Eur J Health Econom 2006 · [Suppl 1] 7:7–104 DOI 10.1007/s10198-006-0368-8 © Springer Medizin Verlag 2005

Oral presentations 0161 Cost-effective of screening for colorectal cancer Aas E. (University of Oslo, Norway) In many European countries screening for colorectal cancer is introduced without a proper evidence of cost-effectiveness. In this paper we present a cost-effectiveness analysis of screening for colorectal cancer with two unique features: A comprehensive dataset is collected from a Nor wegian randomized trial of screening with sigmoidoscopy, and uncertainty is explicitly modelled as a Partially Observed Markov Decision Process. Four questions are examined: What is the impact on cost of screening of including avoided treatment costs and production gains? What is the cost of screening per life year gained? What is the marginal cost per life year gained of including a faecal occult blood test (FOBT)? What is the cost per life year gained according to Dukes’ stages? Data at the level of each individual include information about trial costs, use of health care services, labour market participation, travel expenses, incidences of cancer with stage description, time of death and socioeconomic variables as income, education, age, and gender. The robustness of the results is tested with probabilistic sensitivity analyses. Preliminary results show that the cost per life year gained from sigmoidoscopy is favourable compared with other preventive health measures, while an additional FOBT is not.

0355 Challenges of Estonian health system due to mobility of health professionals Aaviksoo A. (PRAXIS Centre for Policy Studies Policy Programme, Estonia, Tallinn) Votilderk A., Kallaste E., Jesse M. Current paper describes challenges and policy responses that mobility of health professionals has posed on Estonia, a new EU member. Our analysis is based on several surveys of health professionals and our own studies conducted between 2003 and 2005.years have witnessed increased emigration of Estonian doctors, residents and nurses to Scandinavian countries, which is mainly driven by very large wage differentials and expectations for better working conditions abroad. An active recruitment by recipient countries has played a very important role in facilitating emigration.of Estonian health professionals is already having a significant impact on Estonian health system, causing shortage of nurses and doctors in some fields.policy reaction of Estonian government to these problems has been to increase salaries of health professionals by 30-40% in 2005. On the other side of the policy measures are longer waiting lists for health services, and higher government funding for education of health professionals. In the paper we discuss whether these policy measures have had substantial impact on the challenges of Estonian health system and what might be the future options.

0061 Exploring State Dependent Reporting Bias in Self-Assessed Health in a Sub-Saharan Population Abegunde D. (World Health Organization, Geneva, Switzerland,) Lambo E.S., Akande T.M. A number of studies, mainly from developed countries, have illustrated the challenges in cross-population comparability of responses to Self Assessed Health (SAH) questions across individuals who differ in terms of socio-economic, demographic, cultural factors and health experience. This is because of response heterogeneity, which results from response category shift or “true” health effects. A challenge is isolating the “true” health effects from reporting behaviour, particularly in a multicultural population with developing country demographics. Using household health survey data from Nigeria, a typical Sub-Saharan country, we explored state dependent reporting bias across individuals’; cultural backgrounds, household wealth, level of education, age and gender, and compared the results with similar studies from developed countries. By conditioning the SAH response on observable health variable (morbidity experience) and other individual characteristics, our results showed that reporting heterogeneity was more prevalent in our data compared to studies from developed countries. We found evidence for index and threshold shift for cultural groups, household wealth, education, age and gender. Distinctively, reporting heterogeneity was demonstrated even across household wealth and educational level in our data unlike in studies from developed countries. We conclude that individual characteristics influence reporting behaviour differently in our data. The findings have implications for sample size in future studies. To compare SAH responses across subgroups, response behaviour should first be controlled for and results of studies using SAH should be cautiously interpreted for other countries, particularly in sub-Saharan Africa.

0579 Testing the Predictive Validity of TTO and SG Abellan J.M. (Department of Economics, Erasmus University, Rotterdam, The Netherlands) Bleichrodt H. A well-known problem in health utility measurement is that the main methods yield systematically different results. Bleichrodt & Johannesson (1997) found evidence that among the main methods, the time trade-off is most consistent with individual preferences. Their method used only ranking data, however, and was entirely framed in terms of certainty. Besides, in recent year advances in the SG have been proposed that have improved the descriptive accuracy of SG measurements. In particular, allowing that people behave according to prospect theory, currently the main descriptive theory of decision under risk, has removed many inconsistencies in SG measurement. This paper presents a new comparison between TTO and SG measurements both under certainty and under risk and both based on ranking and based on choice data. We find that when all health outcomes are certain the TTO still outperforms the SG. For risky health outcomes, however, the SG performs better than the TTO, especially when we correct the SG for the deviations from expected utility that are modelled by prospect theory. This pattern holds both for ranking and for choice data.

