 sleep items (insomnia: initial, middle and delayed) that were extracted from the Hamilton Depression Rating Scale (HAMD). These 3 items were renamed as ‘‘difficulty in falling asleep’’, ‘‘difficulty in staying asleep’’ and ‘‘early awakening’’. The total scores of the HAMD were calculated based on the PSD/non-PSD groups. The difference between the total scores was detected by the Independent-Samples T Test. PSD patients were then divided into neurological function deficient and non-deficient groups by the item score (0 and non-0) in each of the 10 NIHSS items (except for level of consciousness, motor arm, motor leg). The total scores and the differences between the total scores were calculated. Finally, the correlation between the HAMD and the NIHSS was assessed using the Spearman Correlation.; RESULTS Of the 1,781 restoration stroke patients (age 62.66 ± 10.18 years; 972 male, 54.58 %), significant difference was obtained for the HAMD between the PSD and non-PSD group (p = 0.00). For the PSD patients, significant differences were observed for the NIHSS items ‘‘best gaze’’, ‘‘facial palsy’’, ‘‘dysarthria’’, ‘‘limb ataxia’’, ‘‘sensory’’ and ‘‘extinction and inattention’’ (p = 0.05) between the neurological function deficient and nondeficient groups. The following NIHSS items ‘‘best gaze’’, ‘‘facial palsy’’, ‘‘sensory’’, ‘‘extinction and inattention’’ and ‘‘dysarthria’’ were related to the HAMD, especially the item ‘‘extinction and inattention’’ (r = 0.115, P = 0.00). The correlation coefficient of the items ‘‘best gaze’’, ‘‘facial palsy’’, ‘‘sensory’’, ‘‘extinction and inattention’’ and ‘‘dysarthria’’ were 0.061, 0.072, 0.084, 0.115 and 0.047 respectively.; CONCLUSIONS Conclusions: Different aspects of neurological function deficits in PSD patients result in different effects on dyssomnia, especially for patients with ‘‘extinction and inattention’’. In the future, studies will be conducted to investigate the effect of different infarct sites of stroke on dyssomnia of PSD patients.
(2076) Adherence to long term quality of life questionnaires in localised high risk prostate cancer and causes of non compliance Abdenour Nabid, MD, FRCP(C), Centre hospitalier Universitaire de Sherbrooke, Canada; Nathalie Carrier, Centre hospitalier Universitaire de Sherbrooke; Andre´-Guy Martin, Hoˆtel-Dieu de Que´bec, Canada; Jean-Paul Bahary, M.D., FRCP, Centre Hospitalier de L’Universite de Montreal, Canada; Marie Duclos, Centre Universitaire de Sante´ McGill; Francois Vincent, Centre hospitalier re´gional de Trois-Rivie`res; Sylvie Vass, CSSS de Chicoutimi; Boris
97 Bahoric, Hoˆpital ge´ne´ral juif; Robert Archambault, Hoˆpital de Gatineau; Ce´line Lemaire, Hoˆpital Maisonneuve-Rosemont, Canada AIMS The purpose of the present analysis is to evaluate long term adherence to quality of life questionnaires (QOLQ) in a group of 630 patients with high risk prostate cancer (HRPC) treated in a prospective randomized phase III trial (PCS IV clinical trials, Gov. # NCT 00223171).; METHODS PCS IV evaluates the duration of androgen blockade (AB) in HRPC. Patients were randomised to 1 month of anti-androgen with 36 (36 m) vs 18 (18 m) months of LHRH analog plus radiotherapy (RT). QOLQ used EORTC 30 and PR 25 tools available in French and English, before the beginning of treatments, every 6 months during AB, 4 months after and then once a year for 5 years.; RESULTS From October 2000 to January 2008, 310 patients were randomised to 36 m and 320–18 m of AB. There are no significant differences in pre-treatment characteristics between the 2 groups for median age, initial PSA, Gleason score and clinical stage. With a median follow-up of 72 months, global adherence to QOLQ is 71.8 % (9,638/13,432). For the 9,638 QOLQ filled, 99.9 % of the questions were answered for EORTC 30 and 99.5 % for PR 25. Pre treatment adherence was 86.2 %: 87.1 % (36 m) vs 85.3 % (18 m). At 18 months, adherence was 74.8 % (36 m) vs 80.0 % (18 m), and dropped to 56.0 % for the 36 m group. Five years after AB termination, adherence was 47.0 %: 47.8 (36 m) vs 46.2 % (18 m). Adherence according to age stratification (\60, 60–64, 65–69, 70–74, = 75) varied between 69.8 and 72.6 %, p = 0.77. Adherence amongst the 10 participating centers varied between 48.2 and 85.8 %, p \ 0.001. Of the 3,794 missing QOLQ, ten causes of non-compliance were recorded, mainly: QOLQ not done 28.6 %, death 20.8 %, patient’s refusal to continue AB 15.2 %, lost to follow-up 8.8 %, biochemical failure 8.3 %.; CONCLUSIONS PCS IV evaluates AB duration in HRPC. In a long term setting, we report the rates of adherence to QOLQ and causes of non-compliance. Source of Funding: AstraZeneca Pharmaceuticals grant.
(2077) Assessing quality of care in lung transplant patients Maria-Jose Santana, PhD, University of Calgary, Canada; David Feeny, PhD, University of Alberta, United States; Sunita Ghosh, University of Alberta; Dale C. Lien, University of Alberta, Canada AIMS Patient assessments of quality provide important information about how well health care providers meet the needs of their patients. The present study examines whether patients’ assessments of health care quality varies by transplant status and sociodemographic characteristics using the Consumer Assessment of Health Plans Study (CAHPS).; METHODS Consecutive pre- and postlung transplant patients attending the lung transplant outpatient clinic at a tertiary institution completed: CAHPS; Hospital Anxiety and Depression Scale (HADS); and Health Utilities Index (HUI). CAHPS contains 38 items including one global rating of the patient’s personal doctor on a 0–10 rating scale. The questions are grouped into three composites: access (getting care quickly, getting needed care), provider (communication skills, shared decision making, knowledge of medical history), and provider’s office staff (follow-up on test results, courtesy/respect, and helpfulness). Data were analyzed using linear regression models. The dependent variables were CAHPS global rating (best to worst doctor) and two composites assessing doctors’ and nurses’ communication skills and shared decision making. The independent variables were patients’ transplant status, gender, age, education and self-rated health. The model was replicated using only pre-transplant subjects.; RESULTS The sample included 213 patients (103 pre- and 110 post-transplant) with mean age 53 of whom 50 % were female. In general, pre-
123
98 transplant patients reported lower scores than post-transplant patients. In particular, female pre-transplant patients with less education and poorer health status reported lower scores when assessing doctors’ communication than their male counterparts (p \ 0.05). Further exploration revealed that females pre-transplant patients with higher HADS anxiety scores tended to report lower scores (p \ 0.05).; CONCLUSIONS This study suggests that elevated anxiety in female pre-transplant patients may lead to perceived difficulties in communicating with their doctors. In order to improve quality of care the observed disparities should be addressed.
(2078) Comparing two methods for longitudinal analysis on questionnaire data: multilevel analysis versus multilevel IRT analysis Rosalie Gorter, VU University Medical Centre, Netherlands; Michiel R. De Boer, VU University, Amsterdam, The Netherlands; Martijn W. Heijmans, VU University Medical Centre, Amsterdam, The Netherlands; Jos WR Twisk, VU University Medical Centre, Amsterdam, The Netherlands AIMS Longitudinal questionnaire data assessed on an ordinal scale are mostly analyzed using linear mixed models treating the sum score of the items as a continuous dependent variable. Theoretically it would be better to use IRT methods for ordinal longitudinal data instead of using sum scores. IRT methods are widely used in practice on cross-sectional data to evaluate the psychometric properties of the questionnaire. Recently, statistical software has become available for incorporating Item Response Theory (IRT) methods in mixed models. We can now take the ordinal structure of the data into account, asses the psychometric properties of the questionnaire while taking into account the nested structure of the data (Measurements nested within patients). In this study, we aimed to compare the outcomes of the newer method to the traditional method for the analysis of development over time in different questionnaire assessed outcome measures.; METHODS We used questionnaires from two Dutch longitudinal cohort studies; the Longitudinal Aging Study Amsterdam (LASA) and the Amsterdam Growth and Health Study (AGHLS). In total, data from eight questionnaires, most of which consisted of several sub scales, with different psychometric properties, length, number of participants, and number of measurements were analyzed. Both models were fitted to the data after which we compared the results of tests regarding hypotheses on the development over time.; RESULTS We found different results in some sub-scales of the data sets in the development over time. In several sub scales of data sets, for example on two different measurement occasions of the subscale ‘Sleeping at night’ of the Sleep Wake Experience List, we found different results from the two models on whether or not there was a significant change in scores over time.; CONCLUSIONS Based on the results from this study we can conclude that results of analysis on the development over time of questionnaire assessed constructs might differ when using multilevel IRT methods compared to the linear multilevel methods that have been used so far.
Qual Life Res (2012) 21:1–132 these protocols to the development and evaluation of a quality of care measure for a disease that is poorly understood outside the field of adult hematology: adult sickle cell disease (SCD).; METHODS The following themes emerged based on formative research with 122 adults living with SCD: 1. Lack of knowledge of SCD on the part of providers in primary care and the emergency department (ED), leading to stigmatization of adults with SCD as drug seeking or leading to inappropriate and ineffective care; 2. Severe delay in receiving care in the ED as well as disrespectful behavior on the part of ED providers; 3. Providers who do not consider information provided by the patient, leading to under- or overmedication for pain; and4. As a consequence of the preceding, adults prefer to treat the symptoms of SCD outside of the traditional health care system. Based on the qualitative data from these interviews, we developed a field test instrument with questions that conformed to CAHPS design principles: referred to observable events rooted in the patient experience and triaged respondents to sets of questions based on their responses to a screener (Figs. 1, 2). After cognitive testing and revisions, field test data were collected from 556 patients at 7 different clinical sites throughout the US. Classical psychometric methods were used to evaluate the measurement properties of scales per standard CAHPS methodology.; RESULTS We identified ‘‘Access’’, ‘‘Provider Interaction’’ and ‘‘ED Care’’ composite measures (Table 1). The confirmatory factor analysis of this structure using SEM where correlated errors among questions were specified, resulted in excellent fit indices (CFI [ 0.95; RMSEA \ 0.06) for the three composites. Cronbach’s alpha for these scales ranged from 0.70 to 0.83, correlations with global ratings of care were moderate (0.32) to high (0.83) (Table 2). As anticipated, data showed worse care reported for SCD patients compared to Medicare, Medicaid and privately-insured.; CONCLUSIONS Thirteen items scaled into three composites resulted
Fig. 1 Example questions from Patient-Reported Quality of Care measure
(2079) Measuring the quality of care delivered to patients suffering from a rare disease San Keller, PhD, American Institutes for Research; Christian Evensen, American Institutes for Research; Roger Levine, American Institutes for Research; Marsha Treadwell, Children’s Hospital and Research Center at Oakland AIMS Fort 17 years, CAHPS investigators have developed numerous patient-reported measures of healthcare quality using a rigorous, standardized set of protocols. We describe here our application of
123
Fig. 2 Reports of Care for Adult SCD Patients Compared to Medicare, Medicaid or Fee-for-Service
Qual Life Res (2012) 21:1–132
99
Table 1 Composites and Items for the patient-reported qualityof-care for SCD tool Scale
Item description
Getting needed care
… when you tried to make an appointment to see DR/RN, how often were you able to get one as soon as you wanted? … when you needed care right away, how often did you get it as soon as you wanted? … what is the longest you had to wait before your pain was treated? How important were bad experiences …in your decision to avoid going for care? … how often did this doctor or nurse explain things in a way that is easy to understand?
Provider interaction
… how often did this doctor or nurse listen carefully to you? … how often did this doctor or nurse treat you with courtesy and respect? … how often did this doctor or nurse spend enough time with you? Emergency department care
… how often did the doctors treating you seem to really care about you? … how often did the nurses treating you seem to really care about you? … how often did the clerks and receptionists treat you with courtesy and respect?
understand risks for these symptoms and toxicities, a method of identifying those with the greatest symptomatic response to cancer therapy is needed.; METHODS In a prospective study of patients with CRC (N = 100), symptoms were documented weekly from start of oxaliplatin-based chemotherapy, during and after for up to 12 months using the Gastrointestinal Module of the M. D. Anderson Symptom Inventory. Case-clustering analysis was used to categorizpatients into high-severity versus low-severity symptom phenotypes over time.; RESULTS Over time, the 5 most severe symptoms reported by patients were numbness/tingling, fatigue, drowsiness, disturbed sleep, and distress. Majority (85 %) of patients received 5–14 cycles of chemotherapy, rest patients received 2–4 cycles therapy. Over 12 months of observation, 21 % of patients in high symptom group for a component score of 5 severe symptoms (fatigue, drowsiness, disturbed sleep, distress and poor appetite). Chemotherapy induced peripheral neuropathy (CIPN, reported as numbness/tingling) significantly and rapidly increased as the chemotherapy dose accumulated (p \ .0001), with 64 % of patients belonging to the high numbness/ tingling group. 30 % of all patients were in the high-pain group, significantly increased after 100 days. No effect of age, gender, education level, or employee status on high versus low symptom group membership. Significantly more Hispanic patients belonged to the high-pain group (P = 0.024). Diabetes was a risk factor for higher CIPN (P = .007).; CONCLUSIONS This study presents a method of categorizing longitudinal symptom report data to identify those at greater risk for several treatment-related symptoms. Efforts to identify those at risk for greatest symptomatic response should be an important component of the ongoing, rapid evolution of gastrointestinal cancer management, with the ultimate goals of better tolerance for therapy and improved overall survival.
… how much were the emergency room doctors and nurses able to help your pain? … how much did the emergency room doctors and nurses believe … you had very bad sickle cell pain?
Table 2 Measurement properties of quality of care scales Composite
Alpha
Correlations with global ratings of quality of care regarding… Routine care
Providers
ED care
Scaling success (%)
% at Floor
% at Ceiling
Overall
Access
0.70
0.49
0.32
0.65
0.44
100
2.5
0.6
Provider interaction
0.86
0.65
0.83
0.44
0.39
100
0
44
ED care
0.83
0.45
0.41
0.75
0.56
100
0
7.5
in a psychometrically-sound measure of care provided for this rare disease across providers.
(2081) Identifying cancer patients with high levels of treatment-related neuropathy and other symptoms in response to oxaliplatin-based chemotherapy Xin Shelley Wang, MD MPH, UT MD Anderson Cancer Center; Charles S. Cleeland, PhD, UT-MD Anderson Cancer Center; Mary Sailors, MDACC; Raza Bokhari, MDACC; Cathy Eng, MDACC AIMS Oxaliplatin-based chemotherapy (FOLFOX or FOLFIRI), a standard first-line chemotherapy for patients with colorectal cancer (CRC), is associated with debilitating general symptom burden, such as fatigue, pain, and distress, also with dose-limiting toxicity, such as cumulative peripheral sensory neuropathy. In order to better
(2082) Thanks for the score report—but what does it mean? Helping clinicians interpret patient-reported outcome (PRO) scores by identifying cut-offs representing unmet needs Claire Snyder, PhD, Johns Hopkins; Amanda Blackford, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins; Toru Okuyama, Nagoya City University Graduate School of Medical Sciences; Tatsuo Akechi, Nagoya City University Graduate School of Medical Sciences, Japan; Hiroko Yamashita, Hokkaido University Hospital; Tatsuya Toyama, Nagoya City University Graduate School of Medical Sciences; Michael Carducci, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins; Albert W. Wu, MD MPH, Johns Hopkins Bloomberg School of Public Health AIMS One of the challenges to using PROs in clinical practice is determining which scores represent a problem and require the clinician’s attention. If cut-offs identifying scores associated with patient problems were available, they would help clinicians interpret PRO score reports. In previous research, we used Supportive Care Needs Survey (SCNS) scores to identify patients with and without unmet needs. We then established cut-offs on the EORTC QLQ-C30 associated with patient-reported unmet needs. This study tests the validity of the cut-offs from the original analysis using a new dataset with a larger sample.; METHODS We conducted a secondary analysis of data collected using the Japanese QLQ-C30 and SCNS from 408 ambulatory breast cancer patients (mean age 56 years). We dichotomized SCNS item and domain scores as ‘‘no unmet need’’ versus ‘‘some unmet need’’ and then evaluated the ability of QLQ-C30 scores to discriminate between no/some unmet need using receiver operating characteristic (ROC) analysis. For domains with areas under the ROC curve (AUC) [ .70, we calculated the sensitivity and specificity of different cut-offs. We hypothesized that compared to our original analysis: (1) the same QLQ-C30 domains would have AUC [ .70; (2) the same SCNS items would provide the highest AUC; and (3) the
123
100
Qual Life Res (2012) 21:1–132
Table 1 Hypothesized relationship between QLQ-C30 and SCNS domains and resulting areas under the curve (AUC): original and validation analysis QLQ-C30 domain
SCNS domain/item(s)
AUC Original analysis
Table 2
Sensitivity and specificity of various cut-off scores: original and validation analysis
QLQ-C30 Domain
SCNS Item
Physical Function
Work around the home
Validation analysis
CutOff
Cohort
Sensitivity Specificity
80
Original
90
Hypothesized AUC C.70 Physical function
Physical & daily living needs .69–.81 (overall score and individual items) Work around the home Not .71–.73 being able to do the things you used to
.69–.74
Emotional function
Psychological needs (overall .56–.74 score and individual items)
.61–.75
Pain Fatigue
Pain Lack of energy/tiredness
.78 .74
.74 .75
Global health/ QOL
Feeling unwell a lot of the time
.73
.76
Role function
Role Function
Work around the home
80
90
.70–.70 Emotional Function
Feelings of sadness
90
100
Global Health/ QOL
Feeling unwell a lot of the time
70
80
Hypothesized AUC \.70 Social function
Not being able to do the things .64 you used to
.68
Sleep
Lack of energy/tiredness
.39–.55
.41–.51
Feeling unwell a lot of the time
Pain
20
10
Fatigue
Being given information…about aspects of managing your illness and side-effects at home
Lack of energy/tiredness
30
20
Cognitive function
Feeling unwell a lot of the time
.54–.60
.53–.63
Nausea/ vomiting
Being given .19–.36 information…about aspects of managing your illness and side-effects at home
.22–.27
Dyspnea
.37–.48
.32–.48
Appetite loss Constipation
.47–.49
.32–.49
.31–.37
.32–.40
Diarrhea
.34–.34
.18–.21
sensitivity and specificity of our original cut-off scores would be supported.; RESULTS All hypotheses were confirmed: 6/6 QLQ-C30 domains with AUC [ .70 from the original analyses had AUC [ .70 in the validation sample (Table 1), with the same SCNS item from the original analysis having the highest AUC in the validation sample. The sensitivity and specificity were generally comparable to the original analysis (Table 2). Examples of cut-off scores (sensitivity, specificity) are: physical function \90 (.85, .65); role function \90 (.85, .62); emotional function \90 (.84, .60); global health/QOL \70 (.86, .56); pain [10 (.93, .54); fatigue [30 (.86, .62).; CONCLUSIONS These results confirm QLQ-C30 cut-off scores associated with patients’ unmet needs. The cut-off scores for these 6 domains could be implemented in clinical practice and their effectiveness evaluated in terms of whether they help clinicians identify patients whose PROs may need attention. Further research is needed to identify appropriate cutoff scores for the remaining QLQ-C30 domains.
123
Pain
.65
.83
Validation .40
.92
Original
.85
.58
Validation .85
.65
Original
.69
.79
Validation .69
.79
Original
.85
.69
Validation .85
.62
Original
.89
.53
Validation .84
.60
Original
.94
.35
Validation .92
.42
Original
.71
.69
Validation .86
.56
Original
.89
.58
Validation .89
.45
Original
.66
.84
Validation .70
.81
Original
.91
.66
Validation .93
.54
Original
.77
.71
Validation .86
.62
Original
.91
.55
Validation .97
.42
(2083) Oral health-related quality of life and self-rated health in middle and older aged Thai adults Tewarit Somkotra, Chulalongkorn University; Vasoontara Yiengprugsawan, The Australian National University, Australia; Sam-ang Seubsman, Sukhothai Thammathirat Open University; Adrian Sleigh, The Australian National University AIMS Oral health is an important public health issue as it could directly impact overall health and quality of life in the population. However, there has been limited evidence on such association in low and middle-income countries and this study aims to address this knowledge gap for Thailand.; METHODS Data were derived from a cohort of 87,134 distance learning adult students aged 15–87 years enrolled at Sukhothai Thammathirat Open University who completed a baseline study in 2005. This study analyses for respondents aged 35 years and older (n = 24,720): number of remaining teeth, experiencing ‘discomfort chewing’ and ‘pain’ and associations between these factors and overall self-rated health. Analysis was carried out using multivariate logistic regression reporting adjusted Odds Ratios (OR) and 95 % Confidence Intervals (CI).; RESULTS Among the Thai cohort members aged 35 and older, 5.8 % reported having teeth less than 20 teeth and 4.2 % reported ‘poor’ overall self-rated health. Experiences with discomfort were most commonly reported in chewing (23.2 %) and pain (10.1 %). Poor overall self-rated health was associated with discomfort chewing (OR = 1.81, 95 % CI 1.56–2.09) and pain (OR = 2.17, 95 % CI 1.82–2.59), after adjusting for socio-geo-demographic attributes and number of remaining teeth.; CONCLUSIONS This study provides strong evidence of the
Qual Life Res (2012) 21:1–132 association between oral health-related quality of life indicators and overall health outcomes. Promoting oral heath in middle age and especially in the later life will thus be vital for improving population health. There is a need for continued integration of oral health as part of holistic public health policy and programs.
(2084) Improving the linguistic validation process of patient reported outcomes instruments into South African languages Mary C. Gawlicki, MBA, Corporate Translations, Inc., United States; Shawn McKown, MA, Corporate Translations, Inc.; Matthew Talbert, Corporate Translations; Barbara A. Brandt, MA, Corporate Translations, Inc., United States AIMS The objective of this research project is to obtain greater knowledge of South Africa Bantu languages. Recently, linguistic validation of Patient Reported Outcomes (PRO) instruments in these languages are required more often. However, this language group presents translation challenges, as their linguistic development and grammatical structure is unique compared to languages of Europe, Asia and the Americas.; METHODS A literature review of South African language history and development was conducted, to gauge whether Bantu languages will continue to be spoken in South Africa, and to determine the necessity of future linguistic validation into such languages. Additionally, linguistic validation data was analyzed, providing insight into the translation process of PRO Instruments in South African Bantu languages and to pinpoint anomalies unique to these languages.; RESULTS Today’s South African Bantu languages evolved from pre-colonial languages that were initially unwritten and had the Roman alphabet imposed upon them by colonists. Consequently, this caused distinctions between the verbal and spoken dialects of these languages. This is evident in analysis of cognitive debriefing data, as subjects recognized certain items that originated from a spoken dialect, and suggested appropriate corrections for the written dialect. Compared to other languages, many more South African Bantu spelling corrections were observed during cognitive debriefing. It was discovered that although some spelling corrections may seem to be a stylistic preference, these should be implemented with caution, as a single letter change can have an effect on words and entire sentences.; CONCLUSIONS Upon completion of the literature review, it appears that South African Bantu languages will continue to gain prominence in South Africa. Therefore, linguistic validation of PROs into these languages is essential in achieving a valid sample of South African respondents. Several linguistic validation anomalies unique to South African Bantu languages were identified. Cognitive debriefing in these languages is imperative to ensure cultural appropriateness, conceptual equivalency, and accuracy of dialect. Additionally, because spelling corrections could affect meaning, further linguistic input is required prior to implementation.
(2085) Limitations of the social relationships domain of WHOQOL-Bref Brigida F. Patricio, MSc, Escola Superior de Tecnologia da Saude do Porto, Instituto Politecnico do Porto, Portugal; Luis M.T. Jesus, Universidade de Aveiro; Madeline Cruice, PhD, Dept of Language & Communication Science, United Kingdom; Andreia Hall, Universidade de Aveiro AIMS The aim of this study is to assess the suitability of the social relationships (SR) domain of the WHOQOL-Bref by comparing it with the WHOQOL-100 SR domain. SR are an essential
101 consideration in clinical rehabilitation after stroke, and especially for those who have aphasia (language impairment).; METHODS Two hundred and fifty-five (n = 255) participants from the general Portuguese population have taken part in this research (mean age 43 years, range 25–84 years; 148 female, 107 male). Participants completed the European Portuguese version of the World Health Organization Quality of Life short-form instrument (WHOQOL-Bref, Serra et al., 2004) and the SR domain of WHOQOL-100 (Canavarro et al., 2009). WHOQOL-Bref SR domain has three items and the WHOQOL-100 SR domain has twelve items. Correlation and regression analysis of quality of life (QOL), and the SR domains of WHOQOL-Bref and of WHOQOL-100 (WHOQOL-100-SR) was undertaken.; RESULTS All WHOQOL-Bref domains were significantly correlated with overall QOL. Correlation strength of WHOQOL-Bref domains in a descending order were: physical domain (0.56), psychological domain (0.50), environment (0.45) and SR domain (0.34). The weakest predictor of overall QOL in WHOQOL-Bref was the SR domain. WHOQOL-100-SR better predicts overall QOL than WHOQOL-Bref SR domain. The item of WHOQOL-Bref most correlated with the overall SR results was F15.3—sexual life (0.82). The WHOQOL-100SR item that most explained the domain results was F13 (SR satisfaction) which explained 89 %, followed by F15 (sexual life), which, together, explained 96 % of the variance of QOL results. Item F13.2 (family relationships) of WHOQOL-100-RS strongly predicted WHOQOL-100-SR results, and was followed by F15.3 (sexual life) and F14.2 (friends support). Together, explained 89 % of the WHOQOL-100-RS results.; CONCLUSIONS The SR domain of WHOQOL-100 better explained overall QOL scores than the WHOQOL-Bref SR domain. Data collection in the clinical subgroup of participants with aphasia is being undertaken now. If these findings are reproducible for people with aphasia (PWA), the WHOQOL-100 SR domain is preferable to the WHOQOL-Bref SR domain when assessing SR among PWA as it will reveal more impact on social relationships and be better understood by clinicians.
(2086) Older adults’ quality of life and adjustment to aging: findings from SF-6D and ATAS-33 Sofia Von Humboldt, MSc, Research Unit in Psychology and Health; ISPA—Instituto Universita´rio, Portugal; Isabel Leal, PhD, UIPES Health & Psychology Investigation Unit, Portugal AIMS To examine the associations of older adults’ quality of life (QoL) with adjustment to aging (AtA).; METHODS For the purposes of this study, the Health Survey Questionnaire (SF-6D), one measure of cognitive functioning—the Mini-Mental State Examination (MMSE) and demographics, were included. Measures were completed using a variety of culturally appropriate methods, including self-administration and interviews. All variables had fewer than 1 % missing values and complete data were available for 709 older adults (M = 84.9; SD = 6.65; range 74–102) from four different nationalities. Exploratory and confirmatory factor analysis were run for data reduction and for exploring theoretical structure. Controlling for age, gender and country of origin, we assessed the level of QoL of elderly people, and its association with AtA.; RESULTS SF-6D score was positively associated with AtA score (r = .295, p \ .001). QoL showed a moderate association with AtA and its impact was also mediated through attitudes towards personal accomplishment and health status. CONCLUSIONS This cross-national study enlightens links among QoL, and AtA in older age. Using developmental frameworks in future research could help us better understand how older adults adapt to aging transitions and add quality to their remaining life years. Keywords: Adjustment to Aging; Old Age; Older Adults; Quality of Life.