0549 The Value of Human Life in Hungary Adorján R. (Ministry of Finance, Budapest, Hungary) The main objective of my work was to determine the value of statistical human life in Hungary. For this evaluation I used a consensual method of the so called labour market approach. I carried out ten different regression model specifications on a sample of 1287 elements that repEur J Health Econom Suppl 1 · 2006

| S7

resented the full-time employed adult population. Different forms of income as the dependent variable and various bundles of explanatory variables were used. Income was significantly influenced in a positive direction by education level, gender, whether white collar worker or not, residence (in Budapest or countryside towns), age, the size of the company, independence, and number of children. The nationality of the company had a negative effect. The best models produced the range HUF 78-393 million (in year 1998) for the most probable inter val of the value of a statistical human life, with HUF 250 million being the most probable value. This result can elementary contribute to public policy decisions about a state investment or action, where risk of loosing or gaining human life is existing and comparison between the costs and the benefits is needed. However, in this aspect Hungary has lot to do in the near future.

0261 Health expenditure under demographic uncertainty Ahn N. (FEDEA, Madrid, Spain,) Previous projections of health expenditure have been usually carried out by combining constant age-sex specific health expenditure with deterministic population projection. However, it has been shown that population size and structure in the future is highly uncertain and most projections turned out wrong. We improve upon the existing literature by incorporating uncertainties in population projection. According to the median projection, national health expenditure will increase by 2 to 5 percentage points as a share of GDP during the period 2004-2050. However, demographic uncertainty will entail substantial uncertainties in the future health expenditure. By 80% confidence interval, health expenditure in 2050 as the share of GDP will range from 9.5 to 12% in Belgium, from 9 to 15% in Finland, from 11.2 to 13.1% in Germany, from 11.2 to 12.5% in The Netherlands, and from 6.4 to 7% in Spain. This projection assumes constant real age-specific per-capita health expenditure. On the other hand, some projections carried out under even a slightly higher growth rate of per-capita health expenditure relative to that of per-capita GDP would increase substantially the share of health expenditure in GDP in the medium and long term.

0375 Adverse Selection and optimal health insurance contracts Alary D. (Université Paris Dauphine, Paris, France) Bien F. This paper investigates the impact of adverse selection on health insurance contracts. We use a bi-dimensional utility function (wealth and health status) to take into account the link between health status loss and the indemnity. Health status loss can increase or decrease marginal utility of wealth and then modify optimal contracts. We show that when overinsurance is not forbidden by institutional constraint, first best indemnities are larger than pecuniary loss if the marginal utility of wealth is decreasing in health otherwise the indemnity is lower than the pecuniary losses. The first part of this result provides an explanation for disability pension in health insurance contracts. The intuition is that illness reduce health status and then increases (decreases) the marginal utility of wealth in the bad state of nature when marginal utility is decreasing (increasing) in health. Next, we give sufficient conditions such that first best contracts are optimal with asymmetric information. Indeed high risks prefer higher (resp. lower) indemnity than low risks when marginal utility of wealth is decreasing in health (resp. increasing).

S8 |

Eur J Health Econom Suppl 1 · 2006

0083 Determining the attributes of informal care Al-Janabi H. (University of Bristol, Bristol, United Kingdom) Coast J., Flynn T. Health economists could make more use of the qualitative research that already exists in healthcare to aid experimental design for stated preference discrete choice modelling. Meta-ethnography is a technique to synthesise qualitative research, and bears similarities to quantitative metaanalysis. Meta-ethnography was applied to create attributes for informal care. First, a literature search was conducted for qualitative research on informal care. Second, the studies identified were purposively reduced to a manageable list for synthesis. Third, the findings of the studies were analysed, contrasted and finally combined. Through this process themes were developed to explain the experience of informal care. From these overarching themes a list of informal care attributes will be presented. These attributes will cover reciprocity, burden, respite, activity and control. Meta-ethnography has several advantages over other qualitative methods of creating attributes. It can be used to develop an understanding of a topic beyond that of a standard literature review, but it can also be quicker and more comprehensive than conducting primary research to obtain attributes.

0167 Income-related inequality in physician service use among older people in 10 European countries Allin S. (London School of Economics Health and Social Care, London, UK) Masseria C. Using a unique, comprehensive dataset across 10 European countries, Sur vey on Health Ageing and Retirement in Europe (SHARE), this study measures the level of income-related inequality in physician (GP and specialist) service use among those over age 60. The probability of a visit to the GP and specialist (in the past 12 months) was predicted using multivariate logistic regression models. Income-related horizontal inequity indices were calculated according to the indirect standardization approach. Overall, in all ten countries investigated, the probability of GP visit appears largely equitable while specialist ser vice use appears pro-rich. Only in the Netherlands and Sweden is there evidence of a pro-rich inequity in the probability of a GP visit. After adjusting for differences in need across the income distribution, the distribution of specialist care favours the high income groups in eight out of ten countries. The importance of financial barriers to access (e.g. user charges) and more comprehensive measurements of wealth are discussed.

0213 Estimating smoking cessation relapse rates for policy evaluation Amaya-Amaya M. (Health economics research unit of applied health sciences university of Aberdeen, UK) Goode A., Ludbrook A. Within the UK government’s Tobacco Control policy, smoking cessation interventions have been demonstrated to be effective and cost-effective at the individual level. However, for the purposes of policy evaluation, what is required is an estimate of the expected effectiveness, at a population level, of different components of the strategy. The impact of smoking cessation services on prevalence depends not only on the levels of activity and short term quit rates but also on understanding how many of these quitters will return to smoking. Estimates produced for a review of the effectiveness and cost-effectiveness of smoking cessation interventions are based on a lifetime relapse rate of 40%, with a range of 30% 50%. This paper explores the use of the British Household Panel Survey

(BHPS) data to investigate the determinants of smoking cessation pathways and trajectories. The duration up to a relapse after quitting is represented with several models for the hazard function, including parametric and semi-parametric specifications. Non-parametric analysis is also carried out. Statistical tests are used to choose the best specification. Ultimately this research should produce more robust estimates for the UK of the long-term relapse rates for those who quit smoking

0165 Cognitive Ability and Scale Bias in the Contingent Valuation Method Andersson H. (Swedish National Road & Transport Research Institute, Solna, Sweden) Svensson M. In this study we investigate whether or not the scale bias found in contingent valuation (CVM) studies on mortality risk reductions is a result of cognitive constraints among respondents. Scale bias refers to a rejection of near-proportionality between the respondents’ willingness to pay (WTP) for a risk reduction and the size of the risk reduction, a necessary (but not sufficient) condition for the estimates to be considered valid measures of the marginal WTP. This study is the first, to our best knowledge, which investigates the relationship between cognitive ability and near-proportionality between WTP and the magnitude of risk reductions. In the study 200 Swedish students participated in an experiment where their cognitive ability was tested before they took part in a CVM-study where they were asked about their WTP to reduce bus-mortality risk. In order to avoid anchoring effects the open-ended CVM format is used. The results imply that that the deviation from proportionality is less among those respondents with higher cognitive ability.

0256 The mission is remission – health economic consequences of achieving remission in antidepressant treatment of depression Andlin-Sobocki P. (Stockholm Health Economics, Stockholm, Sweden) Ekman M., Jönsson B. Objectives. There has been an increased argumentation over the past decade suggesting full remission to be the primary goal of the initial phase of therapy. There are however few studies investigating the health economic consequences of attaining remission in pharmacological treatment of depression in clinical practice. The aim of this study was to determine what explains remission in depressed patients in Swedish primary care and to investigate its impact on cost and health-related quality-of-life.. The study was designed as a naturalistic longitudinal survey of patients with depression in primary care setting. Data on patients’ socio-demographics, daily activity, healthcare resource use, quality-oflife (EQ-5D) were collected using questionnaires completed during outpatient GP visit for a follow-up period of approximately 6 months. Patient recruitment and data collection were performed at about 65 primary care centres in Sweden and 447 patients were enrolled. Disease severity was assessed with the Clinical Global Impression Severity scale (CGI-S). Cox regression analysis was used to analyse what factors influence the time to achieve remission.. 53% of the patients achieved remission at end of follow-up. Remitting patients had on average 7 outpatient care visits less per year than non-remitting patients (p<0.0001), and substantially fewer sick-leave days 56 days per year (p=0.029). Remitting patients had a significantly lower total yearly cost of SEK 50 000, compared to non-remitting patients. The health-related quality-of-life was 0.24 higher in remitting patients (p<0.0001). Severely depressed patients had a 70% lower chance of achieving remission quickly than patients with milder cases of depression (p<0.0001).. Remission has a substantial health economic impact and we have shown both statistically significant

reductions in cost as well as improved quality-of-life. This study has convincing evidence for the importance of aiming for full remission in the treatment of depression, and hence indicating that antidepressants that rapidly lead to remission may be cost effective.words