123
102 (2087) How does spinal stenosis affect health-related quality of life of Chinese older people? Chi Yan Yeung, The Open University of Hong Kong, Hong Kong; Doris S.F. Yu, The Chinese University of Hong Kong, Hong Kong AIMS To examine the health-related quality of life (HRQoL) and the relating factors in Chinese older people with spinal stenosis, a common degenerative spinal disease.; METHODS This cross-sectional, descriptive, and correlational study recruited 90 patients with spinal stenosis from a specialty out-patient clinic in Hong Kong. The SF-36 Health Survey and the Oswestry Disability Index were used to assess generic and disease-specific HRQoL respectively. Demographic and clinical variables which might influence HRQoL were also collected.; RESULTS A total of 85 participants were included in the analysis. Majority of them were women (64.7 %), had an average of 69.3 years (SD = 7.2), lived with family (68.2 %), had primary education or below (64.7 %), retired (84.75) and obese (84 %). About 43.5 % of them had been diagnosed for 3 years or above and 90 % were categorized as having moderate or more severe levels of disability. They reported lower SF-36 domain scores, particularly on role functioning, health perception, vitality and bodily pain than the norm value for the population in Hong Kong (Fig. 1). The vitality and mental health subscale score reported by them were even comparable to the local population who have psychological problems, however, only 7 % of them reported psychological problems and were seeking expert psychological care. Moreover, the disease resulted in a detrimental impact on their activities of daily living, with the ability in lifting, standing, and maintaining a sexual life and social life were mostly compromised. Correlational analysis indicated that the participants who were female, employed, used regular analgesics, had a higher Body Mass Index, and a greater number of chronic diseases reported poorer HRQoL.; CONCLUSIONS The prevalence of spinal stenosis is rising with the population aging. This study identified the compromised HRQoL among Chinese older people with spinal stenosis in many perspectives, with the role functioning, physical functioning and mental functioning were most affected. Supporting older people to live with this chronic disease need to use a multi-disciplinary approach to optimize patient’s functional ability and enhance their psychological adaptation.
Qual Life Res (2012) 21:1–132 (2088) Association between smoking status and oral health-related quality of life among Japanese adults: LEMONADE study Mariko Naito, DDS, PhD, Nagoya University Graduate School of Medicine, Japan; Kenji Wakai, Nagoya University Graduate School of Medicine; Toru Naito, Fukuoka Dental College, Japan; Haruo Nakagaki, Aichi-Gakuin University; Nobuhiro Hanada, Tsurumi University; Takashi Kawamura, Kyoto University Health Service AIMS Several studies have revealed negative effects of smoking status on oral health such as tooth loss in adults, but little is known about the impact of smoking and its cessation on oral health relatedquality of life (QOL). This study aimed to assess whether smoking status is associated with oral health-related QOL among apparently healthy people.; METHODS The baseline data from the LEMONADE study was analyzed—which is a survey, conducted from 2003 to 2006 on dentists from the Japanese Dental Association. A selfadministered questionnaire requested that respondents provide their number of teeth lost. The questionnaire included the General Oral Health Assessment Index (GOHAI) for measuring oral health-related QOL. Lifestyle data was also collected. Smoking status was categorized as follows: never, former (quit more than 1 year ago), and current smokers. We used a general linear model to compare the mean GOHAI scores among smoking status categories after adjustment for age.; RESULTS Data was obtained from dentists, aged from 26 to 97 years (mean ± standard deviation [SD] = 52 ± 12; women comprised 8.9 % of the sample. The mean QOL score was 54.7 (SD = 6.1). A total of 9,882 dentists were classified: 3,662 never smoked (37.1 %), 3,462 were former smokers (35.0 %) and 2,758 were current smokers (27.9 %). The adjusted mean QOL scores were 55.1 among the never smokers, 54.8 among former smokers and 53.9 among current smokers. In men, smoking decreased QOL among current and former smokers relative to never smokers (P \ 0.001). Differences in QOL scores by smoking status were attenuated with further adjustment for number of teeth lost, but a significant association between smoking status and QOL scores persisted. In women, a marginally significant association between smoking status and QOL scores was found among subjects 60 years or older (P = 0.08).; CONCLUSIONS Among men of all ages and older women, current smokers are more likely, than former or never smokers to have poor oral health-related QOL. Smoking may give negative effects on not only oral function but also on psychological or social factors in oral health.
(2090) Developing and pilot testing a self-report version of the child health care questionnaire on satisfaction, utilization and needs (CHC-SUN) for children and adolescents with chronic health conditions Silke Schmidt, PhD, University of Greifswald, Germany; Carsten Herrmann, Ernst-Moritz-Arndt-University; Holger Muehlan, Ernst-Moritz-Arndt University Greisfwald, Germany, Germany; Franziska Bomba, Unversity of Luebeck; Ute Thyen, MD, Universitaetsklinikum Schleswig–Holstein, Germany
Fig. 1 HRQoL of Chinese older people with spinal stenosis (n = 85)
123
AIMS Originally, the instrument ‘‘Child Health Care-Satisfaction, Utilization and Needs’’ (CHC-SUN) has been developed cross-culturally to evaluate pediatric health care services for children and adolescents with special health care needs (CSHCN) from the proxy perspective of parents. We aimed at establishing a corresponding selfreport version assessing health care satisfaction, utilization and needs from the perspective of children and adolescents concerned with CSHCN.; METHODS Items of the CHC-SUN first were reformulated in first-person answer format. After cognitive debriefing sessions with
Qual Life Res (2012) 21:1–132 four adolescents the items towards financial burden were eliminated due to a lack of understandability. One item was added to check the suitability of the questions toward the time of diagnosis. The link to the resulting online-questionnaire was spread via doctors, selfhelp-groups and Facebook. Adolescents from 16 to 21 years with any chronic condition were allowed to answer the questionnaire (n = 180). Mental-health-conditions and adipositas were excluded. The sample mainly included patients with diabetes type 1 and multiple sclerosis.; RESULTS Preliminary results indicate sufficient descriptive and psychometric performance for selected characteristics of the measure: Acceptance was good as indicated by low rates of missing data. Internal consistencies of Likert-type satisfactionwith-care scales reach at least .75. Validity could be assumed, e.g. as indicated by differential associations with various facets of health literacy.; CONCLUSIONS Results indicate the suitability of a selfreport-version of the CHC-SUN. More research is needed to further explore psychometric performance of this newly developed version regarding selected validity issues and to investigate sensitivity to change within intervention studies and longitudinal clinical trials. Moreover, according to the cross-cultural DISABKIDS approach in establishing the original proxy measure of the CHC-SUN, future work should be directed to translate this self-report version into further languages.
(2091) Crohn’s disease: symptoms, functional limitations and quality of life Stephen P. McKenna, PhD, Galen Research Ltd., United Kingdom; Jeanette Wilburn, Galen Research Ltd; James Twiss, BSc, Galen Research Ltd, United Kingdom; Mitra Ben-L’amri, BSc, MPhil, Galen Research Ltd, United Kingdom; Karen Kemp, Dept of Gastroenterology, Manchester Royal Infirmary, UK; Simon Campbell, Dept of Gastroenterology, Manchester Royal Infirmary, UK AIMS Research into the impact of Crohn’s Disease (CD) and its treatment on patients has employed generic patient-reported outcome (PRO) measures. These are likely to miss important impacts of the illness and ask questions of limited relevance. Qualitative interviews were conducted with CD patients to determine which issues were of greatest importance.; METHODS Interviews were conducted with CD patients attending outpatient clinics at the Manchester Royal Infirmary. The interviews, which were audiorecorded, covered all aspects of the impact of the illness and its treatment. Interviewees were encouraged to discuss any topic they considered relevant. Transcriptions of the interviews were contentanalysed in relation to the WHO International Classification of Functioning, Disability and Health and the needs-based quality of life (QoL) model.; RESULTS Thirty CD patients (18 (60.0 % female; aged 25–68; mean (SD): 47.9 (14.3) years) were interviewed. Participants had a wide range of CD duration (2–40; mean (SD): 14.3 (13.4) years). 2,641 statements relating to the impact of CD were identified. These fell into 3 major categories of outcome, each with a number of sub-themes. These were Symptoms (including; diarrhoea, constipation, pain, fatigue and emotional impairment), activity limitations (such as difficulties with walking, bending, lifting, carrying and doing jobs around the house) and QoL (including preoccupation with the disease, reduced socialisation, hygiene issues, self-image problems and perceptions of lack of control).; CONCLUSIONS The study was successful in identifying a wide range of outcomes of CD, many of which are not addressed by generic PRO measures. To determine whether new interventions are beneficial from the patient’s perspective evidence is needed that they improve the outcomes identified. It is intended to include these issues in a new PRO specific to CD.
103 (2092) Discrepancies in the onset of disease as shown by quality of life (QOL) among patients with Angina Chen Zhihui, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yu Changhe, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Lv Meijun, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Ma Yuan, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Wang Yang, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS Disease onset discrepancies may lead to different QOL. This study compares disease-specific QOL in patients with different contributing factors using the Seattle Angina Questionnaire (SAQ).; METHODS A cross-sectional study was conducted to collect clinical data. The contributing factors of angina were diagnosed by doctors and placed into three etiological categories: (1) overwork, (2) excessive emotions, (3) cloudy or rainy weather. A descriptive analysis of the above variables was performed. Patients were divided into two groups respectively according to whether they were affected by any of three etiological factor or not, and the difference between groups as they affect the SAQ were examined using the Independent-Samples T Test. And Spearman correlation coefficients was assessed between etiological factors and each dimensions in SAQ.; RESULTS 1,574 patients were recruited (age 59.2 ± 10.0 years, male, 48.5 %). Among the three contributing factors, ‘‘overwork’’ was selected in the ratio 79.2 %; ‘‘excessive emotions’’ 68.0 %; ‘‘cloudy or rainy weather’’ 43.6 %. Patients with etiological factors or not show significant differences in global score and each dimension of the SAQ. The SAQ global scores of patients with ‘‘overwork’’ factor or not are 58.5 ± 13.0 and 63.7 ± 13.3 (p \ 0.05); ‘excessive emotions’ 58.0 ± 13.3 and 62.9 ± 12.5 (p \ 0.05); ‘cloudy or rainy weather’ 57.1 ± 13.4 and 61.4 ± 12.6 (p \ 0.05). As to each dimension, different factors related to different dimensions. The correlations coefficients (CC) of patients with ‘‘overwork’’ ranged from 0.076 to 0.212 for each dimensionsis and was more related to ‘‘physical limitation’’ (r = 0.212, p \ 0.01); for ‘‘excessive emotions’’, the CC ranged from 0.047 to 0.205 for each dimensions and was more related to ‘‘treatment satisfaction’’ (r = 0.205, p \ 0.01); for ‘‘cloudy or rainy weather’’, the CC ranged from 0.013 to 0.193 for each dimensions and was more related to ‘‘disease perception’’ (r = 0.193, p \ 0.01).; CONCLUSIONS The etiological factors affect subjects in a way they performed in their QOL outcomes, this may due to the association between etiological factors and syndromes, a further study about it will be discussed in another paper.
(2093) Comparisons of psychometric properties of the WHOQOL-BREF between mainland China and Taiwan in early adolescence Grace Yao, PhD, Dept of Psychology, Nat’l Taiwan Univ, Taiwan; Zhaoming Guo, College of Educational Science, Yangzhou University, Mainland China AIMS This study compared psychometric properties of the WHOQOLBREF in early adolescent samples between Mainland China and Taiwan.; METHODS Data were collected from 627 junior high school students in Mainland China and 365 junior high school students in Taiwan. Correlation analysis, regression analysis and confirmatory factor analysis were conducted.; RESULTS After deleting three physical domain items (i.e., ‘‘To what extent do you feel that your physical
123
104 pain prevents you from doing what you need to do?’’; ‘‘How much do you need any medical treatment to function in your daily life?’’; ‘‘How well are you able to get around?’’) and one social domain item (i.e., ‘‘How satisfied are you with your sex life?’’), the results from both samples supported the originally four-factor model design, and also showed good item discrimination, internal reliability, content validity, construct validity and predictive validity. In Mainland China and Taiwan versions, the internal consistency (Cronbach’s Alphas) coefficients were 0.62–0.82 and 0.59–0.79, respectively. Content validity coefficient were 0.37–0.68 and 0.33–0.64 for corrected item-domain correlations; and 0.53–0.75 and 0.42–0.62 for inter-domain correlations (all p \ 0.01), respectively. Fit indices revealed that this hierarchical model was acceptable (v2(166) = 653.00, p \ 0.01; NFI = 0.87, NNFI = 0.88, CFI = 0.90, IFI = 0.90, RMR = 0.038; and v2(166) = 452.95, p \ 0.01; NFI = 0.81, NNFI = 0.85, CFI = 0.87, IFI = 0.87, RMR = 0.062). Two-group analyses with unconstrained parameters supported the originally four-factor model design (v2(332) = 1,106.00, p \ 0.01, NFI = 0.85, NNFI = 0.87, IFI = 0.89, CFI = 0.90, RMR = 0.051). But the v2 discrepancy tests showed that two groups did not share the same the first-order factor loadings. The four domains can explain 42 and 34 %of the variance of the Facet G score (measuring overall quality of life and general health).; CONCLUSIONS Psychometric properties of the WHOQOL-BREF for both places in early adolescent samples were agreeable and acceptable.
(2094) Health-related quality of life in persons with COPD attending self-management courses and pulmonary rehabilitation compared with the general population Randi Andenas, PhD, Oslo and Akershus University College, Norway; Kari Hvinden, Glittreklinikken; May Solveig Fagermoen, Oslo University Hospital, Department of Gastroenterology; Anners Lerdal, Lovisenberg Deaconale University College AIMS Chronic obstructive pulmonary disease (COPD) can have a pervasive effect on all areas of life. The disease may cause life-long medical treatment, hospitalizations, absence from work, and activity limitations, and negatively impact psychological well-being and HRQoL. The objective of this study was to explore relationships between socio-demographic variables, paid work, and physical activity, with physical and mental health variables in persons with COPD, and to compare their health-related quality of life (HRQoL) scores with scores from the general population.; METHODS A crosssectional correlational study design was used. Data were collected by self-reported questionnaires from participants within the first 2 days of commencement of educational and rehabilitation courses. Participants from both in- and out-patient rehabilitation were included in the study. HRQoL was measured with the Short Form 12v2 from which physical (PCS) and mental component summary (MCS) scores were computed.; RESULTS Of 127 course attendees, 100 (78.7 %) volunteered to participate in the study. Valid responses on all items were recorded for 88 participants. Mean age was 64.3 years (SD 9.7); 53.4 % were male and 60.2 % lived with a partner. Men reported higher HRQoL than women on four of the SF 12-dimensions and with clinical significant effect sizes; this being bodily pain (d = 0.51), vitality (d = 0.45), emotional role (d = 0.55), and mental health (d = 0.45). Men also reported higher PCS than women (d = 0.45) (all p \ 0.05). Analyzes of bivariate relationship showed that younger participants and those in paid work reported higher PCS scores than those with higher age and not in paid work. Being male, and being physically active were associated with higher MCS. Except for Mental Health, respondents scored lower on all HRQoL subdomains compared with norms.; CONCLUSIONS Being younger, male gender, physical active, and participation in paid work are factors related
123
Qual Life Res (2012) 21:1–132 to HRQoL in this COPD population. Physical HRQoL is reduced compared to the general population.
(2095) Development of a checklist to comply with the FDA review requirements for the translation of PRO instruments Isabelle Mear, Mapi Research Institute, France; Caroline Anfray, MAPI Institute, France; Katrin Conway, MA, MAPI Research Trust, France AIMS In its guidance on the use of PRO measures, the FDA specifies the areas to be addressed in PRO documents provided for review. Regarding language translation and cultural adaptation, four areas are listed: (a) process used to translate and culturally adapt the instrument for populations that will use them in the trial; (b) description of patient testing, language- or culture-specific concerns, and rationale for decisions made to create new versions; (c) copies of translated or adapted versions; and (d) evidence that content validity and other measurement properties are comparable between the original and new instruments. The objective of this study is to present the development of a checklist to organize the evidence generated during the translation of a PRO instrument to comply with the FDA review requirements.; METHODS (1) Review of the process used to translate PRO instruments, and of the evidence provided during the translation process; and (2) organization of the evidence according to the four areas listed in the FDA guidance (excluding measurement properties).; RESULTS The checklist is organized in a table form divided into 3 parts. Part 1 (translation background information) gathers the information requested in areas (a) and (c). Part 2 (translation report) provides the evidence required in area (b). Part 3 (content validity) concerns the comparability of content validity. Information about other measurement properties is not included in the checklist since it only covers the translation process. For the content validity part, it was assumed that comparability will depend on: (1) a clear definition and understanding of the concepts to be translated; and (2) the involvement of trained professionals. Hyperlinks are provided in each part of the table to lead to the evidence documents (e.g., concept list, rationale for decisionmaking, reports, demographic information of respondents, participants’ CVs, etc.) required for each area listed by the FDA.; CONCLUSIONS The checklist is a unique tool which will enable researchers to submit a consistent and comprehensive translation dossier for FDA review.
(2096) Patient reported outcome measures: a revised model-based classification system for research and clinical practice Jose M. Valderas, MD, PhD, MPH, University of Oxford, United Kingdom; Daniela C. Gonc¸alves, University of Oxford, United Kingdom; Jordi Alonso, MD PhD, IMIM-Hospital del Mar Medical Research Institute, Spain AIMS We developed a classification system for Patient Reported Outcome (PROs) measures based on an integrated theoretical framework (models: Wilson and Cleary (W&C) and the International Classification of Functioning Disability and Health (ICF)) and including 3 components: construct, population and measurement model. Here we aimed: (a) to refine the construct component of the classification (including: ‘‘Symptoms’’, ‘‘Functional status’’, ‘‘General Health Perceptions’’, and ‘‘Health related quality of life’’), and (b) to explore its integration with other existing classification systems, including the PROMIS Domain Framework and the WHO Family of International Classifications.; METHODS Model representations and full
Qual Life Res (2012) 21:1–132 specifications for PROMIS and WHO classifications, along with selected widely used instruments (SF-36, EQ5D, and others) were reviewed. The authors then engaged in an iterative process of critical review, adopting a sceptical perspective on both the elements of the model as well as the nature and direction of the associations between them.; RESULTS It was feasible to integrate the construct axis with WHO definition of health and the PROMIS Domain Framework. Physical, mental and social health are now identified as subcategories across all categories, except for ‘‘Symptoms’’, which does not include social health. ‘‘Symptoms’’ and ‘‘Functional status’’ have been expanded to include affect and cognition. The model now acknowledges both the feed-back loop of ‘‘Functional status’’ on ‘‘Symptoms’’ and the possibility of ‘‘Symptoms’’ impacting on ‘‘General Health perceptions’’ without any mediation through ‘‘Functional Status’’. The link with WHO ICD-10 has been made explicit by identifying: ‘‘Symptoms’’, ‘‘Functional status’’, alongside ‘‘Biological and Physiological Variables’’ as the relevant information contributing to diagnoses. The previous linear representation of the categories has been substituted by a matrix in order to account for different layers of measurement along a objective-subjective continuum (e.g.: frequency, intensity, difficulty, bother, satisfaction). We exemplify the use of the classification system with selected PRO instruments.; CONCLUSIONS This classification system of PRO measures, based on a valid integrated conceptual model (W&C, WHO Health, WHO ICD, WHO ICF, PROMIS), should allow the classification of most currently used instruments and may facilitate a more adequate selection and application of these instruments.
(2097) A comprehensive review of the validation and clinical application of the Infant Dermatitis Quality of Life (IDQoL) Index Sam Salek, BSc RPh PhD FFPM, Cardiff University, Redwood Bldg, United Kingdom; Vrutti Gada, Cardiff University; Mohammad K.A. Basra, MD, Cardiff University, United Kingdom; Andrew Finlay, FRCP, Univ. of Wales College of Medicine, United Kingdom AIMS The IDQoL, assessed by parents, measures the impact that Atopic Dermatitis (AD) has on the quality of life (QoL) of infants aged between 0 and 4 years.1 The aim of this study was to appraise all clinical and psychometric data on the use of IDQoL from its inception in 2001 until 2011, in order to create a one-stop reference source.; METHODS A detailed literature search was carried out using the Cardiff University Dermatology Department library, Medline, PubMed and Google Scholar. Articles and abstracts that described the clinical use of the IDQoL and those that investigated its psychometric properties, were included. Articles not in English were excluded. Data were extracted and tabulated using pre-designed templates.; RESULTS 35 articles were identified of which 32 fulfilled the inclusion criteria. Five aspects of the IDQoL were studied: psychometric; descriptive; clinical practice research; drug trials; and therapeutic interventions. The IDQoL has been translated into 15 languages and used in 14 countries as well as being included in two multinational studies. 25 of these studies demonstrated psychometric properties such as test–retest reliability, validity, responsiveness to change and interpretability, though there is no valid score description system yet published. None of the studies investigated the dimensionality and internal consistency of the tool. Eight studies used the IDQoL to assess the effectiveness of therapeutic interventions such as education programmes, consultations and wetwrap therapy. Six studies used the IDQoL in drug trials. No studies focused on its use in clinical practice research. The IDQoL has been used together with other instruments such as the Dermatitis Family Index and a severity assessment tool in most studies.; CONCLUSIONS The findings indicate that IDQoL is a reliable and valid measure that has been used extensively worldwide. This unique review serves as a single reference source allowing potential users of the IDQoL to make an
105 informed decision regarding its use in their clinical studies as well as in routine practice.
(2098) The dermatitis family impact instrument: an evaluation of its psychometric properties and clinical application Sam Salek, BSc RPh PhD FFPM, Cardiff University, Redwood Bldg, United Kingdom; Sean Dodington, Cardiff University; Mohammad K.A. Basra, MD, Cardiff University, United Kingdom; Andrew Finlay, FRCP, Univ. of Wales College of Medicine, United Kingdom AIMS The Dermatitis Family Impact (DFI) questionnaire is a diseasespecific measure to assess the impact of atopic eczema on the quality of life (QoL) of parents and family members of affected children. The aim of this study was to appraise the published articles reporting clinical and psychometric aspects of the DFI questionnaire from its development in 1998–2011 in order to create a single reference source.; METHODS A comprehensive literature search covering 1998–2011 was carried out using Pubmed and Google Scholar as well as a manual search through the Cardiff University Department of Dermatology library and Inter-library journal requests.; RESULTS 45 articles were identified reporting the use of the DFI (29 full articles and 16 abstracts/summaries). Only one study demonstrated test–retest reliability for the DFI (r = 0.95). 3 studies reported the internal consistency of the DFI with Cronbach’s alpha values ranging from 0.85 to 0.90. 14 studies showed sensitivity to change with significant differences between baseline and the end of study. The DFI was compared with other objective and subjective parameters in 22 studies demonstrating its construct (convergent) validity. Currently, there are 19 validated translations of the DFI that have been used in 15 different countries (including 2 multinational studies). The DFI has been used in 9 clinical studies assessing the effectiveness of 5 different topical drugs, 1 probiotic supplement, and in studies of the effectiveness of the care of dermatology nurses and dermatologists.; CONCLUSIONS This work brings together previously scattered DFI validation data supporting its use in both clinical settings and research. The findings also highlight areas requiring further examination.
Saturday Poster Session (3001) Psychometric properties of the KIDSCREEN-27 in Brazilian children and adolescents Silvia M S Rodrigues, MD, MSc, Universidade Federal do Para´, Brazil; Janari S. Pedroso, Universidade Federal do Para´; Fernando Augusto R. Pontes, Universidade Federal do Para´; Christoph O. Ka¨ppler, Technische Universita¨t Dortmund; Celina Maria C. Magalha˜es, Universidade Federal do Para´, Brazil AIMS The aim of this study was to perform the cultural adaptation and identify psychometric properties of the questionnaire KIDSCREEN-27 for the children and adolescents Brazilian population. This instrument was developed in Europe to assess the quality of life related to health of children and adolescents.; METHODS The KIDSCREEN-27 was one of the instruments used in a transcultural and cross-sectional study which aimed to evaluate perspectives about the mental health of Brazilian children and adolescents whose version in Portuguese Lusitanian, was adapted for use in the Brazilian population. The final questionnaire was administered to a sample of 1,115 children and adolescents of both sexes and 489 parents/caregivers of four capital cities from clinical and schools contexts. To identify psychometric properties, the questionnaires were evaluated by indicators for internal consistency and construct, discriminant and convergent validity.; RESULTS The questionnaire KIDSCREEN-27
123
106 was able to discriminate by gender, age groups, clinical and schools settings in its five dimensions. The exploratory and confirmatory factor analyzes confirmed the structure of the five theoretical dimensions. In exploratory analysis the model of five factors explained 56.87 % of the total variance with satisfactory values of internal consistency. The Cronbach a showed coefficients between 0.77 and 0.86. The confirmatory factor analysis showed goodness of fit indices RMSEA 0.077 and CFI 0.838 for the model similar to the original validation. The convergent validity was obtained in five dimensions, with correlation coefficients that represent the monotrait-heteromethod significantly different and greater than 0 in the five dimensions (range of r from 0.252 to 0.424, p \ 0.01).; CONCLUSIONS The translation, cultural adaptation and psychometric properties of KIDSCREEN-27 showed acceptable coefficients of reliability and validity, similar to the original version. This allows the Brazilian researchers have an instrument to assess the quality of life and, simultaneously, to compare their results with other countries using this instrument.
(3002) Applicability of the WHOQOL-BREF on early adolescence in China Zhaoming Guo, College of Educational Science, Yangzhou University, Mainland China; Grace Yao, PhD, Dept of Psychology, Nat’l Taiwan Univ, Taiwan AIMS The purpose of this study was to examine the applicability of the WHOQOL-BREF China version in early adolescent sample.; METHODS A total of 627 junior high school students participated in this study. Correlation analysis, regression analysis and confirmatory factor analysis were used.; RESULTS After deleting three physical domain items (i.e., ‘‘To what extent do you feel that your physical pain prevents you from doing what you need to do?’’; ‘‘How much do you need any medical treatment to function in your daily life?’’; ‘‘How well are you able to get around?’’) and one social domain item (i.e., ‘‘How satisfied are you with your sex life?’’), the results supported the originally four-factor model design, and also showed good item discrimination, internal reliability, content validity, construct validity and predictive validity. The internal consistency (Cronbach’s Alphas) coefficients ranged from 0.62 to 0.82 for the four domains. Content validity coefficient were in the range of 0.37–0.68 for corrected item-domain correlations and 0.53–0.75 for inter-domain correlations (all p \ 0.01). Except for Chi-square test (v2(166) = 653.00, p \ 0.01), fit indices revealed that this hierarchical model was acceptable (NFI = 0.87, NNFI = 0.88, CFI = 0.90, IFI = 0.90, RMR = 0.038). The four domains can explain 42 % of the variance of the Facet G score (measuring overall quality of life and general health).; CONCLUSIONS This study result supports the applicability of the WHOQOL-BREF China version in early adolescence.
(3003) Evaluation of the measurement properties of the Children’s Dermatology Life Quality Index (CDLQI) Sam Salek, BSc RPh PhD FFPM, Cardiff University, Redwood Bldg, United Kingdom; Soora Jung, Cardiff University; Mohammad K.A. Basra, MD, Cardiff University, United Kingdom; Andrew Finlay, FRCP, Univ. of Wales College of Medicine, United Kingdom AIMS The evidence suggests that skin diseases have a major impact on the quality of life (QoL) of children. The Children’s Dermatology Life Quality Index is the most widely used dermatology-specific QoL measure for children aged from 5 to 16 years. The aims of this study were to appraise the psychometric properties of the CDLQI from its
123
Qual Life Res (2012) 21:1–132 inception in 1995–2011, in order to create a single reference source and identify aspects requiring improvement.; METHODS A comprehensive literature search was performed using Pubmed and Google Scholar.; RESULTS 81 published articles fulfilled the inclusion criteria and were reviewed for psychometric data. 51 described validity, 16 sensitivity to change and 5 test–retest reliability. For test–retest reliability studies, Spearman rank correlation coefficient was used as the key statistical test, yielding an average of r = 0.882 (p \ 0.01). Sensitivity to change of the CDLQI was mainly performed using the paired t-test and Wilcoxon signed rank test: only 2 out of 16 trials failed to show statistically significant change. In terms of convergent validity, 21 studies used Spearman’s rank correlation coefficient and 22 used Pearsons correlation. Cross-cultural adaptations were described in 5 articles. Internal consistency reliability was carried out in response to comparison of validation with other language versions of the CDLQI and 3 publications reported high correlations between item scores. None of the studies tested the factor structure or differential item functioning (DIF) of the CDLQI. Only one article described the clinical meaningfulness of the CDLQI scores.; CONCLUSIONS The findings indicate a steady increase in the use of the CDLQI over the last 17 years. This review has brought together previously dispersed data concerning the psychometric properties of the CDLQI, adding to its strength as a robust tool to meet the needs of both clinicians and researchers. It has also identified certain aspects of the CDLQI validation that require further examination including its dimensionality, DIF and clinical meaningfulness.
(3004) Brazilian children and adolescents infected with HIV: the initial stage of development an instrument of health-related quality of life—DISABKIDS group Maria Fernanda Cabral Kourrouski, University of Sa˜o Paulo, Brazil; Isabella S. Abreu, Universidade de Sa˜o Paulo, Brazil; Ana Carolina Gomes Martins Oliveira, University of Sa˜o Paulo, Brazil; Danielle Maria S S dos Santos, MD, University of Sao Paulo, Brazil; Keila C. Deon, MD, University of Sa˜o Paulo, Brazil; Maria Ce´lia Cervi, University of Sa˜o Paulo; Roberta A. Reis, PhD, Universidade Federal do Rio Grande do Sul, Brazil; Lucila Castanheira Nascimento, University of Sa˜o Paulo, Brazil; Claudia Benedicta dos Santos, MISPEERP-USP, Germany; Regina Aparecida Garcia de Lima, University of Sa˜o Paulo, Brazil AIMS To present the first phase of development an instrument of Health related Quality of Life to Brazilian children and adolescents living with HIV using the methodology of the DISABKIDS group.; METHODS Methodological study with data collection carried out from focus groups with children, adolescents and their caregivers, at a specialized service to treatment for people with HIV, at a university hospital in the state of Sao Paulo—Brazil. We conducted eight focus groups with children and adolescents between 8 and 18 years of age and their parents or caregivers. The instrument used to conduct the groups was the same for children, adolescents and their caregivers, which was systematized by the DISABKIDS group. The focus groups’ interviews were recorded, transcribed and organized according to the domains physical, mental and social, using the card sorting method.; RESULTS Children and adolescents reported their experiences in a competent manner. Content concepts related to defining the dimensions of quality of life, according to the DISABKIDS Group were identified.. In relation to their condition, caregivers, children and adolescents referred not commenting about of disease with friends and at school, because they have fear of stigma and discrimination. In relation to their life, they mentioned the electronic media importance and the fact the disease doesn’t cause physical limitation. It was reported by adolescents, who feel the lack of not having a family consisting of father, mother and siblings because of orphaned by AIDS, a fact that did not appear in the
Qual Life Res (2012) 21:1–132 caregivers’ reports. In relation to their treatment, they did not like to come to the hospital every month, as well as the treatment to be for their whole life.; CONCLUSIONS The focus group was effective for data collection, favoring free speech, thoughts, attitudes and discussions from the perspective of its participants, fostering understanding of the experiences, perceptions, and coping with the disease and treatment. The focus group interviews with experts and literature review techniques are appropriate methods for data collection to construct the instrument, to draw up a specific module DISABKIDS related to quality of life for Brazilian children and adolescents infected with HIV.
(3005) Effect of resistance exercise on quality of life of mothers of children and adolescents with cerebral palsy Nivea Morales, MD; MSc; PhD, Federal University of Uberlandia, Brazil; Moise´s Rosa Batista, Sr., Professor, Universidade Federal de Uberlaˆndia, Brazil; Jaqueline Pontes Batista, Federal University of Uberlandia; Jussara Caetano Furtado, Federal University of Uberlandia; Jorge Guto Mendes, Federal University of Uberlandia; Hygor Nunes Araujo, Federal University of Uberlandia; Eduardo Henrique Tavares, Federal University of Uberlandia; Luiz Duarte Rocha, Jr., Federal University of Uberlandia; Roge´rio M C Pinto, Federal University of Uberlandia; Carlos Henrique M. Silva, Federal Univesrity of Uberlandia AIMS Cerebral palsy (CP) is considered the most common physical disability in childhood and involves not only the affected individuals, but also impairs many aspects of life of their caregivers. The objective is to evaluate the effect of a resistance exercise program on quality of life (QL) of mothers caring for children and adolescents with CP.; METHODS Twenty-two healthy and sedentary mothers participated in the intervention, which consisted of two weekly sessions of resistance exercise of moderate intensity for 12 weeks. Two questionnaires were completed through interviews before and after intervention: 36-item Short Form Questionnaire (SF-36) and Caregiver Burden Scale (CBS). The scores obtained before and after intervention were compared using the Wilcoxon test and the magnitude of differences was measured by effect size. Reliability was checked by Cronbach’s alpha coefficient.; RESULTS Maternal age ranged from 18 to 58 years (median 41 years). The median age of children/adolescents was 14 years. The Cronbach’s alpha coefficient ranged from 0.70 to 0.90 for the SF-36 and 0,50–0,76 for the CBS. The SF-36 scores were significantly higher after resistance exercise intervention in six domains and the physical component (p \ 0, 05), with effect size from 0.2 to 1.0. CBS scores were significantly lower after the intervention (p \ 0,05), with effect size of 0,78.; CONCLUSIONS The regular exercise resistance improved the QL and perceived burden of mothers caring for children and adolescents with CP. The results encourage physicians and rehabilitations professionals to propose this kind of intervention for these mothers.
(3006) Young adults with disability benefits Eefje Verhoof, Emma Childrens Hospital; Heleen Maurice-Stam, MSc, Emma’s Children’s Hosp, Academic Medical Ctr, Netherlands; Hugo Heymans, Emma Childrens Hospital; Martha Grootenhuis, PhD, Academic Medical Center, Netherlands AIMS A growing number of young adults with a somatic disease or disability since childhood apply for disability benefits. Despite careful guidance and support, many of them are not successfully integrating into the labour market. The purpose of this study was to assess the health related quality of life (HRQoL), anxiety and depression of young adults with disability benefits because of a somatic disease/ disability compared to reference groups from the general Dutch
107 population and to explore factors related to their HRQoL, anxiety and depression.; METHODS Young adults (22–31 years, N = 377) with disability benefits because of a somatic disease/disability, completed the RAND-36 and the HADS online. Differences between respondents and both reference groups were tested using analysis of variance and logistic regression analysis by group and age (and gender). Regression analyses were conducted to predict HRQoL (Mental and Physical Component Scale; RAND-36) and Anxiety and Depression (HADS) by demographic and disease related variables.; RESULTS The respondents reported worse HRQoL than the reference group, and a higher percentage of them was at risk for an anxiety and depressive disorder. HRQoL, anxiety and depression were associated with the course of the disease and the use of medical devices.; CONCLUSIONS (Paediatric) health care providers should pay systematic attention to socio-emotional functioning of patients growing up with a somatic disease/disability in order to optimise their well-being and adaptation to society. Future research might focus on identifying those most likely to develop difficulties and who would take advantage of psychosocial support, specific to workforce participation.
(3007) Family burden and quality of life of parents of children with a diagnosis of congenital anomaly: the mediating role of parenting stress Marco Pereira, Universidade de Coimbra, Portugal; Sara Albuquerque, University of Coimbra; Ana Fonseca, University of Coimbra; Maria Cristina Canavarro, Universidade de Coimbra, Portugal AIMS The occurrence of a pre- or postnatal diagnosis of a congenital anomaly may imply changes in the family dynamics and functioning, may imply additional requirements at the caregiving level, and may also have an important impact on the parents’ quality of life (QoL). The present study aimed to assess the perceived burden and QoL of parents whose child has a congenital anomaly, and to examine whether this association is mediated by parents’ experience of parenting stress.; METHODS The sample comprised 134 parents, whose child has a congenital anomaly. The assessment protocol included the Impact on Family Scale (IOF), the World Health Organization quality of life instrument (WHOQOL-Bref), and the Parenting Stress Index— Short Form (PSI-SF).; RESULTS Mothers reported higher lower scores on all domains quality of life than fathers, and also reported higher scores on parenting stress and stress in the interaction with the child. No differences were found regarding the parental burden. Controlling for the covariates, the results from mediation models revealed that parenting stress mediated the association between parental burden and parental QoL, particularly the associations between parental burden and Psychological (point estimate = -0.38; CI = -0.70/-0.22), and Social relationships (point estimate = -0.50; CI = -0.85/-0.28) domains.; CONCLUSIONS Findings from this study highlight the importance of understanding the role of internal mechanisms (e.g., parenting stress) in QoL of parents of children with a congenital anomaly, and support the hypothesis that parenting stress is one mechanism through which parental burden decreases parents’ QoL.
(3008) Self and proxy evaluation of HRQOL in Turkish epileptic children by using KINDL Dilek Ergin, celal bayar university, Turkey; Erhan Eser, MD, Celal Bayar University, School of Medicine, Dept. Public Health, Turkey; ¨ niversity; Pinar Yazici, Celal Bayar Muzaffer Polat, Celal Bayar U University
123
108 AIMS Quality of life of assessment has been a popular supplementary method in clinical monitoring of epileptic children. However self and proxy ratings were not adequately compared in the literature. This study aimed to present the consistency between self and proxy QOL assessments in epileptic children.; METHODS Self and proxy (mothers) ratings of KIDNL generic form were obtained from 106 epileptic children (aged 4–16; mean age 9.1) in outpatient clinics of a university hospital in western Turkey. Type1 error level was accepted as 0.05 in the statistical analyses.; RESULTS Of the children, 62.3 % live without support and epileptic seizures were frequent than in every 30 days in 18.9 % of the children. Of the children, 68.9 % have generalized,17.0 % partial and 14.2 mixed type (generalized plus partial seizures) epilepsy and 67.0 % live in a low social class family. Income level is the most sensitive variable on the HRQOL score for both children and their proxies. Disease specific variables such as duration of illness, seizure frequency and type of seizures mostly make a negative effect on physical, psychological, self-respect and school domains of the KINDL. Mothers’ education level has a significantly positive effect on the friends and school domains of the scale. Family support has a positive effect of on psychological self rated score whereas mothers proxy ratings targets the importance of family support on the school domain score. Moderate levels of consistencies are obtained for overall (r = 0.58); physical (r = 0.42); psychological (r = 0.47); self-respect (r = 0.54) and friends (r = 0.47) domains between self and proxy ratings. Very poor consistencies were detected for friends (r = 0.35); school (r = 0.21) and chronic illness (r = 0.28) domains.; CONCLUSIONS KINDL generic version can be regarded as a sensitive instrument for use in the epileptic children. Self rating assessment is suggested rather than a proxy rating where possible.
(3010) Psychometric properties of KIDSCREEN-27 among childhood cancer survivors and age matched peers—a Rasch analysis Anna Jervaeus, Karolinska Institutet, Sweden; Anders Kottorp, Karolinska Institutet; Lena Wettergren, PhD, Karolinska Institute, Sweden AIMS To evaluate the psychometric properties of KIDCSREEN-27 for use among survivors of childhood cancer and age matched peers.; METHODS KIDSCREEN-27 is an generic instrument, consisting of five dimensions (Physical Well-being; Psychological Well-being; Autonomy & Parent Relations; Social Support & Peers; School Environment) measuring health-related quality of life (HRQoL) in healthy and chronically ill children and adolescents. In this study, it was completed by 63 survivors, aged 12–22, and a comparison group (n = 257), aged 11–23. KIDSCREEN-27 was evaluated using a Partial Credit Model (PCM), a Rasch model developed for polytomous data. The aspects studied were the properties of the rating scale, internal scale validity, person response validity, unidimensionality and separation reliability.; RESULTS The rating scales revealed expected patterns of responses. The item goodness-of-fit demonstrated acceptable MnSq values; person goodness-of-fit showed some mixed results in which three dimensions displayed misfitting patterns of person responses above the set criterion of 5 %. Ceiling effects were present for all dimensions and floor effects in one dimension (Social Support & Peers). The explained variance within each dimension was above the set criterion (50 %) for all dimensions except Autonomy & Parent Relations (39.9 %). Separation index showed non-satisfactory values for two dimensions thus, unable to distinguish statistically different subgroups in the sample.; CONCLUSIONS This Rasch analysis revealed that KIDSCREEN-27 works reasonably well for childhood cancer survivors and an age matched comparison group. However, the instrument was not capable of discriminating participants according to perceived
123
Qual Life Res (2012) 21:1–132 HRQoL in two of the dimensions. This might have been because a rather small sample was used, which may have lacked representativeness due to the number of diagnoses included. Furthermore a relatively short time had passed after diagnosis, given that the incidence of health problems increases as time passes. Additionally, we do not know if those who chose to participate (both survivors and the comparison group) were better off regarding socioeconomic factors than those declining participation. Therefore it is recommended that future research should include a larger sample of childhood diagnoses in order to demonstrate different levels and patterns of HRQoL among the item scores in KIDSCREEN-27, and also the impact of time.
(3011) Quality of life in caregivers who take care of family patients with hemodialysis Takashi Mandai, MD, Japanese Society of Quality of Life Research, Japan; Hiromasa Hirai, Japanese Society of Quality of Life Research, Sumaku, Kobe, Hyogo, JAPAN; Takashi Nakano, Japanese Society of Quality of Life Research, Japan; Satio Kawanishi, Japanese Society of Quality of Life Research, Sumaku, Kobe, Hyogo, JAPAN; Kunio Nakayama, Japanese Society of Quality of Life Research, Japan; Yuri Kitamura, Japanese Society of Quality of Life Research, Sumaku, Kobe, Hyogo, JAPAN; Rika Hayashida, MS, Siebold Univ of Nagasaki, Japan; Michiko Kobayashi, MD PhD, Kwassui Women’s College, Japan; Atsuhiko Fukuoka, Japanese Society of Quality of Life Research, Sumaku, Kobe, Hyogo, JAPAN; Kozaburo Adachi, PhD, Agape-Kabutoyama Hospital, Japan AIMS The purpose of this study was to investigate the quality of life (QOL) in caregivers who took care of family patients with hemodialysis.; METHODS Sixty-two caregivers participated in this study. Our new original self-administered QOL questionnaire including 40 questions divided into 13 categories with 16 caregivers who took care of family patients with hemodialysis specific questions was used.; RESULTS Cronbach’s alpha coefficients of our questionnaire were excellent enough to accept for clinical use: 0.87 in mental function, 0.85 in dietary problems, 0.83 in economical condition, etc. before taking care of family patients with hemodialysis, and 0.87 in dietary problems, 0.83 in economical condition, 0.81 in social participation, etc. after one, respectively. Our QOL questionnaire contained 12 main factors and cumulative contribution was 0.85 before taking care of family patients with hemodialysis. Our questionnaire also contained 13 main factors and cumulative contribution was 0.87 after one. There was significant deterioration in mean QOL of caregivers after taking care of family patients with hemodialysis, compared with before one (p \ 0.001). Compared with before taking care of family patients with hemodialysis, significant deteriorations of mean caregivers’ QOLs were demonstrated in 10 categories, i.e. in well-being (p \ 0.05), in dietary problems (p \ 0.05), in work performance (p \ 0.05), etc. after one. There was significantly positive correlation between mean caregivers’ QOLs before taking care of family patients with hemodialysis and mean caregivers’ QOLs after taking care of family patients with one (r = 0.89, p \ 0.01).; CONCLUSIONS These findings indicate that our QOL questionnaire has excellent enough reliability and potency of validity to investigate the QOLs in caregivers who take care of family patients with hemodialysis. The QOLs in caregivers who take care of family patients with hemodialysis were significantly deteriorated by taking care of family patients with hemodialysis. Furthermore these findings also show the potency of prediction on caregivers’ QOL changes by taking care of family patients with hemodialysis in future. Excellent QOLs are indispensable for both patients and caregivers who take care of family patients to continue the excellent medical treatment and care, so we must pay more attention to improve the QOLs not only in patients but also in caregivers who take care of family patients with hemodialysis.
Qual Life Res (2012) 21:1–132 (3012) Differences in mothers’ and fathers’ quality of life (QoL) of children treated for acute leukemia (AL) Patricia Minaya Flores, MA, Laboratoire de Sante Publique EA3279, France; Ge´rard Michel, Pediatric Haematology Departments, University Hospital of Marseille, France; Julie Berbis, Department of Public Health, Medicine University, France; Pascal Auquier, PhD, Laboratoire de Sante Publique, France AIMS Parents of children who survived after AL were very involved in their child’s care. During treatment time their usual activities may have been modified, leading to physical and psychological affliction and impacting on their QOL. Even once treatment completed, the disease and its treatment could have durably shifted the parents’ life. QoL evaluation is necessary in this population, but when QoL is assessed in caregivers who are parents, not always father and mother’s evaluations are accessible. Sometimes, only one respondent complete questionnaire while in other cases both, the father and the mother, answer a same evaluation. Decisions have to been made about which respondents to consider. The purpose of this study was to analyze differences between mothers’ and fathers’ QoL assessments after treatment for AL.; METHODS The sample was composed of caregivers of children treated for acute leukemia recruited from 7 cancer centers in France issued from the French National LEA Cohort. We used the specific CareGiver Oncology Quality of Life questionnaire (CarGOQoL), a self-administered instrument comprising 29 items which is based on the exclusive point of view of caregivers to assess the impact of cancer and its treatment on caregivers’ QoL. Sociodemographic and clinical data were collected. Multiple logisticregression models were used for analyses.; RESULTS Final results on the total sample will be available in September 2012. Currently, a total of 100 mothers, 50 fathers and 120 mixed parents of children aged 2–17 years have completed the specific CarGOQoL.; CONCLUSIONS Statistical analysis is ongoing, but the very important size of our sample allows to expect interesting results concerning differences between mothers and fathers QoL assessments that would be take into account when an evaluation is needed.
(3013) The QOL of dementia caregivers Michiko Kobayashi, MD PhD, Kwassui Women’s College, Japan; Yuzaburo Hirata, Matsuura City Office, Matsuura, Nagasaki, Japan; Takashi Mandai, MD, Japanese Society of Quality of Life Research, Japan AIMS We surveyed the QOL of people taking care of dementia patients to see what is needed for home care of dementia patients.; METHODS We surveyed 57 dementia patient caregivers in M city and T town in N prefecture, by self-reported questionnaire through the mail. We used the questionnaire developed by the Japanese Society of Quality Of Life Research.; RESULTS Change in QOL-Comparison before and after dementia diagnosis: Before average score 7.2, After 6.2—In the 6 dementia items, too, a significant fall (p \ 0.01); Before average score 6.8, After 4.6—Men 6.9–5.6 (p \ 0.01), 6 dementia items 6.7–4.2 (p \ 0.01) A significant decline Women 7.2–6.4 (p \ 0.01), 6 dementia items 6.8–4.7 (p \ 0.01) A significant decline— Caregiver’s relation to patient (1) Spouse 6.9–5.6 (p \ 0.01), 6 dementia items 6.7–4.4 (p \ 0.01). (2) Child 7.4–6.5 (p \ 0.01), 6 dementia items 7.1–4.7 (p \ 0.01)—Living together 7.2–6.2 (p \ 0.01) Not living together 6.6–6.4 Caregiver’s wishes, help requests (1) Fever and bedsores, anxiety about illness (2) Home care support, better treatment/bathing servicePleasant things (1) Day service/day care provides a change of place (2) A smile, a relaxed situation, help from family (3) A ‘‘Thank you’’, having someone to ask for adviceUnpleasant things (1) Anxiety about wandering (2) Worries about the future
109 (3) Night care; CONCLUSIONS There was a significant decrease of the QOL of people taking care of a dementia patient at home, comparing before and after diagnosis of the disease, and especially the fall in the relationship with the patient was large. As 80 % of the caregivers are women, there is a need for concrete physical, mental and financial support for female caregivers. The level of care grade of the patient and the age of the caregiver greatly affect the caregiver’s QOL. In order to allow dementia patients to be able to live with their families, real support to prevent the decrease of caregivers’ QOL is necessary.
(3014) The practice of mindful yoga and breath awareness meditation in promoting health Fusako Nakamachi, MS, Tokyo Kasei-gakuin University, Japan; Ikuko Yamaguchi, Japanese Society for MIndful Living; Yoshiro Konishi, Japanese Society for MIndful Living; Yutaka Haruki, Japanese Society for MIndful Living; Hidehumi Oga, Japanese Society for Mindful Living, Japan AIMS Students sometimes face mental and physical health issues during their university days. By becoming more aware of the importance of taking care of their body and their mind, they may find ways to regain mental and physical health. This study assesses whether practicing yoga and breath awareness meditation might have an positive effect on student well-being.; METHODS The 27 subjects were students at a 4-year women’s college in Japan. A professional yoga instructor guided them in mindful yoga and breath awareness meditation exercises in two 90-min sessions. The first week they practiced yoga. The next week they did breath awareness mediation. Both before and after doing the exercises, the students answered a questionnaire evaluating themselves on six items related to physical and mental changes: vitality, cognitive awareness, emotional state, mood, tenseness, physical well-being.; RESULTS There were statistical differences in four of the mindful yoga items: cognitive awareness (p \ 0.001), emotional state (p \ 0.001), tenseness (p \ 0.002), physical well-being (p \ 0.001). On the other hand, there were statistically significant differences for only two of the breath awareness meditation items: emotional state (p \ 0.011), and mood (p \ 0.034).; CONCLUSIONS Practicing mindful yoga seemed to be more effective in bringing about physical and emotional change than did breath awareness mediation. However, we should not ignore the fact that breath awareness mediation contributed to a calmer and better emotional state of mind. The students had knowledge of yoga but were not familiar with meditation. Less physically active students might find breath awareness mediation attractive.
(3015) Psychological effects of hill walking at Kaminoyama Japan—relations among mood alteration, weathers and trail characteristics Kayoko Machida, Sapporo City University, Japan; Hirofumi Ueda, Sapporo City University; Namiko KAWAMURA, Oita university; Nobuyuki KOSEKI, Kurort Lab AIMS Walking in forest or on hills has been reported to have beneficial effects on person’s psychological state. However such questions as what aspects of moods are improved in what circumstances, how long the enhanced moods last, and how the weather affects on mood changes still remain unanswered. Focusing on these questions, this study evaluated mood alteration of participants of hill walking on different nature trails in various weathers. All the trails in this study are located in Kaminoyama, Japan and designated as climatic-terraintherapy trails.; METHODS A questionnaire including a mood scale of
123
110 12 questions was given to participants of hill walking. They were requested to answer the same mood scale six times (before and after walking, three times during the walking, and the following day of the walking) so that changes in moods can be traced. The mood scale consists of three factors: active, relaxed, and negative mood factors. The same survey was conducted on 5 days in October, 2011. Participants, trails and weathers were different on each survey day. 96 participants answered the questionnaire on the day of walking and 62 of them returned the following-day survey sheet by mail. The mean values of each mood before and after walking were compared using a Student’s paired t-test on SPSS15.0.; RESULTS As a whole, the active mood and the relaxed mood significantly enhanced after walking (t = 4.597, p \ .001 and t = 5.160, p \ .001) and the negative mood was significantly mitigated (t = -5.549, p \ .001). The active and relaxed moods stayed at the increased level until the following day. The negative mood was kept lower on the following day than the mean value of before-walking mood. On a fine day, the active mood significantly went up, while the negative mood showed no difference. On a rainy day, the negative mood was significantly reduced but the active mood remained unchanged.; CONCLUSIONS Walking on a climatic-terrain-therapy trail has beneficial effects on participants’ mood status and the improved moods were kept until the following day. Different weather leads to different mood changes. Even on a rainy day, hill walking has a soothing effect on participants’ negative mood.
(3016) Review of methods used to translate neurocognitive assessments Christelle Giroudet, MAPI Institute; Caroline Anfray, MAPI Institute, France AIMS Neurocognitive assessments explore cognitive functioning which typically include orientation, new-learning/memory, intelligence, language, visuoperception, and executive function. They are usually composed of three elements: the stimulus (e.g., images, digits, letters, words, story, objects from daily life), the instructions to the patients (read by the rater), and the instructions to the rater on how to administer and score the test. The objective of this study is to review the methodologies used to adapt neurocognitive assessments in other languages and cultures.; METHODS A review of projects involving the translation of neurocognitive assessments was conducted focusing on the methodologies used to translate the instructions and the stimulus.; RESULTS Fifteen projects were reviewed. The process used to translate the instructions involved definition of concepts, two forward translations, reconciliation of forward translations, a back translation, resolution of the back and forward translations, a clinician’s review, and proofreading. The process used to translate the stimulus was similar to the process used for the instructions with the addition of two steps: involvement of other experts to help establish and prepare stimuli (e.g., speech therapists) prior to the clinician’s review (at the time of the conceptual definition) and cognitive interviews with subjects. Examples of challenges encountered during the translation process of neurocognitive assessments, such as the National Institutes of Health Stroke Scale (NIHSS) List of Words/ Sentences, the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), the Rey Auditory Verbal Learning Test (RAVLT) or the Alzheimer’s Disease Assessment Scale—Cognitive (ADAS-Cog), will be presented.; CONCLUSIONS The translation of neurocognitive assessments follows a methodology similar to the one used for patient-reported outcome measures. The involvement of experts in the field, such as speech therapists, early in the process is paramount to guarantee the correct adaptation of the stimulus to the target languages and cultures.
123
Qual Life Res (2012) 21:1–132 (3017) Relationship between outcome of hospitalization for the first episode of acute coronary syndrome and baseline health-related quality of life and depression Carina M. Dessotte, University of Sa˜o Paulo; Gislaine P. Gil, University of Sa˜o Paulo; Ma´rcia A. Ciol, University of Washington; Lidia A. Rossi, PhD, University of Sao Paulo, Brazil; Fernanda S. Silva, University of Sa˜o Paulo; Andre´ Schmidt, School of Medicine of Ribeira˜o Preto, University of Sa˜o Paulo; Rosana Spadoti Dantas, PhD, University of Sa˜o Paulo, Brazil AIMS To study the association between discharge outcome and healthrelated quality of life (HRQoL) in the 4 weeks prior to hospitalization and baseline depression in patients hospitalized for a first episode of acute coronary syndrome (ACS).; METHODS Observational prospective study involving 234 cardiac patients who were hospitalized in a Brazilian university hospital for their first episode of ACS (61.8 % had acute myocardial infarction and 38.2 % had unstable angina). HRQoL was measured using the SF-36 (all eight sub-scales) referring to the 4 weeks prior to hospitalization, and depression was measured using the Beck Depression Inventory (BDI) referring to 1 week before hospitalization. Data were analyzed by descriptive statistics for all variables and logistic regression for outcome of hospitalization (alive or dead).; RESULTS Seventy-two participants were females (30.8 %) with mean age 58.8 years (S.D. = 11.1), while 162 were men (69.2 %) with mean age 57.7 years (S.D. = 12.6). During the hospitalization, two women (2.8 %) and five men (3.1 %) died. To assess whether the sub-scales of the SF-36 and the BDI at baseline could explain the outcome ‘‘death or survival’’ after adjusting for the person’s age, we performed a logistic regression analysis with age forced in the model and the subscales of SF-36 and BDI as explanatory variables entered in a forward fashion. Age was statistically significant (p = 0.04), but neither the sub-scales of SF-36 nor the BDI were statistically significant.; CONCLUSIONS Self-report measures of HRQoL, such the SF-36, and of depression, such as the BDI, might be helpful to predict short-term outcomes of patients hospitalized for the first episode of ACS. However, in our population, these measures did not add any more information after age was in the model. These measures may not be good predictors of shortterm outcomes or these measures might not capture the true state of health and depression in our population (such as when the participant responds the question as he/she thinks the investigator wants). It is also possible that the SF-36 and the BDI might be good predictors of longterm outcomes (such as 2-year survival), which will be the focus of our next phase of research.