0257 Resource-use and costs associated with depressive patients in the primary care Andlin-Sobocki P. (Stockholm Health Economics, Stockholm, Sweden) Ekman M., Jönsson B. Background. According to the WHO, depression is the fourth leading cause of disability worldwide, which makes it important to provide effective treatment against the disease. Despite the clinical and economic importance of depression, resource utilization and cost in clinical practice is not well documented. The aim of this study was to investigate medical resource consumption, productivity loss and costs associated with patients treated with antidepressants for depression in primary care. 65 Swedish primary care practitioners recruited 447 patients to this naturalistic longitudinal obser vational study. Patients over 18 years with depressive symptoms, and who initiated a new antidepressant therapy were included in the study. Data on patients’ socio-demographics, daily activity and quality-of-life (EQ-5D) were collected using questionnaires completed during outpatient GP visit for a follow-up period of approximately 6 months. Disease severity was assessed with the Clinical Global Impression of Severity (CGI-S). Regression analysis was used in order to investigate the determinants of cost in depression.. Based on a complete sample of 398 patients, the total annual cost per patient was estimated at SEK 102 000 (SEK 92 000-113 000) in 2005 prices. Direct costs were estimated at SEK 36 000, constituting 35% of the total annual cost per patient. Among direct costs, the cost for medical visits was the largest single cost item, representing about 18%. The costs for antidepressants represented only 4% of the total costs. The indirect costs, i.e. productivity loss due to lost working time, were estimated at SEK 67 000 per patient (SEK 57 000-76 000), or 65% of the total annual costs per patient. No demographic variables were significantly associated with cost of depression. Direct and indirect costs were however correlated positively with achievement of clinical remission. The presence of sick-leave during follow-up was moreover associated with 82% higher costs. The burden of depression to society is high, both in terms of direct treatments costs and indirect costs for sickness absence and early retirement. Because of the high indirect cost per patient, there seems to be a particular need for therapies that have the potential to reduce absenteeism and early retirement. words depression, primary care, cost, cost-of-illness.

0087 Unresolved issues and challenges in efficiency measurement Andrew S. (Centre for Health Economics, York, United Kingdom) Stochastic frontier analysis (SFA) and data envelopment analysis (DEA) have become popular techniques for measuring organisational efficiency. Some key issues require further consideration before these techniques are fit for the purposes of regulation. Statements about relative organisational efficiency cannot be made unless the organisations being compared pursue a common set of objectives to which a uniform set of weights is attached. The appropriateness of the analytical model cannot be evaluated using standard testing procedures, because the aim of efficiency analysis is to extract individual rather than average effects. The evaluative task is more demanding the more complex the production process and the more heterogenous the environmental constraints each organisation faces. The historical context in which organisations are located can rarely be fully appreciated by the analyst, and this lack Eur J Health Econom Suppl 1 · 2006

| S9

of knowledge places constraints on model construction and interpretation. Further development is required before either SFA or and DEA are able to meet these challenges. An alternative approach is illustrated that allows analysis of organisations that pursue multiple objectives, offers greater flexibility in model construction, and is amenable to standard statistical testing.

0596 Expenditure control under DRG-based financing in the English NHS Andrew S. (Centre for Health Economics, York, United Kingdom) Global expenditure control has become a major challenge in the English NHS following the introduction of DRG-based funding (so-called Payment by Results, PbR). Primary Care Trusts (PCTs) – who are responsible for the commissioning of care on behalf of their resident populations – face increased financial risk for two reasons. First, under PbR they are unable to negotiate lower prices, having to pay the set national prices. Second, they have less control over volume, because patients have greater choice about where and when they are treated. This makes it difficult to specify volumes in advance with their contractual partners. We explore the ability of PCTs to manage demand so as to ensure that their budget allocations are not exceeded. We examine the instruments available to PCTs to influence the decisions of (1) general practitioners who influence the level of demand by virtue of their treatment and referral decisions; and (2) providers, particularly hospital consultants, who influence the extent to which demand is converted into activity by their admission and treatment decisions. We conclude that these instruments are insufficiently powered to manage expenditure, and that refinement of the payment system is required.