(3018) Evaluation of the impact of the psychiatric hospitalization on quality of life Gabriela L. Nuernberg, Universidade Federal do Rio Grande do Sul, Brazil; Eduardo Trachtenberg, Universidade Federal do Rio Grande do Sul; Fernanda Baeza, Universidade Federal do Rio Grande do Sul; Thiago V. Freire, Hospital de Clı´nicas de Porto Alegre; Martina Kopittke, Hospital de Clı´nicas de Porto Alegre; Bruno P. Mosqueiro, Hospital de Clı´nicas de Porto Alegre, Brazil; Felipe B.P. Costa, Hospital de Clı´nicas de Porto Alegre; Tatiana K. Sansonowics, Universidade Federal do Rio Grande do Sul; Marcelo P. Fleck, PhD, Hospital de Clinicas de Porto Alegre, Brazil; Neusa Rocha, MD, Universidade Federal do Rio Grande do Sul, Brazil AIMS To evaluate the impact of the psychiatric hospitalization on quality of life of inpatients and compare it to the general population.; METHODS Subjects that were admitted from may/2011 to december/ 2011 to the psychiatric unit of Hospital de Clı´nicas de Porto Alegre (Brazil) were assigned to the study. After an initial evaluation with MINI diagnostic structured interview (MINI), the following
Qual Life Res (2012) 21:1–132 instruments were accessed at admission and at discharge: Brief Psychiatric Rating Scale (BPRS), Hamilton Anxiety Rating Scale (HAMA), Hamilton Rating Scale for Depression (HAM-D), Young Mania Rating Scale (YMRS), Clinical Global Impression (CGI), Global Assessment of Functioning (GAF) and World Health Organization Quality of Life—BREF (WHOQOL-BREF).; RESULTS Ninety eight patients were enrolled to the study. According to MINI, the most frequent diagnosis was major depressive episode (61 %), followed by psychotic episode (20 %) and mania episode (16 %). Every clinical measures improved significantly on hospitalization: mean BPRS scores were 23.1 at admission and 11.1 at discharge (p \ .001); HAM-A mean scores were 25.8 at admission and 8.3 at discharge (p = 0.004); HAMD scores improved from 22.0 to 9.6 (p \ .001) and YMRS scores were 12.5 at admission and .7 at discharge (p \ .0001). The scores of all domains of QOL had statistically significant improvement along hospitalization, except the scores of the environmental domain. On admission, patients had lower scores of QOL in all domains when compared to the general population, while on discharge this difference was only regarding psychological and social domains (p = .05 and .001).; CONCLUSIONS Both symptomatology and QOL of patients with severe mental illnesses improved with hospitalization, but QOL in psychological and social domains still remained poorer when compared to the general population. These findings suggest that interventions should rely not only on symptom improvement, but also focus on improvement of QOL.
(3020) Are patients in the United Kingdom receiving written information about health related quality of life following oesophageal cancer surgery? Sean Strong, Bmed, MRCS, MSC, University of Bristol, United Kingdom; Natalie Blencowe, University of Bristol; Rhiannon C. Macefield, University of Bristol, United Kingdom; Sara Brookes, University of Bristol; Joanne Nicklin, University Hospital Bristol NHS Foundation Trust; Kerry Avery, PhD, Dept of Social Med, Univ of Bristol, United Kingdom; Jane Blazeby, MD, University of Bristol, United Kingdom AIMS Surgery is currently the mainstay of curative treatment for oesophageal cancer. Oesophagectomy, however, is associated with a significant risk of in-hospital death, poor long term survival and has a negative impact on health-related quality of life (HRQL). To reach treatment decisions and satisfy informed consent patients require adequate information provision. Patient information leaflets (PILs) are often provided as an adjunct to hospital consultations with specialists, although their usefulness and quality are currently unknown. The aim of this study, therefore, was to summarise the content of PILs provided prior to oesophageal cancer surgery in hospital trusts throughout England and Wales.; METHODS NHS hospital trusts performing oesophageal cancer surgery were identified from the 2010 National Oesophago-gastric Cancer Audit Report. PILs, pertaining to oesophagectomy, were obtained from trust websites. Where PILs were not accessible online, clinical nurse specialists were contacted by telephone, and paper copies supplied by post. Components reported were selected a priori using information from a systematic review summarising outcome reporting in oesophageal cancer surgery. Supplementary components were added if missing from the original framework. All PILs were independently coded by two surgical research fellows.; RESULTS 43 hospitals were identified as performing oesophageal cancer surgery in England and Wales, of which 41 PILs were obtained for evaluation. Most PILs explained the operative procedure itself (n = 29). Provision of information concerning postoperative symptoms and HRQL was poor with few PILs discussing changes in physical (n = 22), emotional (n = 9), role (n = 11) and sexual (n = 7) function.; CONCLUSIONS Although most PILs
111 provided a detailed description of the operation itself, issues pertaining to HRQL following surgery were less frequently discussed. Together with outcome data obtained from literature reviews of clinical and patient-reported outcomes, results from this review will inform development of a ‘core disclosure set’ for patients undergoing oesophageal cancer surgery. It is therefore recommended that PILs also incorporate these ‘core disclosure’ measures at a minimum, ensuring that all patients receive key information about their surgery.
(3022) Oncologists’ perceptions of psychosocial communication during outpatient visits—barriers and orientation ˚ sa Kettis, Hanna Fagerlind, Uppsala University, Sweden, Sweden; A Uppsala university; Bengt Glimelius, Uppsala university; Lena Ring, PhD, Medical Products Agency (MPA) and Uppsala University, Sweden AIMS The aim was to explore oncologists’ psychosocial orientation, i.e. attitudes and burden associated with treating patients psychosocial problems. Further, the aim was to explore physicians’ perceptions regarding barriers for communicating psychosocial areas during outpatient consultations.; METHODS A postal questionnaire was sent out in end of 2011, to all clinically active oncologist in Sweden (n = 537). The questionnaire included questions regarding the oncologists’ demography, the Physician Psychosocial Beliefs Scale (PPBS) ranging from 32(maximum psychosocial orientation) to 160 (minimal psychosocial orientation). Further the questionnaire included 11 questions regarding barriers for psychosocial communication. Descriptive statistics were used.; RESULTS 344 clinically active oncologists completed the questionnaire. The response rate was 64 %. The mean PPBS value was 85.4 (range 49–123, SD = 13.1). The oncologist perceived about 5 communication barriers with regard to psychosocial aspects. Of these barriers more than 50 % was perceived to impact their clinical handling of patients. The most common barriers affecting oncologist clinical handling of patients were: the consultation time being too short to discuss psychosocial issues, the lack of resources for taking care of the potential problems discovered and the lack of good methods in clinical practice for evaluating patients’ psychosocial health/status. There was a week positive correlation between the PPBS score and the number of perceived barriers, i.e. less psychosocially oriented oncologists perceived more barriers.; CONCLUSIONS The PPBS scores for oncologist correlates well to previous research. However, the consistent findings suggest that oncologist are less psychosocially oriented than other physician specialties. Reasons for this should be further explored. Oncologists perceive many different kinds of barriers impacting psychosocial communication during medical consultations. Interventions aiming to resolve barriers must therefore be multifaceted and to some extent individualized to the context of different clinics as well as for different oncologists. It is important to aim to resolve or minimize the barriers to facilitate optimal care and treatment for patients with cancer.
(3023) Framework and criteria to sort, select and reduce measurement instruments for daily practice in physical therapy Anna Beurskens, PhD, Zuyd University of applied sciences, Netherlands; Guus Meerhoff, Royal Dutch Society of Physical Therapy, Netherlands; Victorine De Graaf-Peters, Royal Dutch Society of Physical Therapy; Raymond Swinkels, Zuyd University of applied sciences AIMS The amount and diversity of instruments in the Dutch physiotherapy guidelines is too extensive: in total 111 instruments. Many
123
112 of these instruments are too long and concern the same domains and concepts. The aim of this bottom-up project is to develop, together with professionals and guideline developers, a framework and criteria that will guide the structure, selection and reduction of feasible instruments for application in daily physical therapy practice.; METHODS The project consist of the next steps: Step 1 Defining the general aim of the framework and review of the (inter)national literature Step 2 Development of a concept framework Step 3 An inventory of domains and overlap of instruments in the Dutch guidelines Step 4 An inventory made by professionals of domains and concepts relevant for musculoskeletal functioning in physical therapy and measurement instruments used Step 5 Consensus meetings with representatives of professionals and guideline developers Step 6 Processing results and preparation of field tests: sorting, selection and reduction of measurement instruments for domains and concepts by using the framework and criteria for feasibility and use in daily practice Step 7 Implementation- and dissemination planning.; RESULTS Conceptual framework based on the structure of clinical reasoning and International Classification of Functioning. An inventory of domains that are relevant for physical therapy and feasibility criteria for the selection of measurement instruments. A guide for the selection and reduction of recommended measurement instruments in guidelines and use in daily practice.; CONCLUSIONS The framework gives professionals and guideline developers the same structure and a clearer understanding about the selection of measurement instruments for daily physical therapy practice. It helps professionals to learn when to use which instrument for what patient.
(3024) Surgical decision in breast cancer: impact on health related quality of life Teresa Sequeira, Portugal; Ariana Santos, Portuguese Oncology Institute, Porto, Portugal; Joaquim Gonc¸alves, Polytechnic Institute of Ca´vado & Ave, Portugal; Cla´udia C. Ribeiro, Universidade Cato´lica Portuguesa, Portugal; Joana Teixeira, Nurse, Portuguese Oncology Institute, Porto, Portugal, Portugal; Joaquim Abreu Sousa, Portuguese Oncology Institute, Porto, Portugal; Augusta Silveira, University Fernando Pessoa, Portugal AIMS The surgical approach for breast carcinoma generally has three equally effective options: modified radical mastectomy (RM), breast conserving surgery (BC) or mastectomy with reconstruction (BR). Although the long-term clinical outcome is comparable there is a growing interest considering Quality of Life (QoL) outcomes in order to find a patient-centered option. The present study aims to compare the functional and burden of women submitted to the three common beast cancer surgical procedures.; METHODS Breast cancer patients (n = 95) from the Breast Clinic of the Portuguese Oncology Institute, Porto, submitted to surgery for less than 1 year and being followed up—between May and November 2011—were enrolled in this study after given their informed consent. Patients (mean age 53.45 years, 11.88 years old) completed the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30) and its specific supplemental module (QLQ-BR23) thought the Quality of Life Informatics Platform (QoLIP) already in use at the institution. Scales were linearly transformed into scores (0–100). Kruskall-Wallis and Mann-Whtiney tests were used as nonparametric tests.; RESULTS The QLQ-C30 functional subscales scores were compared for each surgery type and the higher scores observed were assigned to BC (n = 17) for role functioning (75.81), cognitive functioning (83.13) and social functioning (84.27). BR (n = 33) pointed best for physical functioning, (77.00) and emotional functioning, (69.39). The lowest scores observed were mainly
123
Qual Life Res (2012) 21:1–132 associated with RM (n = 45) for emotional functioning (67.09) and cognitive functioning (76.53) but also in BC for physical functioning (73.76) and in BR for role functioning (56.47). When considering the five functional subscales of QLQ-BR23, BC was related to the best scores for body image (77.88) and future perception (43.50), followed by BR that scored maximum for sexual functioning (25.32.) RM presented again the lowest scores.; CONCLUSIONS The lowest QoL scores were mainly ascribed to patients submitted to RM. As BC and BR present different advantages, patient reported outcome contributes to a better understanding of both surgery and disease course. This knowledge should be shared with patients and may contribute to support clinical decision in order to promote a personalized medicine.
(3025) Does thyroid disease affect work function? Results from a qualitative study Mette A. Nexo, PhD. Student, The National Research Centre for the Working Environment, Denmark; Torquil Watt, MD PhD, Copenhagen University Hospital, Denmark; Ulla Feldt-Rasmussen, ˚ se K. Rasmussen, Copenhagen Copenhagen University Hospital; A University Hospital Rigshospitalet, Denmark; Steen J. Bonnema, Odense University Hospital; Laszlo Hegedu¨s, Odense University Hospital; Mogens Groenvold, MD, PhD, University of Copenhagen, Denmark; Jakob Bjorner, PhD, National Research Centre for the Working Environment, Denmark AIMS Thyroid diseases affect a large portion of the labour force. Thyroid disease can reduce quality of life (QOL), but few studies have focused on the work function component of QOL. These studies point to considerable work impairments in subgroups of persons with thyroid disease. However, further quantitative studies with large samples are needed to reveal whether work function is decreased in persons with thyroid disease in general. Qualitative studies can reveal the possible mechanisms involved. This qualitative study examines how thyroid disease may affect work function.; METHODS Seventeen medically treated persons with thyroid diseases (autoimmune hypo- and hyper-thyroidism) were strategically selected for semi-structured interviews from two main hospitals in Denmark. Interviews focused on how the participants experienced their health, disease, work and daily life. The interviews were analysed using Interpretative Phenomenological Analysis. Themes of importance to the experienced life-worlds were grouped, developed, and categorized in a step-wise procedure.; RESULTS Three main themes emerged regarding the patients’ work life (1) Illness perception: Thyroid disease was not experienced as a disease, rather as a diffuse mental state (e.g. fatigued and woozy). Participants doubted that their diffuse state of mind was a valid manifestation of thyroid disease. (2) Experienced impairments: Work impairments were primarily concerned with mental abilities such as cognitive (e.g. pace, concentration, exhaustion) and problems remaining at work during required hours. (3) Sickness behavior: The need to recover was perceived as an individual responsibility and was not handled by sickness leave.; CONCLUSIONS This study illustrates how the perception of illness may complicate the identification and rehabilitation of the work function of people with thyroid disease. The illness perceptions are easily confused with normal mental phenomena and problems defining these perceptions within a biomedical disease specific framework render the experienced work impairments invalid. In order to improve the QOL component work function for persons with thyroid disease future research needs to identify the need for specific work rehabilitation interventions.
Qual Life Res (2012) 21:1–132 (3026) Reliability and validity of the psoriasis symptom inventory in patients with moderate to severe psoriasis Dennis Revicki, PhD, United BioSource Corporation; Ying Jin, United BioSource Corporation; Dina Chau, Amgen Inc; Hema Viswanathan, Amgen, Inc AIMS The Psoriasis Symptom Inventory (PSI) was developed to assess psoriasis related symptoms consistent with best practices for patient-reported outcome (PRO) instrument development. The PSI was administered in a Phase 2 clinical trial to evaluate the effect of brodalumab on improvement in psoriasis symptoms. The objective of this analysis was to evaluate the psychometric characteristics of the PSI in patients diagnosed with moderate to severe psoriasis.; METHODS Psychometric analyses were completed using selected Phase 2 clinical trial data from 186 patients with moderate to severe plaque psoriasis. Clinical outcomes included the Psoriasis Area and Severity Index (PASI) and Static Physician’s Global Assessment (sPGA). Patients completed the PSI and Dermatology Life Quality Index (DLQI). Unidimensionality was evaluated using confirmatory factor analyses (CFA) and item response theory (IRT) analyses. We examined internal consistency using Cronbach’s alpha, test–retest reliability using intraclass correlation coefficients (ICCs), concurrent validity with Pearson correlations between related constructs, and known groups validity using analysis of variance models. Responsiveness of the PSI total score was evaluated by examining changes in PSI scores by responder groups based on PASI and sPGA.; RESULTS The sample was 36 % female, 89 % Caucasian, and averaged 42.9 (SD = 12.2) years old. In the CFA, the comparative fit index was 0.94, indicating very good fit for the single factor model. For the IRT analysis, slope parameters ranged from 2.10 to 3.58, and threshold values ranged from -3.14 to 1.53. Internal consistency reliability was 0.93–0.98 and test–retest reliability was 0.87–0.91. The PSI total score was moderately to strongly correlated with the DLQI symptom item (0.63–0.88) and total scores (0.55–0.82). PSI total scores varied significantly by PASI (p \ 0.001) and by sPGA severity groups (p \ 0.0001), with greater mean PSI scores observed in more severe rated groups. PSI total scores significantly improved with improvements in PASI (p \ 0.001) and sPGA scores (p \ 0.001).; CONCLUSIONS The PSI showed high reliability, construct validity, and responsiveness to changes in clinical status. The PSI demonstrates potential for use as a PRO measure to assess symptoms of patients with moderate to severe psoriasis in clinical trials.
(3027) Attitudes of Portuguese nurses facing end-of-life decisions Pedro L. Ferreira, PhD, University of Coimbra, Portugal; Cristina R. Nogueira, Coimbra University Hospital, Brazil AIMS Three concepts are usually associated with end-of-life decisions: dysthanasia, assisted suicide and euthanasia. Its use in practice is sometimes confusing and the lack of clarification often influences the public, patients and family and even health professionals. On the other hand, little is known about the experience, opinions and attitudes of nurses against such decisions. The present study sought to understand the attitudes of nurses towards end-of-life decisions.; METHODS The Belgian questionnaire ‘‘Attitudes verpleegkundigen over beslissingen ann het levenseinde bij patie¨enten’’ was implemented, after having been subjected to a processes of translation and cultural validation for Portuguese. The sample was made up of 159 nurses attending post-graduate specialization in 2010.; RESULTS The nurses want to be involved in all decision-making procedures on patients, especially for end-of-life decisions. In fact we obtained 95 % of agreement on the questions related to this issue Almost half of them didn’t agree that nurses have some difficulty in express their opinions to doctors. Although they do not accept clearly euthanasia,
113 these professionals consider that the relief of suffering can justify its practice. We didn’t identify any relationship between personal characteristics or employment of professionals and their attitudes and opinions about decisions related to end of life. However, the religious factors stressed as the most influence the nurses in this matter.; CONCLUSIONS After this project we are aware of the attitudes and opinions of nurses about issues related to end-of-life and we were able to identify some individual and labour factors that may influence the attitudes of nurses. Providing information about the role and the practice of the nurses in end-of-life decisions, we expect that these results can contribute to a deeper discussion and ethical debate in the Portuguese society.
(3028) Developing and pre-testing an item pool relating to the effects of exposure to health websites Laura Kelly, PhD, United Kingdom; Sue Ziebland, University of Oxford, United Kingdom; Crispin Jenkinson, Health Services Research Unit, United Kingdom AIMS The internet is a valuable resource for accessing health information in the UK with 71 % of the population having sourced health information online. Health websites have developed to be much more than information sites: they are used to exchange experiences and find support as well as information and advice. We are developing an e-health impact questionnaire to assess the effects of websites with experiential and factual information. This questionnaire will be used in a randomised control trial examining the effects of online patient’s experiences. This study aimed to develop an item pool to inform the e-health impact questionnaire.; METHODS Items were informed through a review of relevant literature and secondary qualitative analysis of 226 narrative interviews relating to patient and carer experiences of health. Interviews were analysed using the Framework method. Statements representing themes relating to sharing patient and carer experiences online were selected and re-cast as questionnaire items. Items were shown for review to an advisory board consisting of six academics and clinicians working in the area of e-Health. Cognitive debrief interviews (n = 21) were used to assess items for face and content validity using the verbal probing method.; RESULTS Sixty-seven generic items were constructed according to the key themes identified through relevant literature and qualitative analysis as important to sharing health experiences online. Items were amended to improve face validity following reviewer’s comments. Three rounds of cognitive interviewing confirmed (1) the instructions were easy to understand, (2) the acceptability of 62 items, (3) the response options were appropriate and (4) the electronic format of the questionnaire was appropriate for use. Ethical approval was obtained for cognitive testing through the University of Oxford Ethics Committee.; CONCLUSIONS Health related measurement scales require a clear conceptual basis to inform item generation. Developing a measure through the use of qualitative analysis and expert review can enhance both face and content validity. Instructions, items and response options were acceptable to patients and carers. Further tests are currently being conducted to establish the psychometric properties of the scale.
(3029) Identification and classification of individualized patient-reported outcome measures Daniela C. Gonc¸alves, University of Oxford, United Kingdom; Ignacio Ricci-Cabello, University of Oxford; Jose M. Valderas, MD, PhD, MPH, University of Oxford, United Kingdom
123
114 AIMS Individualized patient-reported outcome (PROs) measures allow for active involvement of the individual respondent in specifying their content and/or scoring and are particularly promising for use in clinical practice. The goal of this study was to identify the existing individualized PROs, classifying them according to the level of individualization, measured construct and disease, target population.; METHODS A systematic review of the literature was defined, based on an integrated model for health outcomes, previously published reviews, and direct contact with experts. A snowball technique was employed to detect further relevant references.; RESULTS Seventy three PROs with some degree of individualization were identified, 66 of which were fully classified. Eight families of instruments, defined as different versions of the same instrument and/or specifications for different populations, comprised 43 (58 %) of the PROs. The majority of PROs were developed in English (60 %), mainly in the USA (n = 20), UK (n = 18), and Canada (n = 13). Approximately two thirds of the PROs were individualized at the content level (domains nomination, selection, or both), and a similar proportion had individualized weighting. Eleven measures allowed the respondent to nominate domains and establish weights. Only three permitted the respondent to define the scale, and these were also individualized at the content and weighting levels. Twenty one measures were not condition specific, and of these seventeen targeted adults. As for the constructs measured, functional status was the most common, followed by symptoms, health-related quality of life, and general health perceptions. Several measures focussed on more than one construct. The most cited generic instrument was Ferrans and Powers Quality of Life Index.; CONCLUSIONS Although the advancement of individualized PROs has been considerably slower than standardized, there are already a substantial number of available measures. The existing individualized PROs differ greatly regarding the degree of individualization, the targeted population, and the construct measured.
(3030) Patient-reported quality of life outcomes after cystectomy for bladder cancer: a comparison of open and robotic-assisted laparoscopic techniques Jeffrey S. Montgomery, University of Michigan; John M. Hollingsworth, University of Michigan, Department of Urology, United States; Change He, University of Michigan; Scott Gilbert, University of Florida, Department of Urology; John T. Wei, MD, University of Michigan, United States; Rodney Dunn, MS, University of Michigan Dept of Urology; Brent Hollenbeck, MD, MS, University of Michigan; Alon Z. Weizer, University of Michigan AIMS Conventional wisdom suggests that newer and less invasive surgical approaches will result in improved outcomes, such as healthrelated quality of life (HRQOL). Robotic-assisted laparoscopic cystectomy is a recent innovation for the treatment of bladder cancer that is touted to be less invasive. To date, no study has compared HRQOL recovery between open and robotic cystectomy.; METHODS We conducted a prospective study of 324 consecutive patients (267 open and 57 robotic) who underwent radical cystectomy for bladder cancer at our institution between 2008 and 2012. We assessed baseline and post-operative patient-reported quality of life at 1, 3, 6 and 12 months using the Bladder Cancer Index (BCI) validated questionnaire. The BCI consists of 36 items divided among 3 domains: urinary, bowel and sexual. Each domain contains function and bother components. Domain scores range from 0 to 100, with higher scores corresponding to a better health state. We evaluated HRQOL recovery differences by surgical approach using linear regression models.; RESULTS There were no significant differences in patient age, gender, clinical stage
123
Qual Life Res (2012) 21:1–132
Fig. 1 HRQoL recovery after cystectomy assessed by BCI
and urinary diversion type between the 2 groups. Baseline BCI scores were comparable, and by 1 year after cystectomy, all scores returned to near baseline levels (Fig. 1). Within the first year after cystectomy, the recovery of urinary, bowel and sexual function and bother HRQOL after open and robotic cystectomy was comparable.; CONCLUSIONS The recovery of patient-reported HRQOL was not found to be different following open and robot-assisted laparoscopic cystectomy, despite the minimally-invasive nature of the robotic technique. This raises questions regarding the use of this more expensive technology in the setting of limited healthcare resources. Ongoing multicenter studies should examine other outcomes, including complications and surgical recovery.
(3033) Analysis for DIF in the Spanish PROMIS pain interference item bank Leo Morales, MD, Group Hlth Research Institute; Sylvia H. Paz, PhD, UCLA; Karen L. Spritzer, UCLA; Ronald D. Hays, PhD, UCLA David Geffen AIMS To investigate the newly translated PROMIS pain interference (PI) item bank for language DIF. The PI bank assesses the impact of pain on physical, mental and social functioning aspects of health from the patient’s perspective.; METHODS The study included 527 Spanish-speaking and 678 English-speaking adults all in the United States. The Spanish sample mean age was 36 (SD = 10.5); 63 % were female; and 32 % had a high school diploma or less, 33 % had some college, and 34 % had college degree or more. Nine percent spoke only Spanish, 53 % spoke Spanish better than English, 37 % spoke Spanish and English equally well, and \1 % spoke English better than Spanish. By comparison, the English sample mean age was 51 years (SD = 18.9 years); 54 % were female; and 19 % had a high school diploma or less, 35 % had some college and 46 % had a college degree or more. The PROMIS PI bank consisted of 41 items each with 5 response categories. The PROMIS protocol was used to produce the Spanish translations. The protocol involves 2 initial
Qual Life Res (2012) 21:1–132
115
forward translations, 1 reconciled version, 1 back-translation, comparison of original with back-translations, and 3 reviews by bilingual experts. Cognitive interviews with native Spanish speakers followed to evaluate the comprehension of the items. Analyses included assessments of dimensionality, DIF and DIF impact simulation studies.; RESULTS Coefficient Alpha for the item bank was 0.99; the mean item-rest correlation was 0.82 (range from 0.72 to 0.89); and a one-factor confirmatory model had a CFI of 0.88 and RMSEA of 0.179—all suggestive of one underlying dimension. LORDIF in R was used to assess DIF. Using a criterion of a pseudo-R2 = 0.02, two items were found to show DIF (painin39 and painin48) (see Appendix). CAT simulation studies using Firestar were run to assess DIF impact. Theta estimates for the Spanish sample generated using English and equated-Spanish item parameters were highly correlated (r = 0.97).; CONCLUSIONS Two items in the Spanish pain interference bank showed DIF. The impact of the DIF in CAT simulation studies was small and probably negligible. However, to be consistent with the approach adopted for other PROMIS Spanish item banks, we recommend using equated Spanish item parameters to score these two items and English parameters for all other items.
were found in the translation of some items, which may lead to crosslanguage measurement differences, such as subtle changes in the concept being measured and reordering of item locations relative to others (differential item functioning). For example, in Dutch-Flemish no distinction is made between fatigued and tired. Also, the translation of eight different words for fatigue levels may have resulted in a different ordering (item difficulties) of the fatigue items. In a few cases it was necessary to have separate translations for Belgium and the Netherlands. For example, the word ‘‘walking’’ was translated as ‘‘lopen’’ in Dutch, but had to be translated as ‘‘stappen’’ in Flemish. In addition, some original English items were considered suboptimal. For example, the item ‘‘Being tired made it hard for me to play or go out with my friends as much as I’d like’’ was considered too difficult for kids. Finally, the reading level of the translation seemed rather high, which may be the result of the high education level of the translators.; CONCLUSIONS Cognitive debriefing is being conducted in the general population in Belgium and the Netherlands. A reading level test is also being performed. Subsequently, the item banks need to be calibrated and validated in the Dutch-Flemish general population and in various disease populations.
(3034) Dutch-Flemish translation of the PROMIS item banks
(3035) The use of EuroQoL-5D in pharmacoeconomic studies of dermatolological treatments: is it relevant?
Caroline B. Terwee, PhD, VU University Medical Center, Department of Epidemiology and Biostatistics, Netherlands; Leonard Roorda, MD PhD, Netherlands; Maarten Boers, VU University Medical Center, Department of Epidemiology and Biostatistics; Joost Dekker, VU University Medical Center, Department of Rehabilitation Medicine, Amsterdam, the Netherlands; Rene Westhovens, University Leuven, Belgium, Belgium; Jaap van Leeuwen, CEO Leones Group BV, the Netherlands; Martha Grootenhuis, PhD, Academic Medical Center, Netherlands; Lotte Havermann, Emma Children’s Hospital/AMC, Psychosocial Department, Amsterdam, the Netherlands; Hein Raat, Erasmus MC, University Medical Center, Department of Public Health, Rotterdam, the Netherlands; Eline van Dulmen-den Broeder, VU University Medical Center, Department of Pediatrics, Amsterdam, the Netherlands; Marion AJ van Rossum, Amsterdam Rehabilitation Research Center, and Emma Children’s Hospital/AMC, Psychosocial Department, Amsterdam, the Netherlands; Karel Hoppenbrouwers, Centre of Youth Health Care, KU Leuven, Belgium; Helena Correia, Northwestern University, United States; Benjamin J. Arnold, MA, FACITtrans, Brazil; David Cella, PhD, Northwestern University; Henrica C.W. de Vet, Sr., PhD, EMGO Institute VU Medical Center, Netherlands AIMS The Patient-Reported Outcomes Measurement Information System (PROMIS) contain a large number of item banks about different aspects of health. It is carefully designed using Item Response Theory (IRT) methods, to perform reliable and valid patient-reported outcome (PRO) measurements across a wide variety of (patient) populations. The possibility of computer adaptive testing (CAT) makes the system highly efficient and flexible for use in research and daily clinical practice. The aim of this study was to translate the PROMIS item banks into DutchFlemish language.; METHODS Twelve item banks for adults (464 items) and nine item banks for children (155 items) were translated according to the FACIT methodology, consistent with existing guidance for translation and validation of PRO instruments. The translation process included four forward translations by two native Dutch and two native Flemish speakers, two back-translations of the reconciled version by native English speakers (one specializing in Dutch, one specializing in Flemish), comparison of source and back-translated versions, and three independent reviews by bilingual, native Dutchspeaking experts (two Dutch, one Flemish).; RESULTS Difficulties
Sam Salek, BSc RPh PhD FFPM, Cardiff University, Redwood Bldg, United Kingdom; Mohammad K.A. Basra, MD, Cardiff University, United Kingdom; Flavia Pereira, Cardiff University; Vicent Piquet, Cardiff University; Andrew Finlay, FRCP, Univ. of Wales College of Medicine, United Kingdom AIMS The EQ-5D, despite its lack of specific dermatology origin, is being used as a utility measure in clinical and pharmacoeconomic studies in dermatology. Such use has been driven by health technology assessment (HTA) agencies to inform their decisions for reimbursement approval of new expensive therapies and by increasing emphasis on pharmacoeconomic research in dermatology. The aim of this review was to examine the use of EQ-5D in pharmacoeconomic studies in dermatology.; METHODS A systematic search involving six online sources and hand search of journals in the medical school library was carried out using search terms: EQ-5D, skin disease, dermatology, economic analysis, cost-effectiveness, cost minimisation, cost-utility, cost of illness, cost-benefit, pharmacoeconomics.; RESULTS Nineteen articles and one abstract published between 2003 and 2011 were identified that had EQ-5D as an outcome measure in cost-of-illness, cost-effectiveness and cost-utility analysis studies. Seven were costof-illness studies involving four skin diseases (atopic eczema, herpes zoster, psoriasis, and acne) concluding that different skin conditions are associated with different levels of burden. There was only one costeffectiveness study that used the EQ-5D. The results indicated that home ultraviolet-B phototherapy for psoriasis was not more expensive than phototherapy in an outpatient department. Thirteen cost-utility studies evaluated the cost versus utility of various interventions for five different skin diseases: chronic hand eczema (alitretinoin), malignant melanoma (interferon-alpha), atopic eczema (tacrolimus), herpes zoster (vaccines) and psoriasis (home vs out-patient phototherapy, biologics, MTX, ciclosporin, topical therapy, in-office vs online care).; CONCLUSIONS Although the EQ-5D allows the possibility of measuring utility and comparisons with non-dermatological diseases, it may not detect important and specific issues pertinent to dermatological diseases. In order to understand the true economic burden of skin disease and its treatments, the EQ-5D should be used in conjunction with dermatology or a skin disease-specific quality of life measure in clinical trials.
123
116 (3036) Language and age equivalence in the patient-reported outcomes information system (PROMISÒ) physical functioning items Sylvia H. Paz, PhD, UCLA; Karen L. Spritzer, UCLA; Leo Morales, MD, Group Hlth Research Institute; Ronald D. Hays, PhD, UCLA David Geffen AIMS To evaluate the equivalence of the PROMIS physical functioning item bank by language of administration (English vs. Spanish) and age (50 or older vs. younger).; METHODS The PROMIS wave 1 English-language physical functioning bank consists of 124 items and 114 of these were translated into Spanish. The items were administered to 640 adult Spanish-speaking Latinos.; RESULTS Item means ranged from 2.53 (SD = 1.36) to 4.62 (SD = 0.82). Coefficient alpha was 0.99 and item-rest correlations ranged from 0.41 to 0.89. A one-factor model fit the data well (CFI = 0.971, TLI = 0.970, and RMSEA = 0.052). IRT parameters ranged from ‘‘Are you able to run 10 miles?’’ (lowest slope) to ‘‘Are you able to put on a shirt or blouse?’’ (largest slope); and ‘‘How much do physical health problems now limit your usual physical activities?’’ (smallest category threshold) to ‘‘Are you able to run ten miles (16 km)?’’ (largest category threshold). Fifty of the 114 items were flagged for language DIF and 30 for age DIF based on an R-squared of 0.02 or above criterion. The expected total score was higher for Spanish than English-language respondents, and for those under 50 than for those 50 or older. Limiting the analysis to items without language DIF, revealed no items flagged for age DIF either.; CONCLUSIONS: English versus Spanish-speaking subjects and those 50 or older versus younger, with the same level of physical function, respond differently to the PROMIS physical function items. This study yields essential information about the equivalence of the physical functioning items in Spanish versus English as well as older versus younger individuals.
(3037) Using the PROMs data to understand and interpret changes in EQ-5D scores Fang-Ju Lin, University of Illinois at Chicago, United States; Jennifer C. Samp, University of Illinois at Chicago; Alexis R. Munoz, University of Illinois at Chicago; Pei Shieen Wong, University of Illinois at Chicago; Simon Pickard, PhD, University of Illinois at Chicago, College of Pharmacy, United States AIMS Using the Patient Reported Outcome Measures (PROMs) data collected by the UK National Health Service, we aimed to examine changes in Oxford Knee Score (OKS), EQ-5D Index and EQ-5D Visual Analog Scale (EQ-VAS) scores, and to benchmark minimally important difference (MID) for the EQ-5D in patients undergoing knee replacement surgery.; METHODS Patients who received knee replacements from 2009 to 2011 completed the EQ-5D and OKS before and after interventions. Responsiveness of the measures was compared using effect size (ES). MIDs for EQ-5D Index and EQ-VAS were estimated using both distribution-based (0.5 standard deviation, 0.5 SD) and anchor-based approaches. Published MIDs for the OKS (3–5) were used as an anchor-based criterion to identify patients who experienced minimal clinically important change, and their corresponding EQ-5D change scores were estimated as MIDs. EQ-5D changes scores were also examined among subgroups of patients according to OKS baseline and change scores.; RESULTS A total of 54,486 patients were included in the analysis. After knee replacement, all measures captured improvements in mean change scores of 0.30 (SD 0.33) for EQ-5D Index, 3.31 (SD 21.03) for EQ-VAS and 14.88 (SD 9.89) for OKS (all p values \0.001). OKS and EQ-5D Index were the most responsive (ES = 1.92 and 0.96, respectively), while EQ-VAS demonstrated limited responsiveness
123
Qual Life Res (2012) 21:1–132 (ES = 0.17). Using the distribution-based method, the range of MID was estimated to be 0.10–0.15 for EQ-5D Index score and 9–12 for EQ-VAS. Using the anchor-based approach, MIDs for EQ-5D Index ranged from 0.08 to 0.12. By stratifying patients according to OKS change score, it revealed that substantial improvement must be seen in OKS before mean EQ-VAS change scores are positive.; CONCLUSIONS Although EQ-VAS appears to be less responsive than EQ-5D Index and OKS, it is misleading to conclude that the measure is not responsive. Instead, patients rate their health systematically more negatively using EQ-VAS compared to EQ-5D Index and OKS. This study emphasizes the importance of using several approaches to eliciting patient-reported outcomes, and also creates a dilemma for decision-makers by providing conflicting information about the desirability of a clinical intervention such as knee replacement.
(3038) Estimating the Portuguese EQ-5D value set using the TTO Lara N. Ferreira, PhD, University of the Algarve-ESGHT, Portugal; Pedro L. Ferreira, PhD, University of Coimbra, Portugal; Luis N. Pereira, University of the Algarve-ESGHT; CIEO AIMS This study aimed at deriving an EQ-5D value set using the time trade-off (TTO) for Portugal. It also intends to measure and analyse national EQ-5D data to provide norms for the Portuguese general population.; METHODS A representative sample of the Portuguese general population (n = 450) stratified by age and gender valued 24 health states defined by the EQ-5D. Face-to-face interviews were conducted by trained interviewers. Each respondent ranked and valued 9 health states using the TTO. Several models were estimated at both the individual and aggregate levels for predicting health state valuations. A national survey was carried out by phone interview by trained interviewers. A representative sample of the Portuguese general population (n = 1,500) stratified by age, gender and region answered to EQ-5D. The Portuguese TTO value set was used to obtain the single index EQ-5D score and therefore derive the population EQ-5D norms.; RESULTS The estimated EQ-5D models were analyzed in terms of their coefficients, overall fit and the ability for predicting the TTO values. Random effects models were estimated using generalized least squares and were robust across model specification. The best fitted model is generally consistent with other European value sets. Descriptive analyses were performed by gender and 10-year age groups, marital status, educational groups and region. Health status declines with advancing age, and women reported worse health status than men. These results are similar to other EQ-5D population health studies.; CONCLUSIONS This research provides the Portuguese EQ-5D value set and is therefore recommended for use in cost-utility analyses conducted in Portugal. This study also provides Portuguese population health-related quality of life data measured by the EQ-5D, based on a national representative sample that can be used as population norms as reference values when comparing different groups’ health status with the Portuguese population. The main findings for different subgroups are consistent with results from EQ-5D population studies in other countries.
(3039) Quality of life predictors for Portuguese general population Brigida F. Patricio, MSc, Escola Superior de Tecnologia da Saude do Porto, Instituto Politecnico do Porto, Portugal; Luis M.T. Jesus, Universidade de Aveiro; Madeline Cruice, PhD, Dept of Language & Communication Science, United Kingdom; Andreia Hall, Universidade de Aveiro
Qual Life Res (2012) 21:1–132 AIMS The purpose of this research was to assess the quality of life of Portuguese adults (general population sample) and identify predictors of QOL in this population; METHODS Two hundred and fifty-five (n = 255) participants from the general Portuguese population have taken part in this research (mean age 43 years, range 25–84 years; 148 female, 107 male). Participants completed a postal European Portuguese version of the World Health Organization Quality of Life shortform instrument (WHOQOL-Bref, Serra et al., 2006) and a European Portuguese version of the Center for Epidemiologic Studies Depression Scale (CES-D, Gonc¸alves & Fagulha, 2004). Demographic information was also collected.; RESULTS Portuguese adults reported their QOL as good. The physical and psychological domains strongly predicted overall QOL (65 %). The weakest predictor domain was social relationships. Age, educational level, socioeconomic status and emotional status were significantly correlated to QOL and explain 50 % of the variance in QOL scores. From those, the strongest predictor of QOL was emotional status followed by educational level and age. QOL was significantly different among: marital status; living place (continent or islands); people they live with; occupation; health; CONCLUSIONS The Portuguese population report high levels of QOL. The strongest predictors of QOL were: physical domain; psychological domain; educational level; environmental domain; age. The results of the domains were similar to those of Serra et al. (2006), but other predictors were not studied in Portuguese Population (Serra et al., 2006). This data is important for heath care policies and therapy planning since it reveals the most important variables that influence the QOL of people who attend health care services.
(3040) Quality of life and subjective age in Southern European older adults: a comparative multiple correspondence analysis for latent constructs Sofia Von Humboldt, MSc, Research Unit in Psychology and Health; ISPA—Instituto Universita´rio, Portugal; Isabel Leal, PhD, UIPES Health & Psychology Investigation Unit, Portugal AIMS To analyze the determinants of quality of life (QoL) and subjective age (SA) identified by older adults and to investigate the latent constructs that can work as major determinants in QoL and SA in an older Portuguese and Romanian population; METHODS Measures were completed, using a variety of culturally appropriate methods, including demographics and interviews. Complete data were available for 64 older adults aged between 72 and 99 years (M = 80.1; SD = 5.8). Data was subjected to content analysis. Representation of the associations and latent constructs were analyzed by a Multiple Correspondence Analysis (MCA).; RESULTS The most prevalent response of the interviewed participants for QoL was ‘Accomplishment, personal fulfillment, and future projects’ (24.1 %). ‘With apprehension’ (33.3 %) was identified as the most frequent SA response. Findings showed a model for each nationality. AtA and SA for Portuguese elderly were explained by a three-factor model: ‘conciliated’, ‘young-at-heart’ and ‘involved’. A three-dimension model formed by ‘satisfied’, ‘attentive’ and ‘concerned’ was indicated as a best-fit solution for Romanian elderly.; CONCLUSIONS QoL and SA are strongly explained by increased likelihood of specific constructs in its definition. QoL was differently related to SA in older adults in both samples. Key Words: Multiple Correspondence Analysis; Older Adults; Portuguese; Quality of life; Romanian; Subjective Age.
(3042) Dyspnea and quality of life (QOL) in patients with COPD Ljiljana M. Danojevic, MD, Health Care Centre Grocka, Serbia; Srdjan R. Dimitrijevic, no; Zeljka D. Bozalo, Spomenak AIMS Dyspnea is a common symptom that accompanies a diagnosis of chronic obstructive pulmonary disease (COPD). It is simply
117 defined as the uncomfortable awareness of breathing and the need to increase the respiratory effort.; METHODS In order to determine if the intensity of dyspnea correlate with objective pulmonary function measurements and HRQOL, a cross-sectional, observational study of thirty one patients (6 female, 25 male) with COPD was undertaken. Level of dyspnea was measured with Medical Research Council (MRC) scale and Borg scale. Assessment included spirometry, 6-min walk test (6MWT), the Saint George’s Respiratory Questionnaire (SGRQ) and the 10 Item Respiratory Questionnaire Monitoring 10 (RIQ-MON 10).; RESULTS Relationships between parameters of the 6MWT (mean 363 ± 96.9 m), FEV1(mean 57.83 ± 20.34 %), FVC(mean 85.14 ± 19.10 %), the MRC scale (mean 1.68 ± 0.98), the Borg scale (mean 3.79 ± 2.60), the SGRQ scores (mean 51.79 ± 20.6), the RIQ-MON 10 (mean 18.68 ± 6.78) were evaluated with Spearman rank correlation coefficients. The correlations between 6 min walk distance (6MWD) and FVC ([rho] = 0.481), the 6MWD and FEV 1 ([rho] = 0,579), the 6MWD and MRC scale ([rho] = -0.660), the MRC scale and the SGRQ total scores ([rho] = 0.713), the MRC scale and the RIQ-MON 10 total scores ([rho] = 0,523), the Borg scale and total SGRQ scores ([rho] = 0.762) the Borg scale and the RIQ-MON 10 total scores ([rho] = 0.742) were all significant level of 0.01.; CONCLUSIONS Reduction in objective pulmonary function measurements such as FEV1 were not well correlate with the patient‘s perception of dyspnea and HRQOL. Impaired functional exercise capacity inversely correlate with dyspnea.
(3043) An extended study of the psychometric properties TCM-SAQ-QBS: Traditional Chinese Medicine Syndrome of Angina Questionnaire-Phlegm and Blood Stasis syndrome pattern Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Chen Zhihui, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Lv Meijun, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yu Changhe, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Ma Yuan, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Wang Yang, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS Due to the significant importance that syndrome pattern plays in the clinical practice of Traditional Chinese Medicine (TCM), a syndrome-specific questionnaire for assessing patients’ outcomes has already been established: it has been shown to be both reliable and valid. This study was conducted to refine the structure of TCM-SAQPBS for its extended use.; METHODS On the basis of the established version, we reorganized the ‘‘activity’’ domain. The already existed 10 items were combined into 6 items according clinical practice. Then a refined version with 30 items was applied in a cross-sectional study to collect data. Patients were recruited from Phlegm and Blood Stasis syndrome (PBS), and other syndromes (N-PBS). The Seattle Angina Questionnaire (SAQ) also tested for criterion validity. 240 patients were re-tested after 24 h. Internal consistency was tested by Cronbach’s alpha. The construct validity was assessed by exploratory factor analysis (EFA).; RESULTS After the adjustment of the structure, the 10 items in ‘‘activity’’ domain were reduced to 6 items: item 1 and 2 were combined into one item due to their similar exercise intensity; the wording of item 3 was changed from ‘‘walking 100 meters’’ to ‘‘daily activity, such as walking’’, which is more close to daily life; item 4 and 5 were combined into ‘‘daily physical activity,
123
118 such as climbing one stair or walking 500 meters’’; item 6, 7 were combined into item 10 as illustration part. Thus, a new version with 30 items was formed to apply in cross-sectional study. 997 (99.6 %) patients were eligible, of which 799 (age 59.1 ± 10.4 years; male, 51.3 %) in PBS group and 198 (age 58.2 ± 10.7 years; male 49.5 %) in the N-PBS group. Questionnaire was completed in 12.1 ± 7.5 min in PBS group, while 9.9 ± 7.3 min in N-PBS group. The test–retest correlations coefficients (CC) are greater than 0.874 between the same items and 0.972 for the global scores. Cronbach’s alpha coefficient is 0.850, and ranging from 0.817 to 0.955 for each domain. The CC between TCM-SAQ-PBS and SAQ was 0.536. Nine factors were extracted in EFA and were stratified into 6 domains as expected by the Conceptual Framework. The scores of syndrome domain between the two groups PBS 31.6 ± 4.8 versus NPBS 29.1 ± 5.7 were significantly different (p \ 0.001),which displays specificity.; CONCLUSIONS The new version TCM-SAQ-PBS was shown to have good psychometric properties. It can be extended into TCM clinical practice for angina patients’ outcomes.
(3044) Instrument and translation errors that can be caused by use of upgraded Indian fonts Kimberley Ahchak, Health Research Asia; Beng Li Ting, Health Research Asia, Singapore; Mona L. Martin, RN MPA, Health Research Associates, Inc AIMS Developing Clinical Outcome Assessments (COA’s) in Indian languages is often fraught with difficult font issues. The upgrading from Windows XP operating system to Windows 7 has brought about a number of improvements and upgrades in many of these Indian fonts. These font upgrades have also brought to light other issues that could affect the display of characters and document layout of existing linguistically validated instruments that were developed using fonts from Windows XP. This study aims to highlight the types of issues faced when a font is upgraded to a newer version.; METHODS The Short Form 36 version 2 (SF-36v2) was used in this study. A total of 4 languages and 4 Indian fonts with varying characteristics and properties were assessed using visual before-and-after comparisons made by experienced language consultants. All differences found between the old and upgraded font versions were documented. The font changes for each language were identified and then sorted into categories for: word conjugation, font rendering and visual display (such as character size, character thickness and word spacing).; RESULTS Results showed that 1 out of 4 languages tested had word conjugation corrected; 2 out of 4 languages had font rendering corrected. All 4 languages had changes in visual display that required manual adjustments before the document could be appropriately presented.; CONCLUSIONS In order to maintain the appropriate readability and comprehension of their instruments, developers of quality of life and clinical outcome instruments need to employ quality assurance checks when their measures use the upgraded Indian fonts in order to ensure that Indian characters and words are properly formed and that the visual display is appropriately presented.
(3045) Return to work and quality-of-life in workers with traumatic limb injuries: a 2-year repeated measurements study Wen-Hsuan Hou, E-Da Hospital & I-Shou University, Taiwan; Chung-Yi Li, Taiwan; Huey-Wen Liang, National Taiwan University Hospital, Taiwan AIMS To assess the impact of return-to-work (RTW) status on healthrelated quality-of-life (HRQoL) over a 2-year period in workers with traumatic limb injuries, and to elucidate factors that may contribute to
123
Qual Life Res (2012) 21:1–132 the association of RTW with HRQoL.; METHODS A sample of 966 injured workers was repeatedly surveyed on their RTW status, HRQoL (assessed by the 5-item EuroQol instrument), and activity/ participation at 1, 3, 6, 12, 18, and 24 months after injury. A series of regression models were used to examine the associations between HRQoL and RTW with sequential adjustment for explanatory variables such as personal, environmental factors, body structure and function, activity/participation, and post-injury period. The generalized estimating equations (GEE) approach was used for model fitting so as to account for within-subject correlations of HRQoL.; RESULTS Over a 2-year study period, 81.2 % of the study participants had one or more RTW episodes; 38.2 % of them successfully maintained their RTW status till the end. A significant positive association was found between RTW status and HRQoL. The association could largely be explained by the domains of activity/ participation. A higher HRQoL was associated with a shorter length of hospital stay, a lower level of perceived stress, frequent conduct of instrumental activities of daily living, and a longer post-injury period. A reduced HRQoL, on the other hand, was observed for participants with more depressive symptoms.; CONCLUSIONS RTW showed a positive and independent influence on HRQoL in workers with limb injury. Additionally, the activity/participation domains and the elapsed time since injury largely explained the association between RTW and HRQoL.
(3046) Health related quality of life as an independent predictor of long-term disability for patients with relapsing remitting multiple sclerosis Karine Baumstarck, France; Jean Pelletier, Neurology, Timone University Hospital, Marseille, France; Helmut Butzkueven, MS Unit, Melbourne, Victoria, Australia; Oscar Ferna´ndez, Institute of Clinical Neurosciences, Ma´laga, Spain; Peter Flachenecker, Neurological Rehabilitation Center Quellenhof, Bad Wildbad, Germany; Egemen Idiman, Department of Neurology, Dokuz Eylu¨l University, Izmir, Turkey; Sergio Stecchi, Multiple Sclerosis Unit, Department of Neuroscience AUSL, Bologna, Italy; Pascal Auquier, PhD, Laboratoire de Sante Publique, France AIMS To determine the value of the initial QoL level, assessed using both a specific-disease (the MusiQoL) questionnaire and a generic (the SF-36) questionnaire, to predict changes in the disability status of patients with multiple sclerosis (MS) while considering the key sociodemographic and clinical confounding factors, and to determine if specific-MS questionnaires predict more accuracy these changes than generic questionnaires.; METHODS Design: multicenter, multiregional, and longitudinal study. Main inclusion criteria: patients with RR-MS subtype and an EDSS score lower than 7.0. Sociodemographic and clinical data were recorded at baseline. Every 6 months up to month 24, QoL (MusiQoL and SF-36) was recorded. At 24-month, individuals were defined into two ‘disability change’ groups according to the neurological standards (EDSS change): the worsened and not worsened patients.; RESULTS Five hundred twenty-six patients were enrolled: 386 (83.7 %) not worsened and 75 (16.3 %) worsened patients at 24-month. The activity of daily living and the relationship with health care workers dimensions of MusiQoL questionnaire were independent predictors of change in the EDSS score after 24 months. The physical-functioning dimension of the SF-36 questionnaire predicted independently disability change after 24 months.; CONCLUSIONS Patient-reported baseline QoL levels provide additional prognostic information on MS disability beyond traditional clinical or sociodemographic factors. These findings reinforce the importance of incorporating a patient’s evaluation of their own QoL level during patient monitoring and the assessment of treatment effects.
Qual Life Res (2012) 21:1–132
119
(3047) Measuring subjective symptoms in COPD and their impact on quality of life Muirne CS Paap, University of Twente, Netherlands; Cees AW Glas, University of Twente; Benjamin Forstreuter, University of Twente; Job Van Der Palen, Medisch Spectrum Twente AIMS The St George’s Respiratory Questionnaire (SGRQ) is a very popular measurement instrument for assessing Quality of Life (QoL) in patients with chronic obstructive pulmonary disease (COPD). The SGRQ uses two steps to obtain an indication of the level of QoL: first it is assessed which complaints the patients suffers from, subsequently weights are assigned to indicate the distress associated with the complaints. The weights reported in the manual were based on the average distress associated with a symptom as reported by a sample of patients on a visual analogue scale (VAS) over two decades ago. Using fixed weights has some drawbacks; most importantly, it is assumed that each patient experiences the same amount of distress if a certain symptom is present. Our main aim is to model the answers to the multiple-choice items and the VAS scales simultaneously, allowing for individual differences in distress associated with respiratory symptoms.; METHODS Five-hundred patients with COPD are recruited by pulmonary physicians in six clinics in the Netherlands. The SGRQ enhanced with VAS scales is administered at home in a digital format accessible online. Multidimensional item response theory (MIRT) will be used to take into account the dependence structure of the responses to the SGRQ items and the continuous data derived from the VAS scales measuring associated distress. Models of different complexity will be compared and the one that shows the best fit will be selected. Finally, the resulting distress estimates will be compared with the estimates based on the SGRQ manual (using fixed weights).; RESULTS The results will be available at the time of the conference.; CONCLUSIONS This study illustrates the potential of using MIRT to model QoL that is assessed in a two-step manner. The advantage of this novel approach is that it can provide an indication of the experienced symptoms as well as the accompanying distress simultaneously and on person-level. If patients show a large variation in distress associated with a given symptom, it is advised to use personalized weights based on MIRT instead of the fixed weights that are currently in use.
(3048) A review of the existing EORTC quality of life measures in Spanish: response scales evaluation Maria Arnott, EORTC; Michael Koller, PhD, Center for Clinical Studies, Germany; Eva Elfriede Greimel, PhD, Medical University Graz, Austria; Andrew Bottomley, PhD, EORTC Headquarters, Belgium; Cheryl Whittaker, EORTC; Dagmara Kulis, EORTC, Belgium AIMS While EORTC’s standard translation procedure is followed in all new translations of EORTC questionnaires, cultural adaptations are performed when a language is spoken in more than one country. Spanish requires the most adaptations since it is officially spoken in 21 countries. We analysed the number and nature of these cultural
Argentina Chile Colombia Guatemala Mexico Peru Puerto Rico
USA Total
Not at all 2
3
3
–
3
3
3
4
A little
–
–
–
–
–
–
–
–
21 –
Quite a bit
–
–
–
–
–
–
–
–
–
Very much
–
2
–
1
–
–
–
1
4
adaptations based on the response scales.; METHODS We analysed 55 different Spanish adaptations of the EORTC measures available up to December 2011, performed with both the old approach of adapting the Spanish (Spain) version to the target country (policy before 2010), and with the new approach of pilot-testing this version first and then analysing the comments (policy from 2010 onwards). A comparison of the response scales was conducted to detect the most important differences and to assess whether they are necessary or can be avoided to limit the number of variants.; RESULTS A total of 55 Spanish cultural adaptations were performed: Argentina (9), Chile (6), Colombia (7), Costa Rica (2), Guatemala (2), Mexico (14), Peru (3), Puerto Rico (3) and USA (9). When analysing the response scales, we found the following different versions: (the results are presented in Table 1)The answer ‘‘Not at all’’ required 21 adaptations; ‘‘Very much’’ presented 4. ‘‘A little’’ and ‘‘Quite a bit’’ showed no need for adaptation. When we analysed the approach conducted for each case, the results showed there was consistency in the adaptations for the answers when using our new approach. In Spanish for Mexico, half of the variants (6 out of 12) adopted exactly the same change for answer 1 (‘‘En absoluto’’ to ‘‘Para nada’’) when using this approach.; CONCLUSIONS The large variety of adaptations makes it difficult to prove whether they are necessary, and therefore strict rules are now defined in the EORTC Translation Manual. With this new approach (pilot-testing the European version) we have tracked the most significant differences and it is the most recurrent that should be used for consistency in EORTC questionnaires. Our method may provide insight to standardise the cultural adaptations for Spanish-speaking countries.
(3049) Has the quality of systematic reviews of measurement properties improved over time? Caroline B. Terwee, PhD, VU University Medical Center, Department of Epidemiology and Biostatistics, Netherlands; Lidwine B. Mokkink, VU University Medical Center, Netherlands; Arnela Haagmans-Suman, VU University Medical Center, Department of Epidemiology and Biostatistics; Henrica C.W. de Vet, Sr., PhD, EMGO Institute VU Medical Center, Netherlands AIMS Systematic reviews of measurement properties of health status measurement instruments are important tools for selecting instruments for research and clinical practice. In a previous study on 148 reviews published up to 2007 we concluded that the quality of these reviews left much to be desired and should be improved. The aim of this study was to evaluate whether the quality of these reviews has improved over the past 5 years.; METHODS A systematic search was performed in Medline and Embase in March 2012 to identify all published systematic reviews of measurement properties of health status measurement instruments. The 100 most recent published reviews were appraised using a standardized quality checklist by two independent reviewers. The appraisal included the search strategy, article selection, quality assessment of the included studies, quality assessment of the instruments and data synthesis.; RESULTS Preliminary results of 27 reviews showed that reporting of the search strategy was improved (100 vs. 84 % in 2007). However, in 56 % of the recent reviews the search strategy was considered suboptimal with regard to search terms for instruments and measurement properties. Fewer studies evaluated the quality of the included studies (44 vs. 56 %), but more often a standardized checklist for quality assessment was used. The quality of the instruments was more often evaluated (59 vs. 52 %) and in 69 % a standardized set of quality criteria was used. Data synthesis (qualitative or quantitative pooling) was performed in 37 % of the studies. In 81 % of the studies a recommendation was given for the best instrument for a given purpose.; CONCLUSIONS Systematic reviews of measurement properties are
123
120 becoming increasingly important for evidence-based instrument selection, and support recent initiatives on outcome measurement such as the COMET initiative. The quality of these reviews seems to have improved over the past 5 years on some aspects, but there is still much room for improvement. Recently developed methodological tools, such as search filters for studies on measurement properties, and the COSMIN checklist for assessing the quality of included studies, should help to increase the quality of these reviews.
(3050) Assessing quality of life in new regions of the world: further development of the PsAQoL Mitra Ben-L’amri, BSc, MPhil, Galen Research Ltd, United Kingdom; Stephen P. McKenna, PhD, Galen Research Ltd., United Kingdom; Jeanette Wilburn, Galen Research Ltd; James Twiss, BSc, Galen Research Ltd, United Kingdom AIMS The PsAQoL is a measure of quality of life (QoL) specific to psoriatic arthritis (PsA). Measure content was derived from qualitative interviews conducted with UK patients. New language versions have since been developed for Hungary and Sweden. Interest in the PsAQoL has increased due to the need to determine changes in QoL associated with new biological treatments in multi-national clinical trials. There has been a trend in recent years for such trials to be conducted in new regions of the World rather than Europe and North America. An important question remains to be answered; can such measures provide valid assessment of QoL in these different cultures?; METHODS New adaptations of the PsAQoL are currently being produced for Eastern Europe (4), Europe (1), the Middle East (1), Central and South America (2), Australasia (1) and Asia (5). The measures were translated into the new languages using the two panel methodology required for needs-based measures and tested with local patients by means of cognitive debriefing interviews.; RESULTS Overall translations of the questionnaires into the new languages were successful. Cultural and language difficulties were experienced in some countries, particularly in the Far East. However, the translation panels were able to overcome these difficulties adequately. Cognitive debriefing interviews with 10–15 PsA patients in each country have confirmed that the adapted PsAQoL is acceptable to patients who found it easy to understand and complete.; CONCLUSIONS It remains to be seen whether these new adaptations meet the standards set by the existing European language versions (internal consistency [0.85), test–retest reliability coefficients: [0.85 and construct validity exhibited by their ability to distinguish groups of PsA patients that varied by perceived disease severity and general health and by correlating as expected with comparator measures. To test true cultural equivalence it is intended to use Item Response Theory analyses to determine whether respondents in the new regions answer the PsAQoL in the same way as those in Western countries. This will show whether the scales will work validly in the developing countries and whether data can be combined from different regions of the World.
Qual Life Res (2012) 21:1–132 Germany; Carolyn C. Gotay, PhD, UBC School Pop & Pub Health, Canada; Eva Elfriede Greimel, PhD, Medical University Graz, Austria; Madeleine T. King, PhD, Psycho-oncology Cooperative Research Group (PoCoG), Australia; David Osoba, FRCPC, QOL Consulting, Canada; Joseph Schmucker von Koch, PhD, University of Regensburg, Medical Ethics, Regensburg, Germany; Martin Taphoorn, Department of Neurology, MC Haaglanden, The Hague, The Netherlands; Joachim Weis, PhD, University of Freiburg, Psychooncology, Freiburg, Germany; Andrew Bottonley, PhD, EORTC, Quality of Life Department, Brussels, Belgium, on behalf of the European Organisation for Research and Treatment of Cancer (EORTC) and the Patient-Reported Outcome and Behavioural Evidence (PROBE) AIMS The Patient-Reported Outcomes and Behavioural Evidence (PROBE) project brings together investigators from a variety of disciplines to conduct retrospective quantitative analysis of data pooled from closed international randomised clinical trials (RCTs). Its aim is to understand the burden of cancer on patients’ lives from the patient’s perspective and to create better interpretive tools to guide clinicians in their decision making for improving their patients’ Health-Related Quality of Life (HRQOL).; METHODS All research analyses resulted from closed RCTs of the European Organisation for Research and Treatment of Cancer (EORTC), the German Gynaecological Oncology Working Group (AGO) and the National Cancer Institute of Canada-Clinical Trials Group (NCIC-CTG). Thirty EORTC, eighteen NCIC and two AGO RCTs contributed to the creation of a database covering 18 different cancer sites and 17,239 patients. Merged data from this international collaboration made use of the EORTC Quality of Life Questionnaire (EORTC QLQ-C30) and associated modules as well as clinical and survival data.; RESULTS Since 2009, over 20 pooled analyses have been conducted. We have presented 33 abstracts at major conferences (e.g. ASCO, ESMO, and ECCO), and published 6 papers (with 5 more submitted or in preparation) in high impact factor (IF) oncology journals (IF [ 12). Among other analyses, the studies of baseline HRQOL as a prognostic indicator suggest that HRQOL data increase the predictive accuracy by 6 % over clinical data alone. Pooled analyses of minimal important differences (MID) confirmed the estimate for the EORTC QLQ-C30 as [10, similar across cancer types. While such analyses provided valuable insights, occasionally some constraints still exist, such as data access, data ownership, transfer of data across member states, the need to standardise databases to facilitate pooled analysis, and funding, are long-term barriers.; CONCLUSIONS The results of the international PROBE project demonstrate the power of collaborative research in HRQOL to maximise the value of already collected patient-reported data. Our existing findings can more precisely inform health care providers, patients and their families. PROBE presents many future opportunities, including efforts to map predictive variables for each cancer site.
(3052) Impact of socioeconomic factors and co-morbidity on health-related quality of life in a multi-ethnic HIV cohort
(3051) The added clinical value of health-related quality of life research: major outcomes of the pooled analysis of phase III randomised clinical trials
Sherly George, PhD Student, University College Dublin, Ireland, Ireland; Michael Gibney, Institute of Food and Nutrition, University College Dublin; Fiona Mulcahy, St James’s Hospital, Dublin, Ireland; Mary Codd, University College Dublin
Efstathios Zikos, MSc, MA, EORTC HQ, Belgium; Divine Ewane Ediebah, MSc., EORTC HQ, Belgium; Corneel Coens, MSc, EORTC HQ, Belgium; Chantal Quinten, MSc, EORTC, Belgium; Jolie Ringash, MD, The Princess Margaret Hospital, Canada; Bryce B. Reeve, Ph.D., University of North Carolina at Chapel Hill, United States; Charles S. Cleeland, PhD, UT-MD Anderson Cancer Center; Henning H. Flechtner, PhD, Otto-von-Guericke University (KKJP),
AIMS To investigate the Health-related quality of life(HRQoL) of patients enrolled in the Dublin HIV Cohort study, with particular emphasis on socio economic variables and co-morbidity.; METHODS Administrative, demographic, socioeconomic, clinical and HRQoL data were collected from 528 HIV infected patients enrolled in Dublin HIV Cohort study. Patients completed Medical Outcomes Study HIV (MOS-HIV) Survey during their routine clinical visit. This is a
123
Qual Life Res (2012) 21:1–132 validated, self-administered and disease specific questionnaire. HRQoL was assessed under physical and mental health domains as physical health summary score (PHS), and mental health summary score (MHS). Scores ranged from 0 to 100, higher the score, better the HRQoL.; RESULTS Mean (±SD) age 41(±9) years, 68.9 % were male, 73 % Caucasian, 24 % Sub-Saharan African (SSA), 55 % employed, and 52 % with one or more co-morbid conditions. 89 % are on antiretroviral therapy (ART). Overall, median PHS was 56 (range 15–65) and median MHS 51 (range 11–67). Compared to males, females reported significantly lower MHS (52 vs. 47; p \ 0.05). An ethnic comparison within the gender identified statistically significant lower PHS and MHS among Caucasian females compared SSA females (PHS median 49 vs 58, p \ 0.001; MHS median 44 vs 51, p \ 0.01). Further exploration of the data identified that co-morbidity and unemployed had a negative impact on both PHS and MHS among Caucasian patients. Caucasian unemployed females with co-morbidity reported lowest PHS 38 (range 16–60); and MHS 36 (range 11–61). Caucasian unemployed males with co-morbidity also had lower HRQoL scores: PHS median (range) is 43 (15–63) and MHS median (range) is 41 (12–66). Among SSA, MHS was lower in unemployed men and women with co-morbidity (male 48; female 47). However they reported above average PHS irrespective of co-morbidity or employment status; CONCLUSIONS Co-morbidity and unemployment have negative impact on HRQoL of HIV patients in this cohort. Lower HRQoL among indigenous population compared to migrant population from SSA highlights the need for further exploration of data set in order to inform management strategies. Routine assessments of HRQoL using measuring tools relevant to HIV patients will identify risk factors to provide personalised care and attain best possible health and quality of life.
(3053) Quality of life (QoL) in older adults: survey data based on self-assessment versus face-to-face interview Liv Halvorsrud, Diakonova University College, Norway, Norway; Mary Kalfoss, DrPH, Diakonova University College, Norway AIMS The aim of this study was to explore the quality of data obtained on QoL assessment among older adults with respect to two modes of instrument administration; self assessment and face-to-face interviews.; METHODS Using quota stratified sampling (gender, age, geographical representation), 490 respondents were drawn from Statistics Norway. 401 completed postal surveys (mean 75.1 years, female 55 %, partner 62 %) and 89 completed face-to-face interviews (mean 78.6 years, female 73 %, partner 28 %).; RESULTS Approximately half of the items on the WHOQoL Old showed significantly higher mean values in the postal sample; however, there were fewer missing answers in the interview sample. Three multiple regression analyses were conducted to explore statistical associations based on the mode of administration and other variables. By adjusting for sociodemographic variables, we found that reduced energy made the strongest contribution to overall QoL in both administration methods (self-assessed: B = 0.33, b = 0.36, p = 0.00, explaining 18 % of the variance; interview: B = 0.26, b = 0.29, p = 0.01, explaining 11 % of the variance). By using the same adjustments, reduced memory was shown to have a significant contribution only in the interview sample (self-assessed: B = 0.33, b = 0.36, p = 0.75, explaining 3 % of the variance; interview B = 0.26, b = 0.29, p = 0.03, explaining 9 % of the variance). Finally, by using the same adjustments in a third model, with the independent variables memory and energy, reduced energy made the most significant contribution in both samples (self-assessed: B = -0.35, b = -0.38, p = 0.00, explaining 17 % of the variance; interview sample: B = 0.21, b = 0.24, p = 0.05¸ explaining 14 % of the variance).; CONCLUSIONS Data suggest that there may be more vulnerability towards giving socially desirable answers, especially for questions of a sensitive nature among those who were interviewed. Problems with
121 reduced energy were especially relevant with both modes of administration. Results underline the importance of taking into consideration energy and memory problems as potential threats to data quality when interpreting QoL scores using both self assessment and interviews.
(3054) Development of asthma-specific quality of life items for item banking Cathy Sherbourne, PhD, The RAND Corporation; Nicole K. Eberhart, RAND Corporation; Maria Orlando Edelen, PhD, RAND Corp.; Brian D. Stucky, RAND Corporation, United States; Marielena Lara-Greenberg, RAND Corporation; Nancy Sin, UC Riverside AIMS Asthma is a common chronic disease that can have a substantial impact on those who suffer from it. A key goal in asthma treatment is the improvement of health-related quality of life (QOL), defined in terms of how much symptoms bother or matter to the patient in different areas of life. However, existing measures of asthma-specific QOL often confound QOL with asthma symptoms and functional impairment. The current study is addressing these limitations by using state-of-the-art methods to develop, evaluate, and test the validity of item banks that assess QOL in adult asthma patients. The qualitative item review process will be discussed, including themes that emerged from focus groups and a comprehensive literature review of asthma measures.; METHODS We conducted five focus groups with 50 asthmatic adults and systematically reviewed the literature to collect existing asthma-specific scales and items. Using text from the focus groups, we employed the method of free pile sorting to identify key QOL dimensions and used a ‘‘binning and winnowing’’ process to develop a pool of items based on what people said during the focus groups, supplemented with items from the literature. Cognitive interviews were conducted to refine item wording and format before field testing via the internet on 2000 asthmatic adults.; RESULTS Focus group participants varied in terms of demographics and severity of asthma. A total of 643 QOL statements were identified from transcripts and sorted into 24 initial bins which were subsequently used to generate a pool of about 115 items in 20 conceptual domains that were retained to carry forward to the field test. Cognitive interviews identified few problems in item wording and supported a 4-week recall period, and 5-choice response options.; CONCLUSIONS We used rigorous qualitative methodology to develop a set of items that are being fielded to a large national sample of adults with asthma. Analyses of these data will confirm conceptual domains and identify items to be incorporated in item banks. The measures can be used to monitor QOL in asthmatic patients, to compare the effectiveness of different treatments, and to clarify relationships between asthma control, impairments, and QOL.
(3055) Satisfaction with the drug therapy among patients with coronary heart disease: psychometric performance of the Brazilian Version of the Treatment Satisfaction Questionnaire for Medication—TSQM (version 1.4) Ana Carolina S. Liberato, Master’s Student, University of Campinas, Brazil; Roberta CM Rodrigues, University of Campinas; Thaı´s M. Spana, University of Campinas; Fernanda F. Januzzi, University of Campinas; Maria Cecı´lia BJ Gallani, Universite´ Laval AIMS This study was aimed at evaluating the practicability, acceptability, ceiling and floor effects, reliability, construct and criterion validity of the Brazilian version of the Treatment Satisfaction Questionnaire for Medication—TSQM (version 1.4).; METHODS The study was conducted among 100 patients with coronary heart disease in an outpatient cardiologic clinic in Sa˜o Paulo, Brazil. Data were
123
122 collected between June/2010–May/2011 at two different moments. At baseline, sociodemographic and clinical data were gathered. TSQM and the scale for measuring generic satisfaction with medication were applied by interview in order to measure Satisfaction with the drug therapy. Adherence to drug use was measured by the Morisky Self-Reported Scale and by quantifying the proportion of adherence. Seven days after baseline, the TSQM was reapplied (test–retest) in a major portion of individuals (n = 73) who participated of the first application. Statistical analyses concerned: ceiling and floor effects (10 % best and worst possible results of the scale), practicability (time spent during TSQM application), acceptability (percentage of unanswered items and proportion of patients who did not answer all items), reliability (regarding internal consistency and stability of the measure) and validity (Spearman correlation’s coefficient to test the relation between the scores of the Brazilian TSQM and the Morisky Scale, proportion of drug adherence and the generic measure of satisfaction with drug therapy).; RESULTS Regarding practicability, short time for application of the TSQM was evidenced (4.6 min). As for acceptability, all patients answered 100 % of the items. Ceiling effect was verified in Convenience and Global Satisfaction domains. Floor effect was not observed. Evidences of reliability were demonstrated in all domains of the Brazilian TSQM except for Convenience. Significant positive correlations of low-moderate magnitude were verified between the TSQM scores and the general measure of satisfaction.; CONCLUSIONS The Brazilian version of TSQM presents evidence of practicability, acceptability, reliability and validity. Further studies with diverse populations are recommended to abroad its validity.
(3057) Development of a computer-adaptive version of the Forgotten Joint Score for outcome assessment after total knee arthroplasty Johannes M. Giesinger, PhD, Innsbruck Medical University, Austria; Markus S. Kuster, Department of Orthopaedic Surgery, Royal Perth Hospital; Bernhard Holzner, PhD BE, Univ. Hospital, Innsbruck, Austria; Karlmeinrad Giesinger, Department of Orthopaedic Surgery, Kantonsspital St. Gallen AIMS To develop a computer-adaptive version of the Forgotten Joint Score (FJS), a patient-reported outcome (PRO) measure for the assessment of outcome after total knee arthroplasty (TKA).; METHODS Item response theory analysis for the current paper– pencil version of the FJS was performed to obtain psychometric item characteristics required for computer-adaptive testing (CAT) of the FJS. Data from 426 patients after TKA were available for analysis. Within a feasibility study we administered the new FJS-CAT measure to a sample of 60 patients and assessed the patients’ perspective on electronic CAT administration; RESULTS Mean age at surgery was 68.1 years (SD 10.4) and 56.7 % were women. The FJS items showed good unidimensionality and 10 items could be fitted to a Rating Scale Model (RMSEA = 0.06). No substantial differential item functioning was found for these items. The full item bank provided a measurement precision of 0.46 logits, compared to a standard error of 0.50 for a 6-item CAT, 0.61 for a 3-item CAT and 1.05 for a single item. In a feasibility study 90 % of the patients preferred electronic questionnaires administered on a tablet PC over paper–pencil questionnaires.; CONCLUSIONS The FJS was developed to overcome limited discriminatory abilities that have been shown for several PRO measures (e.g. WOMAC Score or Knee Society Score) in patients with good to excellent outcome. The new computer-adaptive FJS version showed high measurement precision with a low number of items. Electronic CAT administration showed high feasibility and patient satisfaction.
123
Qual Life Res (2012) 21:1–132 (3058) Clinician and patient ratings of severity compared with short assessment tools for urinary incontinence Janet E. Sansoni, MSc, University of Wollongong, Australia; Glenn Fleming, University of Wollongong; Graeme Hawthorne, PhD, Univ of Melbourne, Australia; Nicholas Marroszeky, University of New South Wales Global, Australia AIMS A 5 item tool for assessing urinary incontinence, the Revised Urinary Incontinence Scale (RUIS), was developed from an examination of data drawn from a community sample (n = 2915) using classical psychometric analyses. Recently, a clinical validation of this tool was undertaken and this presentation examines the relationship between short incontinence assessment scales, clinician and patient ratings of severity.; METHODS Patients seeking treatment were recruited from 7 Australian incontinence clinics (specialist and community). The study examined clinical and patient definitions of continence status and treatment outcomes. The study protocols contained the RUIS, three other continence specific questionnaires, clinician and patient ratings of severity and other clinical indicators. There were 195 urinary patients recruited at baseline.; RESULTS At baseline all the short scales for the assessment of urinary incontinence correlated significantly with each other (r =/[ 0.72) and with patient ratings of severity (r = 0.50–0.70) but all had lower correlations (r = 0.29–0.37) with clinician rated severity (all correlations; p \ 0.01). The correlation between clinician and patient rated severity was also of a lower magnitude (r = 0.36) sharing only 13 % of the variance. The item concerning frequency of urinary leakage had a higher correlation with clinician rated severity (r = 0.40) than did the volume of urinary leakage (r = 0.26) whereas there were similar correlations (r = 0.58 and 0.52) for these elements with patient rated severity. There were significant differences between the correlations for the clinician and patient ratings for these items. Clinician rated severity in the urinary sample also generally had higher and more significant associations with variables such as BMI, weight, and the SF-36 global assessment of health status than did the urinary incontinence scales and patient rated severity.; CONCLUSIONS Although clinician and patient severity ratings are associated with the frequency of urinary leakage clinician ratings have a lower association with the volume of leakage. Clinician ratings are slightly more associated with the overall perception of the health characteristics of the client. Given these differences in approach it is important that the patient perspective is also captured by the use of patient reported measures for incontinence.
(3059) Linking of at-work disability scales to the ICF Vanitha Arumugam, PT, MPT, University of Western Ontario, Canada; Joy MacDermid, PhD, McMaster University, Canada; Ruby Grewal, University of Westren Ontario AIMS The purpose of the current study is to analyse the content of (1) 25-Items Work Limitations Questionnaire (WLQ-25); (2) Stanford Presenteeism Scale (SPS-6) and (3) Rheumatoid Arthritis Work Instability Scale (RA-WIS) by linking them to the International Classification of Functioning Disability and Health.; METHODS Two researchers linked the items of the WLQ-25, SPS-6 and RA-WIS to the ICF independently. The percentage agreement was calculated between the researchers. The linking procedures were carried out following the eight standardised linking rules proposed by Cieza and colleagues.; RESULTS The 25 items of WLQ-25 were linked to 70 ICF categories from 10 chapters; the 6 items of SPS-6 were linked to 11 ICF categories from 2 chapters and RA-WIS’s 23 items were linked to 35 ICF categories from 6 chapters. Most of the meaningful concepts from all 3 questionnaires were linked to activity and participation domain while a substantial number of concepts were not
Qual Life Res (2012) 21:1–132 defined by the ICF.; CONCLUSIONS The results infer that specific content of the 3 questionnaires are covered by ICF. However, to get a global picture of health status of the employee measures of health and disability should be used as an adjunct to presenteeism scales.
(3060) Quality of life in people living with HIV in a resource limited setting-Uganda Doris M. Mwesigire, MD, MAKERERE UNIVERSITY, Uganda AIMS The availability of ART and improved access in resource limited settings has delayed disease progression and prolonged survival. The recent WHO guidelines increased the cut off to initiate ART to a CD4 count of 350 and below in these settings. The aim of this study is to compare the Global Quality of life between HIV patients eligible for antiretroviral therapy (ART) and patients that are not eligible for antiretroviral therapy.; METHODS We recruited patients attending the HIV clinic at Mulago Hospital, the national referral hospital in Uganda. Consenting adults above the age of 18 years were enrolled over a period of 5 days. Patients eligible for ART had CD4 count of 350 or less and were scheduled to start taking ART on the day of the interview. The non-ART patients had a CD4 count above 450 and were receiving supportive care in the same clinic, these were randomly sampled. All the patients were assisted by a research assistant to complete the Global Person Generated Index (GPGI) questionnaire.; RESULTS Of the 25 patients 68 % were female, 36 % eligible for ART and 64 % not eligible for ART. These proportions were representative of this clinic population. The following we given as what was important to the patients at that time; their health, well being of the children, well being of their parents, well being of other family members, relationships, employment/work, housing, property and food. Patient’s health, well being of family members and employment/work were the areas mentioned almost by everyone as important attributes of their life. The average GPGI score for those eligible for ART was 61 % and those not eligible for ART 69 %; CONCLUSIONS Preliminary results from this study show that economic and social aspects are very important factors in the lives of people living with HIV and should be underscored in their management. There is very little difference in the Global quality of life between patients eligible for ART and those that are not eligible for ART.
(3061) What do we assess about health when we assess self-reported general health? Cristina N. Bostan, University of Lucerne, University of Lucerne, Lucerne, Switzerland, Switzerland; Cornelia Oberhauser, Ludwig-Maximilians-Universita¨t, Germany; Alarcos Cieza, Ludwig-Maximilians-Universita¨t, Munich, Germany AIMS According the World Health Organization’s (WHO’s) International Classification of Functioning, Disability and Health (ICF), capacity is the level of functioning in health and health-related domains measured as the internal capacity of a person and performance is the level of functioning of a person in the current environment. The primary aim of this study was test the WHO assumption, that health, in particular self-reported general health, is more related to capacity than to performance, using empirical data and modern statistical techniques. The secondary aim was to explore the role of the environment in explaining performance.; METHODS Psychometric study on data collected within the 2008 Spanish Survey on Disabilities, Independence and Dependency Situations. The study population consisted of two residence-based population samples, one community-dwelling and one institutionalized. The following analysis steps were performed: (1) the
123 Samejima’s Graded Response Model (GRM) was employed to select the items that can be integrated in two different scales, one for capacity and one for performance, (2) the Spearman correlation coefficient was used to compare the strength of the association between self-reported general health and capacity persons‘ level with the strength of association between self-reported general health and performance persons‘ level, (3) the Generalized Additive Model was used to identify the environmental factor that best explain performance persons‘ level.; RESULTS 26 capacity items and 18 performance items satisfied the GRM assumptions and two different scales could be built based on them. In both samples included in the analysis, self-reported general health was more related to persons‘ performance levels (r = 0.41— community-dwelling sample, r = 0.37—institutionalized sample) than to persons‘ capacity levels (r = 0.35—community-dwelling sample, r = 0.30—institutionalized sample). The access to the personal and technical aids, the social support of the family and friends are significant environmental factors that explain performance persons ‘levels.; CONCLUSIONS Against our hypothesis, this study shows that performance is more related to self-reported general health than capacity. We were able to identify a number of environmental factors associated to performance. This study also shed light on the challenges of assessing capacity and performance at the general population level.
(3062) Surgical decision-making in women with hereditary breast cancer risk who develop breast cancer Marlene H. Frost, PhD, Mayo Clinic; Tanya L. Hoskin, Mayo Clinic; Constance I. Williams, Mayo Clinic; Teresa M. Allers, Mayo Clinic; Amy C. Degnim, Mayo Clinic; Lynn C. Hartmann, Mayo Clinic; Judy C. Boughey, Mayo Clinic AIMS Women with hereditary breast cancer risk who develop breast cancer face complex surgical decisions for both their index breast and their contralateral breast. The aims of this study were to examine women’s decision-making related to their surgical choice and the difficulty they experience with this choice.; METHODS 73 women with hereditary breast cancer risk diagnosed with unilateral breast cancer completed the Stage of Decision Making and Decisional Conflict Scale prior to their Breast Clinic appointment and again following their consultations. Demographic and surgical information was obtained from the medical record.; RESULTS Despite hereditary risk, most women came to their consultation with limited knowledge about surgical options (from separate decision readiness questions: 59 % did not know options, 70 % did not know option benefits, 80 % did not know option limitations). Prior to consultation, 43 % of women were still considering options, 26 % felt close to making a decision, 15 % had not begun to think about their options, and 16 % had made their surgical decision before arrival (although 75 % of these were still willing to consider other options). After consultation, most women had made a decision they felt was unlikely to change (90 %), with 7 % who would still consider a change. Women were slightly more likely to select mastectomy (54, 44 % of whom also selected contralateral prophylactic mastectomy) than breast conservation (46 %). The minority who had made their surgical decision prior to their consultation were likely to follow through with that surgical option (71 %). While over half of the women (58 %) initially reported concern, distress, wavering in decision, or questioning what was important in relation to decision-making, fewer women expressed these difficulties after their consultation (26 %, p B 0.0001).; CONCLUSIONS Most women with hereditary risk diagnosed with breast cancer present to the Breast Clinic with limited information or preconceived ideas about surgical options. The information obtained during their Breast Clinic visits was helpful in moving the majority of women to a decision that seemed unlikely to change.
123
124 (3063) Sarcoma patient’s quality of life: understanding clinical and social-demographical aspects Ana Maria Almeida, Instituto Portugueˆs de Oncologia, Portugal; Teresa Sequeira, Portugal; Joaquim Gonc¸alves, Polytechnic Institute of Ca´vado & Ave, Portugal; Cla´udia C. Ribeiro, Universidade Cato´lica Portuguesa, Portugal; Matilde Ribeiro, Portuguese Oncology Institute, Porto, Portugal; Augusta Silveira, University Fernando Pessoa, Portugal AIMS Sarcoma is a heterogeneous group of tumors affecting commonly young patients and is related with both endogenous and exogenous risk factors. Therapeutic modalities for these patients include: function-preserving surgery and adjuvant therapies (radiotherapy and chemotherapy) and more recently intercepted with cellular and molecular biology approaches in order to minimize the negative therapeutic impact in patient’s quality of life (QoL). The present study aims to relate sarcoma patient’s QoL scores with clinical and socio-demographic aspects.; METHODS Patients (n = 53) from the Skin Clinic of the Portuguese Oncology Institute participated in this study after given their informed consent. Social and demographical data were collected both using semi-structured interviews and clinical processes consultation. Sarcoma patient’s QoL was assessed using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30) using the validated informatics platform (Quality of Life Informatics Platform). Scales were linearly transformed into scores (0–100). Kruskall-Wallis and Mann-Whtiney tests were used as nonparametric tests.; RESULTS Non-metastatic patients (83 %) when compared with the metastatic presented always the highest QoL scores for all functional domains exception made for Cognitive function. The lowest functional score found was ascribed to metastatic patients Role function (69.4). When considering age distribution, a pattern was observed: QoL scores increased until a maximum and decreased linearly then on. It was observed that 40–49 year old patients presented the highest QoL scores for Physical (93.5), Role (90.4) and Social (91.7) function subscales. Emotional function presented a maximum (89.1) in younger patients (30–39 years old) and the Cognitive function even earlier—the highest scores (96.4) were found in the youngest patients (20–29 years old). Married patients score better for most functional scales: Physical (90.0), Role (80.0) and Cognitive (90.1) and revealed to be similar for Emotional and Social function.; CONCLUSIONS Most functional domains are severely impaired in metastatic patients suggesting that early diagnosis does make a difference considering QoL outcomes. The most significant QoL functional burden observed was Role domain, indicating to be age-dependent. Patient reported outcomes such QoL should be analyzed with socio-demographic and clinical variables in order to better understand the sarcoma disease and the sarcoma patient.
(3064) Sense of belonging to the community of care and quality of life in chronic illness Chiara Guglielmetti, Universita` degli Studi di Milano, Italy; Silvia Gilardi; Universita` degli Studi di Milano, Italy; Marco Lauriola, Universita` degli Studi di Roma Italy AIMS Although the application of the construct of sense of community in organizational settings, especially in the educational area, has shown that the perception of belonging to a community contributes to increasing social and subjective well-being of users, this construct has not been used in health organizations. Referring to a previous research on Beta Thalassemia Major (Guglielmetti et al. 2012), the present study aims to analyze whether and under what
123
Qual Life Res (2012) 21:1–132 conditions the construct ‘‘sense of belonging to the community of care’’ positively affects the quality of life of a heterogeneous group of chronic patients. METHODS A sample of 316 patients with various chronic diseases treated in outpatient clinics at 7 hospitals in Italy were given a self-report questionnaire consisting of (a) a scale of 14 items measuring the construct of sense of belonging to the community care (SACC)—operationally defined as the feeling of being part of a supportive network of relationships characterized by positive links and a sense of mutual responsibility for the wellbeing of the members, (b) two batteries of FACIT (quality of life and satisfaction with the quality of care). RESULTS The analyses highlight the mediating role of the sense of belonging to the community care in the relationship between factors of quality of care and quality of life of different types of chronic patients. In particular, the sense of belonging to the community care is associated with a better quality of life, especially in terms of emotional well-being and social relationships. CONCLUSIONS Encouraging the development of a sense of belonging has important implications for the welfare of chronically ill patients.
(3065) Liker-t or number rate scale? A comparison study on Seattle Angina Questionnaire Yu Changhe, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Yang Guanlin, MD, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Chen Zhihui, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Zhang Huiyong, MD, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; LV Meijuan, Liaoning university of Traditional Chinese medicine, Peoples Republic of China; Zhang Zhe, Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Peoples Republic of China; Ma Yuan, Liaoning University of Traditional Chinese Medicine, Peoples Republic of China AIMS Response options are considered vital for the measurement to evaluate the Quality of life. This study aims to compare two response options, Liker-t and Number Rate Scale (NRS), in Seattle Angina Questionnaire (SAQ).; METHODS A cross-sectional study was conducted to collect data. The angina patients completed Liker-t and NRS SAQs. The Liker-t scale is an ordinal value, where value 1 implies the worst QOL, and 5 (or 6) implies the best QOL. The NRS is a linear scale from 0 to 100 marked at intervals of 10. Verbal descriptions are indicated at either end and at the middle of the scale to assist patient comprehension. Patients can thus indicate their perceived QOL on the NRS where value 0 implies the worst QOL and value 100 implies the best. The survey investigates which response options patients prefer to use (Liker-t or NRS SAQ). The missing data were used to evaluate the feasibility of each option. The correlations between the same domains were examined to compare their consistency with the whole questionnaire. The reliability and validity were assessed.; RESULTS 343 patients (age 59.23 ± 20.02 years; 120 male, 35.0 %) were recruited. There are 2 missing data in Liker-t SAQ, but 84 missing data in NRS. Of the 84 missing data 11 are from Item 3 (angina frequency), and 71 are from Item 4 (doses of nitrate). The Liker-t score contained the option ‘‘none’’. This option was not available in the NRS. NRS should be measured from greatest to 0 frequency for Item 3, and from highest to 0 dose of nitrate for Item 4. In the survey, 203 (59.18 %) patients preferred the Liker-t scale mainly due to its simplicity (107, 52.7 %) and clarity (68, 33.5 %). 138 (40.23 %) preferred the NRS for its accuracy (85, 61.6 %) and subjective nature (24, 17.4 %). The correlation coefficient of Liker-t and NRS SAQ is 0.950 (p \ 0.01), ranging from 0.934 to 0.982 for each domain (except angina frequency domain). Both Liker-t and NRS SAQ presented good reliability and validity.; CONCLUSIONS
Qual Life Res (2012) 21:1–132 Both Liker-t and NRS presented good psychometric properties. The Liker-t SAQ is easier to administrate, but the NRS has a board range to be more accurate. However, we may adjust NRS verbal description for Items 3 and 4 and then combine both Liker-t and NRS to form a new SAQ.
(3066) Values of health in patients with systemic lupus erythematosus Cla´udia C. Ribeiro, Universidade Cato´lica Portuguesa, Portugal; Augusta Silveira, University Fernando Pessoa, Portugal; Teresa Sequeira, Portugal; Isabel Santos, Hospital Santa Maria da Feira; Catarina Ribeiro, Universidade Fernando Pessoa; Carlos Vasconcelos, Hospital Santo Antonio Centro Hospitalar do Porto AIMS Systemic lupus erythematosus (SLE) is a pervasive disease with wide-ranging effects on physical, psychological and social wellbeing. Health economics researchers are concerned with the measurement of the health related quality of life, especially eliciting individuals’ preferences regarding specific health outcomes. This type of measurement can be integrated in clinical decision in patients with SLE, planning and health policy models. Example of these instruments is SF-6D questionnaire, a new single summary preferencebased measure of health derived from the SF-36. AIM The aim of this study was to assess health utilities for patients with SLE; METHODS 71 patients with SLE, from Unidade de Imunologia Clinica—Hospital Santo Anto´nio-Centro Hospitalar do Porto who responded to the SF-36v2 were assigned to an SF-6D score, provided the 11 items used in the SF-36, to obtain the self assessment of their health status and quality of life. We used the Portuguese version of the measurement instrument SF-6D and the econometric algorithm created by Brazier to generate utility values for different health state.; RESULTS The SF-6D and preferences derived from the SF-36v2 showed significant correlations with each other, with Pearson coefficients varying from 0.70 to 0.85 (p \ 0.01). The mean utility value obtained by the SF-6D to the whole patients with SLE was 0.76, ranging from 0.47 to 1.00 and associated to a standard deviation of 0.079. Larger deficit levels were found in ‘‘role’’, ‘‘pain’’, ‘‘mental health’’ and ‘‘vitality dimensions’’. Lower utility values were assigned by women and elderly. Less educated individuals, living in rural areas, divorced and separated also showed, in general, lower utility values.; CONCLUSIONS We conclude that the measurement instrument SF-6D is useful and effective in measuring the health-related quality of life in patients with SLE.
(3067) Translation and psychometric testing of the German version of the Health Education Impact Questionnaire (heiQ) Michael Schuler, University of Wuerzburg; Sandra Nolte, PhD, Deakin University, Australia, Germany; Katja Spanier, M.A., Hannover Medical School, Coordination Centre for Applied Rehabilitation Research, Germany; Gunda Musekamp, University of Wuerzburg, Germany; Inge Ehlebracht-Koenig, Rehazentrum Bad Eilsen der Deutschen Rentenversicherung Braunschweig-Hannover; Christoph Gutenbrunner, Hanover Medical School; Hermann Faller, University of Wuerzburg; Richard H. Osborne, Deakin University; Monika Schwarze, Medical School Hannover, Germany AIMS To improve the evaluation of German patient education interventions, the Federal Ministry of Education and Research (BMBF) funded a 3-year project to translate, adapt, and implement the Australian Health Education Impact Questionnaire (heiQ) in Germany. The heiQ has been used in several countries and across
125 languages as a national quality and monitoring tool for patient education programs and applied as an outcomes assessment instrument in clinical trials. It consists of the following eight separate domains: Health directed activities, Positive and active engagement in life, Emotional distress, Self-monitoring and insight, Constructive attitudes and approaches, Skill and technique acquisition, Social integration and support, and Health service navigation. The main aim of this paper is to describe the outcomes of the comprehensive psychometric evaluation of the German version of the heiQ.; METHODS The German heiQ was tested in a sample of 1,202 patients using confirmatory factor analyses (CFA); the sample was split into a calibration (n = 603) and a validation sample (n = 599). Reliability was tested using Raykov’s Composite Reliability Coefficient.; RESULTS CFA revealed that five of the eight heiQ scales initially showed inadequate fit in at least one fit index; however, after freeing one error covariance in respective models, fit indices were acceptable. Factor loadings were mostly acceptable (between 0.5 and 0.9); the only exception was heiQ scale Self-monitoring and insight showing some coefficients between 0.4 and 0.5. Raykov’s Composite Reliability Coefficients were moderate, ranging from 0.71 to 0.87.; CONCLUSIONS Overall, the German version of the heiQ has good to very good validity and reliability. It can be recommended as a measure for patient education program evaluation in Germany. It will deliver important data for the stakeholders of Germany’s patient education programs, generating new evidence of the value of these interventions to inform quality improvement initiatives and determine program effectiveness.
(3068) Towards a meaningful patient report outcomes measures (PROMs) for patients with MPSIII Vanessa M. Ferreira, Lysogene, France AIMS Mucopolysaccharidosis Type III (MPS III) (Sanfilippo Syndrome) are a group of severe genetic diseases that lead to a wide range of disabilities (physical, emotional and psychosocial). An experimental treatment aimed at treating this pathology is currently being investigated by LYSOGENE. Patient-reported outcome measures (PROMs) provide a means of gaining an insight into the way patients perceive their health and the impact that treatments or adjustments to lifestyle have on their quality of life. In the last decade, Health-related Quality of Life (HRQoL) measures have been increasingly included as outcome measures in clinical trials. We aim to assess MPSIII quality of life outcomes in the near future of both, patients and carers.; METHODS A structured review of existing PROMs instruments was performed.; RESULTS Of all the papers included in the study, several were generic HRQoL instruments for use in children with rare diseases, such as the Child health and illness profile (CHIP), KIDSCREEN and PEDSQL. Others, like ConQoL or CFQ are disease specific for Congenital cardiac diseases and Cystic fibrosis, respectively. Three questionnaires covering specific symptomatic aspects of MPS III were identified. Importantly, there is no instrument to evaluate the overall HRQoL in MPSIII patients.; CONCLUSIONS This study provides a detailed description of several Health-related Quality of Life measures as outcomes in the domain of Rare Diseases. Taking into account the burden of these diseases on the quality of life of families, it is likely that a tool targeted to MPSIII would provide innovative information about the best care for these vulnerable children and their families, helpful information in designing clinical trials and on how to assess the impact on parental quality of life (QoL) of children’s affected by this pathology. In addition, there is still an enormous area to explore at the level of the implementation of disease-specific HRQoL instruments targeted to Rare Diseases. Our methodology can be used by other Rare Diseases.
123
126 (3069) Revision of an instrument for measuring patient experiences with injury care Niklas JH Bobrovitz, BHSc(h), University of Calgary, Canada; Maria-Jose Santana, PhD, University of Calgary, Canada; Henry T. Stelfox, University of Calgary AIMS There are very few patient-centered measures of the quality of injury care. The aim of the current project was to conduct cognitive interviews and test a newly developed survey designed to measure patient experiences with acute injury care.; METHODS Cognitive interviews were conducted with ten adult injured patients and ten family members. We used a semi-structured interview guide and asked retrospective probes to identify: participant item interpretation, the information participants recalled when answering items, and participants preferences for item wording.; RESULTS After three rounds of cognitive interviewing we identified four issues. (1) Four items contained ambiguous terminology that was inconsistently interpreted by participants: experiencing ‘well-coordinated transfers’, being ‘oriented to new departments’ after transfers, experiencing ‘organized care’, and experiencing ‘unsafe’ care. Participant feedback was used to revise these items. (2) Two items did not measure the constructs they were intended to measure and were eliminated. First, participants did not think of technical skill when assessing ‘skill and competence of caregivers’. They assumed providers were skilled and instead assessed other provider traits (e.g. communication skills) when responding. Second, participants interpreted being ‘included in decision making’ as simply being provided information. (3) One item, ‘information about the recovery process and timeline’ was eliminated because healthcare providers were often unable to provide timeframes for patient recovery and in these instances the item did not function well. (4) Two items were found to be redundant and were eliminated. An item on respecting patient ‘values and personal preferences for care’ overlapped with an item on respecting patient ‘cultural, spiritual, and religious preferences’ and an item on ‘kind and friendly’ providers overlapped with an item on how often healthcare providers ‘introduced themselves and explained their role in patient care’. Using these results we revised our instrument. The measure now consists of 7 domains (information and communication, transfers and patient transport, pain management, comfort, interpersonal care, safety, and equality) and 35 items.; CONCLUSIONS After three rounds of cognitive interviews we have revised our measure to assess patient experiences with injury care. A multi-centre study has been initiated and results will be available in the fall of 2012.
(3070) Analysis of symptom grading thresholds for patients and clinicians using the common terminology criteria for adverse events (CTCAE) Thomas Atkinson, Memorial Sloan-Kettering Cancer Center, United States; Yuelin Li, PhD, Memorial Sloan-Kettering Cancer Ctr., United States; Lauren Rogak, Memorial Sloan-Kettering Cancer Center; Natalie Barragan, Memorial Sloan-Kettering Cancer Center; Mary Shaw, Memorial Sloan-Kettering Cancer Center; Liora Pollick, Memorial Sloan-Kettering Cancer Center; Sean Ryan, Memorial Sloan-Kettering Cancer Center; Narre Heon, Memorial SloanKettering Cancer Center; Antonia V. Bennett, PhD, Memorial Sloan-Kettering Cancer Center; Ethan Basch, MD MSc, Memorial Sloan-Kettering Cancer Center AIMS Adverse symptom reporting is essential in clinical trials and drug labeling to assess and ensure patient safety. The standard approach to collecting this information in cancer trials is clinician reporting using the Common Terminology Criteria for Adverse Events (CTCAE), which rates symptoms based upon descriptive clinical criteria. The assignment of a given symptom grade
123
Qual Life Res (2012) 21:1–132 (i.e., Grade 3, ‘disabling’ nausea) has implications for patient treatment and/or participation in clinical trials. Since prior research using Kappa statistics has demonstrated that clinician CTCAE ratings are not in perfect alignment with patient-reported adverse events, the present study aims to model and further clarify the relationship between these clinician and patient ratings using the Graded Item Response Model (GRM).; METHODS From the medical charts of 393 patients aged 26–91 (M = 62.39; 43 % male) receiving chemotherapy between March 2005 and August 2009, we retrospectively extracted patient and clinician CTCAE ratings of constipation, diarrhea, dyspnea, fatigue, nausea, and vomiting. Twenty-six attending oncologists comprised the clinician rating group, with patient cancer diagnoses including lung (n = 134), prostate (n = 113), and gynecologic (n = 137). Patients reported using previously developed CTCAE plain language adaptations called STAR (Symptom Tracking and Reporting) items. A GRM was fitted to calculate the latent grading thresholds between clinicians and patients.; RESULTS Clinicians have overall higher average thresholds than patients when assessing constipation, diarrhea, dyspnea, nausea and vomiting (range 0.23–0.58 logit units; p \ 0.05 across mild, moderate, and moderately severe symptoms). For fatigue, however, clinicians are more likely to overestimate patient-reported low-grade mild fatigue (lower threshold by -0.58 logit, p \ 0.01) but underestimate grade 2 or higher fatigue (higher threshold by +1.05 logit, p \ 0.0001). Further, the GRM model shows higher variability in patients’ thresholds in assessing CTCAE than those obtained from clinicians.; CONCLUSIONS Clinicians underestimate patient-reported constipation, diarrhea, dyspnea, nausea, vomiting, and high-grade fatigue. Clinicians overestimate low-grade fatigue. This provides evidence to support the notion that patients report some symptoms that clinicians might not consider noteworthy until they are more severe. The observed pattern of reporting for fatigue suggests there may be discordant understanding of this concept that should be further clarified between the doctor and patient.
(3071) Patient-reported satisfaction with psychosocial and sexual well-being and the nipple areola complex: comparison of preservation and reconstruction Alison N. Price, Memorial Sloan Kettering Cancer Center, United States; Amie M. Scott, MPH, Memorial Sloan-Kettering Cancer Center; Anne Klassen, D Phil, McMaster University, Canada; Stefan Cano, BSc PhD, Peninsula College of Medicine & Dentristry, United Kingdom; Babak J. Mehrara, Memorial Sloan Kettering Cancer Center; Joseph J. Disa, Memorial Sloan-Kettering Cancer Center; Colleen McCarthy, MD MS, Memorial Sloan Kettering Cancer Center; Evan Matros, MD, Plastic Surgery Department, Memorial Sloan Kettering Cancer Center, United States; Peter Cordeiro, Memorial Sloan Kettering Cancer Center; Virgilio Sacchini, MD, Memorial Sloan Kettering Cancer Center; Andrea Pusic, MD MHS FRCSC, Memorial Sloan Kettering Cancer Center, United States AIMS Breast cancer patients undergoing mastectomy and immediate expander-implant reconstruction may opt for a nipple sparing mastectomy (NSM) or nipple reconstruction with skin grafts and tattooing. As many patients are candidates for both, valid and reliable patient-centered outcomes data are crucial to shared medical decisionmaking. The objective of this study is to determine whether patient satisfaction with psychosocial and social well-being is impacted by nipple areola complex (NAC) preservation.; METHODS Eligible patients were identified from a prospectively maintained database of patients who completed the BREAST-Q, a patient-reported outcome measure for breast surgery patients, at our institution between January 2011 and December 2011. The dependent variable was the BREASTQ Psychosocial well-being score, a 10-item scale scored from 0 to 100
Qual Life Res (2012) 21:1–132 and the BREAST-Q Sexual well-being score, a 6-item scale scored from 0 to 100. Univariate analysis and clinical judgment were used to identify variables included in the model, and multivariate linear regression models were constructed to control for confounders.; RESULTS The study sample consisted of 254 women, 52 NSM patients and 202 nipple reconstruction patients. The sample was on average aged 45.6 ± 8.2 (range 23–72). All patients underwent an immediate reconstruction with expander-implants. Nipple sparing mastectomy, marital status and body mass index (BMI) were found to be variables that predicted higher patient sexual well-being and psychosocial well-being after controlling for radiation therapy, original cup size, prophylactic mastectomy, age and implant type (marital status p = 0.002; BMI p = 0.037, type of mastectomy p \ 0.001 R-square = 0.14).; CONCLUSIONS As there is a growing population of breast cancer patients seeking NSM, understanding how a woman’s quality of life is impacted by NAC preservation compared with NAC reconstruction is essential. This study suggests that NSM has a positive impact on patient satisfaction with psychosocial and sexual wellbeing in the setting of breast reconstruction. This patientcentered outcome data can be used to enhance shared medical decision-making by providing patients with information about realistic expectations for quality of life related to type of surgery chosen.
(3072) Handling missing item data on a multi-item instrument Iris Eekhout, VU University Medical Center, Netherlands; Henrica C.W. de Vet, Sr., PhD, EMGO Institute VU Medical Center, Netherlands; Michiel R. De Boer, VU University, Amsterdam, The Netherlands; Jos WR Twisk, VU University Medical Centre, Amsterdam, The Netherlands; Martijn W. Heymans, VU University Medical Center AIMS In many missing data situations researchers do not make a distinction in the application of missing data methods when they use a multi-item instrument. Although advanced techniques as multiple imputation are available, complete-case analysis is most frequently applied, regardless of the number of missing item scores. The objective of this study is to explore the performance of simple and more advanced methods on missing data in case some, many or all item scores are missing in a multi-item instrument based on a reflective measurement model.; METHODS Real life missing data situations were simulated in a multi-item instrument according to specific mechanisms as missing items in a completely random subsample of the data (missing completely at random [MCAR]); missing items dependent on other observed items or on variables outside the instrument (missing at random [MAR]); and missing items dependent on the value of the missing itself (missing not at random [MNAR]). Each missing data situation was created 500 times and the percentage of missing item scores was varied. Subsequently, several techniques to handle missing data were applied as mean imputation, regression imputation and multiple imputation, as well as methods designed for multi-item instruments as two-way and multivariate normal imputation to decide on the most optimal technique. Their implications were studied by comparing regression coefficients fitted on the data using the standardized bias and root mean square error.; RESULTS This simulation study is ongoing. Preliminary results show that when a large amount of item scores are missing, coefficients had more bias compared to the situations where a small amount of items scores are missing. A complete picture of the results also considering specific mechanisms and techniques, will be presented at the conference; CONCLUSIONS According to a recently published systematic review (1), we advise researchers who use multi-item instruments to distinguish between missing item scores and missing total scores when they are considering methods to deal with the missing data. Based on the results of this simulation study, we aim to refine this
127 advice depending on the number of missing item scores and optimal methods in these cases. 1. Eekhout, I., De Boer, M. R., Twisk, J. W., De Vet, H. C., & Heymans, M. W. (2012). Missing Data: A Systematic Review of How They Are Reported and Handled. [Brief Report]. Epidemiology, 23(4), 4.
(3073) Attitudes to ageing: do gender and culture make a difference in older age? Gail Low, PhD, Univ of Alberta, Faculty of Nursing, Canada; Anita Molzahn, PhD RN, University of Alberta, Canada; Mary Kalfoss, DrPH, Diakonova University College, Norway AIMS Examining the effects of gender and country of origin on attitudes toward ageing among older adults from Canada (n = 202) and Norway (n = 482) who took part in the 2004 WHOQOL Group Field Trials.; METHODS In a MANOVA, we examined differences in participants’ attitudes toward psychosocial loss and growth, and the physical changes of ageing by two grouping factors: gender and country of origin. And thereafter, for personal (self-positivity, independence and autonomy) and social (social relationships and activities) contextual variables that, in theory, are important to consider in cross-cultural research. Group comparisons of contextual variable effect sizes upon attitudes to ageing were made and modelled using AMOS 20.0.; RESULTS Canadians had more positive attitudes than did Norwegians toward psychosocial loss (F = 256.60, df = 1, p = .000) and physical change (F = 14.23, df = 1, p = .000). Otherwise women in either country were more satisfied with their social relationships (F = 7.99, df = 1, p = .005). Among older Canadians and Norwegians alike, social participation had a larger effect size than self-positivity (CR = 3.55) and a more marked influence on psychosocial loss than independence (CR = 3.75) in the Canadian sample alone. Social relationships had the same effect size as self-positivity in both groups (CR = 1.67) and were as important as independence to older Canadians (CR = 1.29). Independence mattered little to Norwegians’ sentiments about loss (p = .792). In the physical change model, social participation had a significantly smaller effect than self-positivity (CR = 2.53) and independence (CR = 2.64) in either country. There were marked variations in autonomy and social relationships between Canadians and Norwegians; however both variables had negligible effects (p = .258 and p = .758, respectively). Autonomy was also of little significance to psychosocial loss (p = .605 for Canada and Norway alike). These constrained and freely estimated effects yielded goodness-of-fit with the Canadian and Norwegian data for psychosocial loss (Model 92 = 38.63, df = 30, p = .134, AGFI = .97, RMSEA = .021, CFI = .995) and physical change (Model 92 = 35.44, df = 29, p = .190; AGFI = .972, RMSEA = .018, CFI = .997).; CONCLUSIONS Among two groups of older people supposedly viewing older age through an individualistic frame of reference, attitudes to loss were equally or more heavily attributable to social contextual factors. Qualitative investigations about the nature of social networks and their impact on aging, and social traditions around relationships and activities are warranted.
(3074) A longitudinal analysis of satisfaction with care of patients treated in ambulatory oncology: the influence of health related quality of life Thanh Vaˆn Nguyen, Clinical Research Department EA3181, France; Anne Bredart, PhD, Institut Curie, France; Jean-Franc¸ois Bosset, Radiotherapy Department, Besanc¸on University Hospital, 25030 Besanc¸on, France; Alain Monnier, Radiotherapy Department,
123
128 Montbe´liard Hospital, 25200 Montbe´liard, France; Mariette Mercier, Cellular and molecular biology laboratory, University Hospital of Besanc¸on AIMS In the oncology setting, there has been an emphasis in evaluating outcomes of treatments in terms of quality of life and patient satisfaction. The aim of our study was to investigate determinants of patient satisfaction and to assess the relationship between quality of life and satisfaction with care and their changes over time in curative treatment of cancer out-patients.; METHODS Patients undergoing ambulatory chemotherapy or radiotherapy in 2 centers in France were invited, at the beginning of the treatment, at the end and 3 months after, to complete the OUT-PATSAT35, a questionnaire composed of 35 items and 13 scales, evaluating perception of doctors, nurses, as well as aspects of care organization and services. All measures range from 0 to 100, a higher score reflecting a higher level of satisfaction. Additionally, for each patient, data were collected on socio-demographic, clinical characteristics and quality of life (EORTC QLQ-C30).; RESULTS Among 691 patients initially included, 561 answered to the 3 assessment timings. In multivariate analysis, at the end of the treatment, patients who underwent a deterioration of their global health, reported less satisfaction in most scales, and 3 months after, the same patients had lower satisfaction scores only in doctors’ evaluation. The longitudinal analysis evidenced a significant relationship between a deterioration of global health and a decrease of satisfaction in doctors’ domains and, conversely, between an improvement of global health and an increase of satisfaction in the overall satisfaction scale. Furthermore, perceived global health at the beginning of the treatment remained largely and significantly associated with all satisfaction scores in the following timings. Radiotherapy (vs. chemotherapy) was linked to less satisfaction with doctors’ and nurses’ provision of information and waiting-time, and head and neck cancer patients were less satisfied with the hospital environment; CONCLUSIONS Initial self evaluated global health was found to be the major determinant of patient satisfaction in ambulatory oncology. The subsequent deterioration of global health, during and after the treatment, emphasized the decrease of satisfaction scores, mainly in doctors’ evaluation. Initiatives targeting these patients with poorer health status, before starting the treatment, should improve their pattern of care.
(3075) An international prospective study establishing minimal clinically important differences in the EORTC QLQ-BM22 and QLQ-C30 in cancer patients with bone metastases Liang Zeng, Canada; Edward Chow, University of Toronto; Liying Zhang, University of Toronto, Canada; Ling-Ming Tseng, National Yang-Ming University; Ming-Feng Hou, Kaohsiung Medical University Hospital; Alysa Fairchild, BSc MD FRCPC, Cross Cancer Institute, Canada; Vassilios Vassiliou, Bank of Cyprus Oncology Centre; Reynaldo Jesus-Garcia, Federal University of Sao Paulo; Mohamed A. Alm El-Din, Tanta University Hospital; Aswin Kumar, Regional Cancer Center, Trivandrum AIMS Quality of life (QOL) is frequently an endpoint in clinical trials involving patients with advanced cancer. Statistical significance of minimal differences can be achieved with sufficient sample size, yet the actual clinical relevance is unknown. The purpose of this study was to establish the minimal clinically important difference (MCID) for the EORTC bone metastases module (EORTC QLQ-BM22); METHODS Patients with bone metastases across seven countries were prospectively enrolled in a trial validating the EORTC QLQ-BM22 and completed the QLQ-BM22 and core measure (QLQ-C30) at baseline and 1-month follow up. MCIDs were calculated for each QOL scale for both improvement and deterioration using both an anchor-
123
Qual Life Res (2012) 21:1–132 (performance status) and distribution-based approach; RESULTS A total of 93 patients completed both baseline and follow-up QOL and had recorded performance status at both intervals. Statistically significant meaningful differences were seen in 7 scales. Improvements of 30.5, 20.1, 30.5 and 19.6 in the pain, painful site, painful characteristic and functional interference scales, respectively, demonstrated clinical relevance. Decreases of 12.4, 22.4 and 13.4 were required to represent clinically relevant deterioration in emotional functioning, global health status and financial issues, respectively. Minimal differences for improvement were closest to 0.5 standard deviations (SD) while for deterioration, closer to 0.3 SD on the QLQ-BM22.; CONCLUSIONS Identification of requirements for clinical significance can assist in determining the relevance of QOL changes after treatment and in sample size determination in future trials. Future studies should continue to determine MCID and confirm our findings using a variety of appropriate anchors.
(3076) The impact of psychotherapeutic support on quality of life in patients with gastrointestinal cancer Beate Bestmann, PhD, WINEG, Scientific Institute of TK for Benefit and Efficiency in Health Care, Germany; Thomas Kuechler, PhD, Referenzzentrum Lebensqualitat, Germany; Doris Henne-Bruns, Department of Visceral and Transplantation Surgery, University Hospital of Ulm; Frank Verheyen, WINEG, Scientific Institute of TK for Benefit and Efficiency in Health Care; Sharon Wood-Dauphinee, PhD, McGill Univ.-School of Physical Therapy, Canada AIMS In order to analyze the impact of psychotherapeutic support on survival and quality of life in patients undergoing surgery for gastrointestinal cancer, a randomized controlled trial was conducted in Hamburg, Germany, from January 1991 to January 1993. Survival analysis showed a significant benefit for patients receiving psychotherapeutic support (Kuechler et al. 2007). The aim of this current study was to explore the impact on the patient’s quality of life facing a higher proportion of dropouts in the control group.; METHODS A total of 271 patients were randomized in two groups: patients receiving psychosocial support during the hospital stay (N = 136) and control group (N = 135). The sample consisted of patients with a preliminary diagnosis of cancer of the esophagus, stomach, liver/gallbladder, pancreas, or colon/rectum. Quality of life (QoL) was assessed using the EORTC QLQ-C30 plus a disease-specific module. Mapping onto the EQ5D was performed according to McKenzie et al. (2009). The questionnaires were administered at baseline and at discharge as well as at the 3-, 6-, 12- and 24 months follow-up. Survival was followed up to 10 years after treatment. Statistical analysis included descriptive statistics and t-tests for independent samples. A global p-value of p \ 0.05 was considered statistically significant, whereas a mean difference of 10 points and more on the EORTC scales was considered clinically significant/relevant; RESULTS Looking at the EORTC QLQ-C30, the control group showed slightly better QoL at discharge. During the follow-up no statistically and clinically significant differences were found. Due to the fact that patients in the intervention group showed significant better survival, these QoL results might be seriously biased from non-random missing data and therefore overestimated (Moinpur et al. 2000). For that reason we combined QoL and survival data. We mapped the EORTC QLQ-C30 onto the EQ-5D and then calculated quality-adjusted life-years (QALYs). In terms of QALYs, patients undergoing psychotherapeutic support showed significant better outcomes than patients in the control group.; CONCLUSIONS Cancer clinical trials often have to deal with non-random missing data, which have the potential of seriously biasing the results. Therefore, survival data should be taken into account when interpreting QoL scores.
Qual Life Res (2012) 21:1–132 (3077) Two-part joint regression modeling to analyze longitudinal left censored patient reported outcomes Alette Spriensma, VU University Medical Center, Amsterdam, the Netherlands, Netherlands; Jos WR Twisk, VU University Medical Centre, Amsterdam, The Netherlands AIMS Patient reported outcomes that measure health status over time often result into scores that have a lower and an upper limit. These outcomes are mostly analyzed using ordinary least square models such as longitudinal linear regression. The true score of a patient, however, can lie outside of the chosen range, which results into censoring of the outcome variable. In some situations left censoring occurs as an excess of zeros in the outcome variable. Within econometrics, several methods have been created to deal with an excess of zeros. These methods are called two-part statistical models or mixed distribution models. The outcome variable contains two distributions: A binomial distribution to distinguish between left censored observations and uncensored observations, and a normal distribution (or other continuous distribution) that deals with the uncensored part of the outcome variable. These two-part models result into two separate regression coefficients. Despite this, the scores of patient reported outcomes should be seen as one process and one regression coefficient would be preferred. The purpose of this study is to compare the results of standard approaches with the results obtained from a two-part joint modeling approach in analyzing longitudinal patient reported outcomes.; METHODS A linear mixed model and a logarithmic transformation of the linear mixed model (ln(x + 1)) approach were compared with a two-part joint modeling approach. In order to compare the models, model fit indicators and differences between observed and predicted values were used. All analyses were performed within STATA using the GLLAMM procedure. A health assessment questionnaire example dataset was used to compare the different methods.; RESULTS Preliminary results showed that a twopart joint approach performs better in analyzing the outcome variables in comparison with the standard used approaches. The model fit indicators show a better score for the two-part joint approach, furthermore the scatter plots indicate that the approach is better equipped in estimating the excess of zeros.; CONCLUSIONS Even though there is difficulty in interpreting the regression coefficients of a two-part joint model, this approach has shown to be a flexible tool in analyzing longitudinal patient reported outcome variables that are left censored.
(3078) Evaluating the response category translations of the EORTC QLQ-C30 questionnaire Neil W. Scott, MA MSc PhD, University of Aberdeen, United Kingdom; Josephine A. Etta, University of Aberdeen, United Kingdom; Neil K. Aaronson, PhD, The Netherlands Cancer Institute, Netherlands; Andrew Bottomley, PhD, EORTC Headquarters, Belgium; Peter Fayers, PhD, Aberdeen Medical School, United Kingdom; Mogens Groenvold, MD, PhD, University of Copenhagen, Denmark; Michael Koller, PhD, Center for Clinical Studies, Germany; Dagmara Kulis, EORTC, Belgium; Debbi Marais, University of Aberdeen; Morten Aa Petersen, MSCi, Bispedjerg Hospital, Dept of Palliative Medicine, Denmark; Mirjam AG Sprangers, University of Amsterdam, Netherlands AIMS The aim of this study was to evaluate the translated response categories used in eight versions of the European Organisation for Research and Treatment of Cancer (EORTC) QLQ-C30 questionnaire, which is used in many international clinical trials. Twenty-eight of the 30 items in the questionnaire have the same four response categories: ‘‘Not at all’’, ‘‘A little’’, ‘‘Quite a bit’’ and ‘‘Very Much’’; METHODS People with good knowledge of both English and one of
129 eight European languages were asked to participate in an Internet survey. The relative strength (intensity) of the English and translated categories were assessed using two separate approaches: one using a verbal response scale (VRS), the other a visual analogue scale (VAS).; RESULTS Two hundred and seventy-nine people participated in the survey. Most translations were rated similarly to English. The largest differences were for the German translation of ‘‘Quite a bit’’, which was rated 16.3 points lower than the corresponding English category on a 0–100 VAS. Two Danish translations were associated with VAS differences of over five points. CONCLUSIONS Most of the translated categories were found to be similar to the English versions and should continue to be used. We recommend that three translated categories should be considered for revision. Similar surveys could be used to assess the categories used in other translated quality of life instruments.
(3079) Physical activity matters: the associations between body mass index, physical activity and health-related quality of life trajectories over 10 Years David Feeny, PhD, University of Alberta, United States; Rochelle Garner, Statistics Canada; Julie Bernier, MSc, Statistics Canada, Canada; Amanda Thompson, Statistics Canada; Bentson H. McFarland, Oregon Health & Science University; Nathalie Huguet, Portland State University; Mark S. Kaplan, Portland State University; Nancy A. Ross, McGill University; Chris Blanchard, Dalhousie University AIMS The objective of this study was to assess the associations between body mass index (BMI), leisure time physical activity (LTPA) and health-related quality of life (HRQL) trajectories among adults.; METHODS Data were drawn from the Canadian National Population Health Survey, with respondents being interviewed every 2 years between 1996/97 and 2006/07. Using growth curve modeling. HRQL trajectories for individuals aged 18 and over were associated with measures of BMI and LTPA. Growth models were conducted separately for males and females.; RESULTS Findings suggested that, for males, BMI categories had little impact on baseline HRQL, and no impact on the rate of change in HRQL as men aged. Among women, higher BMI categories were associated with significantly lower baseline HRQL. However, BMI had no impact on the rate of change of HRQL. Conversely, LTPA had significant impacts on baseline HRQL, as well as the rate of change in HRQL, with individuals who were inactive or sedentary having much steeper declines in HRQL as they aged, as compared to individuals who were active in their leisure time. This was true for both men and women, regardless of BMI category.; CONCLUSIONS The results have implications both for public health policy and clinical practice. One implication is the importance of the built environment, pedestrian and cycling friendly environments, and public transit to provide opportunities for engaging in physical activity both for recreation and commuting. Another is the inclusion of physical education in kindergarten through high school programs, perhaps with an emphasis on physical education programs that teach activities and skills that promote life-long participation and activity, rather than achieving competitive excellence. Further, perhaps there needs to be an emphasis on community recreation programs and helping families to promote physically active lifestyles. For clinical practice, it may be important for clinicians to ask how physically active their patients are and to provide counseling to equip patients with the tools needed to engage in a sustained program of regular physical activity. Further, counseling by physical activity counselors can enhance the effectiveness of strategies to promote behavior change. For both men and women, it seems more important that individuals remain active in their leisure time as they age, and perhaps less important that they lose weight.
123
130 (3081) Self-report quality of life measure is reliable and valid in patients suffering from schizophrenia with executive impairment Laurent Boyer, PhD, Aix-Marseille University, France; Karine Baumstarck, France; Romain Padovani, EA 3279; Julie Berbis, Department of Public Health, Medicine University, France; Julien Blanc, Aix-Marseille University, France; Christophe Lancon, MD, Laboratoire de Sante´ Publique, Faculte´ de Me´decine, France; Pascal Auquier, PhD, Laboratoire de Sante Publique, France AIMS Impaired executive functions are among the most widely observed in patients suffering from schizophrenia. The use of selfreported outcomes for evaluating treatment and managing care of these patients has been questioned. The aim of this study was to provide new evidence about the suitability of self-reported outcome for use in this specific population by exploring the internal structure, reliability and external validity of a specific quality of life (QoL) instrument, the Schizophrenia Quality of Life questionnaire (SQoL18).; METHODS Design: cross-sectional study. Inclusion criteria: diagnosis of schizophrenia according to the DSM-IV criteria. Data collection: sociodemographic (age, gender, and education level) and clinical data (duration of illness, Positive and Negative Syndrome Scale, Calgary Depression Scale for Schizophrenia); QoL (SQoL18); and executive performance (Stroop test, lexical and verbal fluency, and trail making test). Non-impaired and impaired populations were defined for each test. For the 6 groups, psychometric properties were compared to those reported from the reference population assessed in the validation study.; RESULTS One hundred and thirteen consecutive patients were enrolled. The confirmatory factor analysis performed in the impaired groups showed that the questionnaire structure adequately matched the initial structure of the SQoL18. The unidimensionality of the dimensions was preserved, and the internal/external validity indices were close to those of the non-impaired groups and the reference population.; CONCLUSIONS Our study suggests that executive dysfunction did not compromise the reliability or validity of self-reported QoL questionnaire.
(3083) Validating a measurement model for the CASP-19 quality of life scale in an older Irish cohort Eithne Sexton, Royal College of Surgeons in Ireland, Ireland; Bellinda L. King-Kallimanis, PhD, TILDA, Trinity College Dublin, Ireland; Anne Hickey, Royal College of Surgeons in Ireland AIMS This study aims to evaluate the validity of existing measurement models for the CASP (Control, Autonomy, Self-Realisation, Pleasure) measure of quality of life—the original CASP-19 model and a modified CASP-12 version—in a recent population survey of older people. Previous studies have reported poor fit for these measurement models. The study also aims to re-develop the model and propose a well-fitting alternative.; METHODS To evaluate the existing measurement models, confirmatory factor analysis (CFA) was used. Data came from a crosssectional sample (n = 6, 823) representative of the Irish community dwelling population aged 50 and over, obtained from the Irish Longitudinal Study of Ageing (TILDA). Model revision was based on descriptive statistics, exploratory factor analysis (EFA) and examination of fit diagnostic statistics. Revised models were tested using CFA.; RESULTS The results of the CFA did not support the validity of the four factor 19 item (CASP-19) and three factor 12 item (CASP-12) models in the TILDA sample. A revised two factor model used 12 of the original 19 items, and combined the control and autonomy dimensions, and the self-realisation and pleasure dimensions. Correlated errors were added to negatively worded items to adjust for an apparent method effect. This model had excellent overall fit to the data (chi-sq 161.90, df = 44, p \ .001; RMSEA = 0.03, 90 % CI 0.02–0.03). A one factor model, with the same 12 items and correlated errors for negatively worded
123
Qual Life Res (2012) 21:1–132 items, also had excellent overall fit (chi-sq 150.86, df = 43, p \ 0.001; RMSEA = 0.03, 90 % CI 0.02–0.03).; CONCLUSIONS Both the redeveloped one and two factor models showed satisfactory fit to the data. We recommend use of the one factor model when examining overall quality of life among older people, while control/autonomy and selfrealisation/pleasure can be examined separately by researchers interested in those constructs. We recommend that researchers adjust for the method effect when using structural equation modelling approaches. Further validation of the proposed models in other samples is required.
(3084) What characterizes middle aged and elderly individuals who do not answer questions on sexuality? Results from a quality of life study Gudrun Rohde, PhD, University of Agder, Norway; Kari Hansen Berg, Sorlandet Hospital, Kristiansand, Department of Rheumatology; Glenn Haugeberg, Sorlandet Hospital, Kristiansand, Department of Rheumatology AIMS Intimacy and sexuality are considered as important parts of Quality of Life (QOL). With increasing age health status may decline and thus have a negative impact on sexual activity. There is a lack of data illuminating the impact of health on sexuality in middle aged and elderly individuals. Aim: To explore the characteristics of middle aged and elderly individuals participating in a in a quality of life study and not answering questions on sexuality.; METHODS Three hundred and ninety seven individuals (353 women, 54 men) aged = 50 years from a case control QOL life study of distal radius fracture patients were studied (181 patients and 226 controls). Three QOL questionnaires were applied: Short Form-36, Quality of Life Scale and 15D. The 15D questionnaire includes one item regarding the impact of health status on sexual activity. Only minor differences was observed between patients and controls, thus all were pooled into one group, and categorized as responders (n = 308) and non-responders (n = 101). Chi-square, independent samples t tests and multiple logistic regression analyses were applied.; RESULTS Mean age for of participants was 67 (8) years, and 101 participants (25 %) did not respond to the 15D questionnaire on sexuality. Non-respondents were significantly older (p \ 0.001), more were females (p \ 0.001), had lower weight (p = 0.001), had less education (p = 0.001), lived alone (p = 0.001), exercised less (p = 0.022), suffered from more diseases (p \ 0.001), had more osteoporosis (p \ 0.001), reported lower MHAQ scores (p = 0.015), lower SF-36-physical component summary scores (p = 0.001) and lower QOLS scores (p = 0.042). In the multiple regression analyses younger age (70–79 years; OR = 0.40, 95 % CI 0.40–0.89, 80 years and older; OR = 0.17, 95 % CI = 0.05–0.54), male gender (OR = 10.3, 95 % CI = 1.32– 79.8) and being married\cohabiting (OR = 3.46, 95 % CI = 1.87– 7.39) were independently associated with being a respondent vs nonrespondent to the 15D question on sexuality.; CONCLUSIONS The results from our study reveals that responders to questions on sexuality in QOL studies in middle aged and elderly individuals are younger, are living in relationship as married or as cohabitants and are more likely to be men. Further studies are needed to illuminate the role of sexuality for QOL in middle aged and elderly individuals.
(3085) Quality of life and psychological factors in manufacturing workers in Mexico Jose Guadalupe Salazar Estrada, Sr., Psic, PhD, Uiniversidad de Guadalajara, Mexico; Norma Alicia Ruvalcaba Romero, MA, Mexico; Norma S. Figueroa Villasen˜or, Universidad de Guadalajara AIMS The aim is to identify the relationship between mental health status and the Sense of Coherence (SC) and psychosocial and quality
Qual Life Res (2012) 21:1–132 of life in relation to labor perceptual characteristics of manufacturing workers in Jalisco, Mexico; METHODS Through a cross-sectional study with descriptive and the application of various questionnaires that identify sense of coherence, quality of life and psychosocial factors in a random sample of workers in the manufacturing industry with more than 3 years old in the same factory; RESULTS The highest overall score it obtained SC men with 124.44. The presence of a ‘‘poor mental health’’ is presented in 17.61 % of men and 29.45 % of women, being male where statistically significant differences with respect to the higher average age of the worker and the time working. the decreased quality of life is related to the working time; CONCLUSIONS SC scores were higher in subjects with better mental health in both sexes, and is inversely correlated with scores on the General Health Questionnaire (GHQ) in all subjects studied and the quality of life of workers .
(3086) Patient-reported outcome measures in studies of colorectal cancer surgery: a systematic review Robert N. Whistance, MB ChB BSc MRCS, University of Bristol, United Kingdom; Rachael O. Forsythe, University Hospitals Bristol NHS Foundation Trust; Angus McNair, MRCS, University of Bristol, United Kingdom; Sara Brookes, University of Bristol; Kerry Avery, PhD, Dept of Social Med, Univ of Bristol, United Kingdom; Anne M. Pullyblank, North Bristol NHS Trust; Jane Blazeby, MD, University of Bristol, United Kingdom AIMS Advances in the treatment of colorectal cancer (CRC) require evaluation with outcomes provided by patients themselves (patientreported outcomes; PROs). Inconsistent use of PRO questionnaires is a major barrier to comparisons between trials and data syntheses in meta-analyses. This study aimed to systematically identify PRO questionnaires reported in studies of CRC surgery, extract the individual questions (items) from them and then group items pertaining to similar constructs into domains. This review will inform the development of a core outcome set (COS) for CRC surgery; a set of endpoints reported as a minimum in all trials of CRC surgery.; METHODS Medical databases were searched using terms for ‘colorectal cancer’, ‘surgery’ and ‘PROs’. Included were original research articles (from 2009 to 2010) and systematic reviews (from 2001 to 2010) of PROs in patients undergoing CRC surgery. Excluded were studies of palliative therapy, screening, non-biomedical interventions and treatment of CRC metastases. PRO questionnaires were obtained by searching the Internet or by contacting authors. All items were extracted and compiled into a longlist. Two researchers blinded to the source questionnaire categorised the items into domains, which was independently checked by two CRC patients and a colorectal surgeon.; RESULTS A total of 5,644 abstracts were screened and, of these, 31 original research articles and 6 systematic reviews were included. Fifty-eight different PRO questionnaires were reported on 184 occasions in 103 publications. The EORTC QLQ-C30 (n = 50) and the EORTC QLQ-CR38 (n = 33) were the most frequently reported questionnaires, while another 40 questionnaires (69.0 %) were reported once. Fifty (86.2 %) full questionnaires were available comprising 920 individual items. The items were grouped into 51 domains. The generic domains comprising most items were ‘Anxiety’ (n = 85) and ‘Fatigue’ (n = 67), while the disease-specific domains with most items were ‘Faecal incontinence’ and ‘Stoma problems’.; CONCLUSIONS There is evidence for inconsistency in the choice of PRO questionnaires in studies of CRC surgery. Standardising the choice of domains measured would facilitate comparisons between different trials and improve data assimilation in systematic reviews. Future work will survey key stakeholders’ views on the most important PRO domains to include in the core outcome set for CRC surgery.
131 (3087) Endorsing high quality standards in developing measures to assess QoL in cancer patients: the data repository initiative on behalf of the EORTC QLG Francesca Martinelli, MSc, EORTC, Belgium; Galina Velikova, MD, University of Leeds, United Kingdom; Colin Johnson, Mchir, Dept of Surgery, United Kingdom; Mogens Groenvold, MD, PhD, University of Copenhagen, Denmark; Corneel Coens, MSc, EORTC HQ, Belgium; Andrew Bottomley, PhD, EORTC Headquarters, Belgium AIMS To meet cancer patients’ and researchers’ needs, one of the aims of the European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Group (QLG) is to develop reliable instruments (modules) for measuring the quality of life (QoL) of cancer patients participating in international clinical trials. The module development process involves four phases, the final one being a large-scale test or field test. The aims of the Data Repository initiative are to establish guidelines in order to harmonize the way field studies are conducted and the way data for field studies are collected and to create a central dataset for storage of all field study data.; METHODS The Data Repository initiative was started in May 2011 and is currently ongoing. Protocols and data from previous field studies are being collected and confidentially stored on the EORTC server, in a secure folder with restricted access. Protocols are being reviewed in order to identify and develop accurate and complete standards for conducting field studies. Data are being harmonized in order to create a central dataset.; RESULTS As a first step of this initiative, a Data Preservation and Sharing Policy has been prepared and finalized. All EORTC QLG Principal Investigators have been very supportive: ten out of thirteen expected protocols have already been collected and are currently being used to prepare a Standard Protocol Template for future studies; almost seventy per cent of expected data have also been collected and are currently being harmonized in order to be added to the central dataset.; CONCLUSIONS As a result of this initiative, the field study process will be harmonized and simplified and data quality will be ensured. Harmonized data will be an asset in terms of quality control. The maintenance of the central dataset will prevent data loss. The long-term added value of data will be enhanced by giving researchers the opportunity to carry out new research on existing data, in accordance with the rules defined in the Data Sharing and Preservation Policy.
(3088) Health-related quality of life in patients with alopecia areata—a report from the national alopecia areata registry Qiuling Shi, PhD, Uiniversity of Texas M.D. Anderson Cancer Center; Joyce S. Osei, UT MD Anderson Cancer Center; Tito R. Mendoza, PhD, Univ. of Texas M.D. Anderson Cancer Ctr.; Madeleine Duvic, UT MD Anderson Cancer Center AIMS Alopecia areata (AA) may significantly affect patient healthrelated quality of life (HRQOL). A better understanding of AA-related HRQOL should lead to improved care for this patient population.; METHODS We conducted a secondary analysis of data from the National Alopecia Areata Registry. AA patients’ HRQOL were measured using three validated instruments: Skindex 16, brief version of the Fear of Negative Evaluation Scale (BFNE), and Dermatology Life Quality Index (DLQI). Mean HRQOL scores were calculated and compared with previously reported HRQOL levels from healthy controls and patients with other skin diseases. A two-step clustering method was used to divide patients into good-HRQOL or poor-HRQOL groups on the basis of their scores. Logistic regression models were used to identify risk factors for poor HRQOL from the pool of demographic and clinical characteristics.; RESULTS Data from 842 patients with AA were analyzed. The mean BFNE score was 35.5 ± 9.9, and the mean
123
132 DLQI score was 6.8 ± 4.8. The three Skindex 16 scores were: symptom scale 18.8 ± 24.2, emotional scale 43.3 ± 33.3, and function scale 30.2 ± 30.8. Scores from the three instruments were significantly correlated, but most correlation coefficients were low (0.1–0.4). BFNE and DLQI were more correlated with the emotion and function subscales of Skindex 16, but less correlated with the symptom scale. Twostep cluster analysis of Skindex 16 scales divided patients into two groups: 481 (57.2 %) with good HRQOL (symptom: 9.1 ± 13.0, emotional: 18.9 ± 18.5, function: 10.0 ± 13.1); 361 (42.8 %) with poor HRQOL (symptom: 31.8 ± 29.2, emotional: 75.7 ± 17.2, function: 56.9 ± 26.9). Multivariate logistic regression modeling revealed a set of risk factors for poor HRQOL: age younger than 50 (OR 3.99, 95 % CI 1.66–9.58), female gender (OR 2.74, 95 % CI 1.73–4.34), hair loss 25–99 % (OR 2.47, 95 % CI 1.12–5.45), family stress (OR 1.8, 95 % CI 1.13–2.86), and job change (OR 2.01, 95 % CI 1.02–3.94).; CONCLUSIONS In this nation-wide study of AA patients’ HRQOL, we found that AA patients’ HRQOL was impaired compared with that reported by other populations. More than 40 % of AA patients reported relatively poor HRQOL. Risk factors for poor HRQOL included younger age, female gender, more hair loss, and distress. The current analysis provides an overview of the HRQOL status of AA patients and may guide patient care in the future.
(3089) Qualitative development and content validation of a new patient reported outcome measure in hyperhidrosis, the Hyperhidrosis Quality of Life Index (Hidroqol) Paul Kamudoni, WSP Centre for Socioeconomic Research, Cardiff University, Germany; Berno Mu¨ller, Riemser Arzneimittel AG, Medical Science and Operations, Germany; Christian Mu¨ller, 2Riemser Arzneimittel AG, Medical Science and Operations, Germany; Sam Salek, BSc RPh PhD FFPM, Cardiff University, Redwood Bldg, United Kingdom AIMS The aims of this study were to develop and validate a diseasespecific instrument for assessing health related quality of life (HRQoL) in hyperhidrosis. Also, to investigate the impact of hyperhidrosis on patients’ HRQoL using a mixed methods approach.; METHODS Online focus group discussion (n = 9), semi-structured interviews (n = 32) and an online surveys (n = 30) were applied for the collection of qualitative data. The sample, comprising of patients with hyperhidrosis, was drawn from online social networking sites (SNS) related to hyperhidrosis. The patients gave consent prior to data collection. Content analysis of the data transcripts was carried out using NVIVO 8 software for qualitative analysis. Expert opinion was sought through panel discussions with dermatologist (n = 5) and patients (n = 7) to evaluate the content validity of the draft instrument.; RESULTS Sixteen online SNS related to hyperhidrosis were used for patient recruitment. The mean age of participants (males = 21, female = 50) was 35 years (SD = 13.2). Mean duration of disease was 23 years (range 3–60 years). A total of 104 HRQoL issues were identified, covering: daily life (mentioned by 96 % of patients), psychological functioning (92 %), social functioning (90 %), professional life (75 %), dealing with the condition (75 %), unmet health care needs (65 %) and physical impact (54 %). For instance, participants described effects on their lifestyle e.g. choice of hobbies. Also commonly reported were negative emotions e.g. embarrassment, and; concerns about whether sweating was noticeable. Ratings provided by expert panel supported the content validity of the new instrument, the HidroQol, with suggestions made on a few items including the recall period for response; CONCLUSIONS This study has identified major issues influencing HRQoL of patients with hyperhidrosis, providing a rich source for the development of a new hyperhidrosis HRQoL
123
Qual Life Res (2012) 21:1–132 instrument. The relevance of its content was supported by a panel of patients and dermatologists. Online SNS provided an adequate sample for the study, pointing towards their use in sourcing patients with skin diseases for outcomes research. Psychometric properties of the Hidroqol will be examined using classical test theory (CTT) and ‘item response theory (IRT)’ approaches.
(3090) Evaluation of severity and frequency levels that optimally differentiate between levels of interference for symptoms in the National Cancer Institute’s Patient-Reported Outcomes Version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) Tito R. Mendoza, PhD; University of Texas M. D. Anderson Cancer Center, Houston, TX, USA; Amylou C. Dueck, PhD; Mayo Clinic, Scottsdale, AZ, USA; Lauren J. Rogak; Memorial Sloan-Kettering Cancer Center, New York, NY, USA; Laura Sit; George Mason University, Fairfax, VA, USA; Yuelin Li, PhD; Memorial SloanKettering Cancer Center, New York, NY, USA; Antonia Bennett, PhD; Memorial Sloan-Kettering Cancer Center, New York, NY, USA; Thomas Atkinson, PhD; Memorial Sloan-Kettering Cancer Center, New York, NY, USA; Steven B. Clauser, PhD; Outcomes Research Branch, NCI, Bethesda, MD, USA; Sandra A. Mitchell, PhD, CRNP; Outcomes Research Branch, NCI, Bethesda, MD, USA; Ethan Basch, MD; Memorial Sloan-Kettering Cancer Center, New York, NY, USA AIMS The National Cancer Institute (NCI) Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) is a library consisting of self-report items representing 78 toxicity symptoms. For each symptom, up to 3 items are included to capture the attributes of severity, frequency, and/or interference. Nine items in the library include all three attributes. To explore the extent to which including multiple attributes for a given symptom is informative, we evaluated what severity and frequency levels optimally differentiate between levels of interference for these nine PRO-CTCAE symptoms. METHODS These data were derived from an NCI-sponsored validation study for the PRO-CTCAE, which included 869 patients (56 % female; median age 59; 32 % minority; 34 % high school or less) with varied cancer types (27 % breast, 20 % lung), and were receiving systemic treatment at 4 NCI-designated cancer centers or 5 NCI Community Cancer Centers Program sites(HHSN261201000043C). To determine optimal cut points for frequency and severity relative to interference, we performed ANOVAs with interference as the dependent variable and various severity or frequency cutpoint boundaries as the independent variable. The cutpoint categorization of frequency and severity which maximizes the difference in interference is determined by inspection of the F-ratio. RESULTS For severity, 67 % (6/9) of symptoms had an optimal cutpoint between ‘mild’ and ‘moderate’ (score 1 and 2 on a 0–4 scale), whereas 33 % (3/9) were between ‘moderate’ and ‘severe’ (score 2 and 3). For frequency, 56 % (5/9) were between ‘rarely’ and ‘occasionally’ (score 1 and 2 on a 0–4 scale), whereas 44 % (4/9) were between ‘occasionally’ and ‘frequently’ (score 2 and 3). Maximal differences in symptom interference were associated with small differences in symptom severity (‘mild’ to ‘moderate’) (e.g., F = 323.1, p \ 0.001, for anxiety). By contrast, maximal differences in interference were associated with moderate differences in symptom frequencies (‘occasionally’ to ‘frequently’) (e.g., F = 281.6, p \ 0.001 for pain). CONCLUSION Cut point variability relative to interference, particularly if these cut points are also found to vary between individuals, suggests that including multiple attributes may be important for understanding the symptom experience. This question is being further evaluated via additional methods.