0367 Price Regulation and Competition in Japanese Pharmaceutical Market-A Case of Anti-infective Anegawa T. (Keio University, Graduate School of Business Administration, Yokohama, Japan) This paper studies Japanese pharmaceutical price regulation using data on popular anti-infective products between 1990 and 2002. The government regulates the official price of products for the National Health Insurance reimbursement, while it leaves the wholesale price to market competition. To evaluate the policy, this study focuses on the role of “price difference ratio” and the effects of product profiles including “dose form”, “product age”, “strength”, “package volume”. Products with higher strength, higher volume package, and co-promoted product are found to have higher. One percent increase in the raised demand by 3 percent in our sample, which effect has become even larger lately. This is due to the fact that pharmaceutical suppliers tend to lower the wholesale price of the “higher strength” and “larger volume package” of “capsule & tablet” products. Although the government applies the uniform pricing rule to all products, pricing of the wholesale price and demand for products vary with product profiles. Suppliers can increase demand by lowering the wholesale prices of products of the capsule and tablet form, larger package, and/or higher strength. The regulators are advised to take the products profiles into consideration to implement price regulation.

0211 Demand management when consumers are not expected utility maximizers Ania A. B. (University of Vienna of Economics, Vienna, Austria) Demand management is the standard way to alleviate the moral hazard problem inherent to health insurance markets. Consumers’; use of health services is disciplined through deductibles and co-payments. This is done under the assumption that consumers fulfill the axioms underlying expected utility theory. Extensive experimental literature suggests that consumers tend to over weight unlikely and to underestimate likely events. Moreover, their attitudes towards risk depend on whether the outcome involves gains or losses. This behavior does not conform to standard expected utility theory. Alternatives that would explain those anomalies have been proposed by rank-dependent utility and prospect theory. In the present paper we review insurance behavior and the consequences of demand management when consumers are affected by behavioral anomalies of the type described above. We show that (1) due to probability distortion alone, even risk averse consumers may leave likely contingencies uninsured, independent of how big the loss is; (2) optimistic consumers, who tend to underestimate the probability of severe illness, will buy more insurance for mild than for severe illnesses; and (3) for optimistic consumers, the problem of moral hazard is alleviated, i.e. they may indeed prefer partial to full insurance.

0343 Ambulatory autologous transplant Le Corroller Soriano AG. (INSERM, Marseille, France) Faucher C., Maraninchi D., Blaise D. and Moatti JP. We present the first randomised study comparing outpatient versus standard hospitalisation follow up after autologous peripheral blood stem cell transplantation for cancer patients. Endpoint of the study was the total number of days spent in hospital (visits in outpatient clinic + hospitalization) during the study period (from graft up to day60). Quality of life (QLQ-C30), anxiety (Spielberger’s State-Trait Anxiety), symptoms, satisfaction with the treatment (specific) and cost minimization (including indirect costs) were measured. 131 patients have been randomized (66 outpatient & 65 standard). Outpatient patients spent fewer days in hospital. 26 of them were not discharged at day 0 for non-medical reasons (no caregiver or refusal) and 10 patients for medical reasons. Only 30 patients were discharged at day 0. Quality of life results were in favor of conventional hospitalization for physical capacity, pain and fatigue (P=0.05). We concluded that outpatient autograft is feasible in a routine practice from the medical and hospital points of view. It appeared not desirable for the patient’s point of view. The anxiety / satisfaction evaluation (currently ongoing) will complete this assertion. The economic analysis will determine the putative benefit of the ambulatory design from several points of view (patients, health care system and collectivity).

0364 What Medical Conditions Mostly Drive Hospital Costs in Canada Between 1998-1999 and 2003-2004? Ariste R. (Canadian Institute for Health Information (CIHI), Ottawa, Ontario, Canada) Glussich A., Panait D. New medical technologies could drive both components of spending, which are price and quantity. There are some conditions related to medical treatments known to undergo substantial technological changes. We assume that hospital spending for such conditions grow faster than spending for conditions with less innovative treatments. We would like

S10 |

Eur J Health Econom Suppl 1 · 2006

to know whether empirical evidence supports this hypothesis. Objective of this study is to calculate the growth in hospital spending by category of diagnosis and to determine which component of costs is mainly responsible for driving hospital costs. Start by estimating the level and change in nominal hospital spending for the fiscal years 1998 and 2003 for twenty of the ICD-10-CA chapters (broad categories) and about sixty categories of diagnosis. We use a bottom-up approach, meaning the cost of a given condition is estimated with micro-level data. The cost data are obtained from the databases of Canadian provinces with major costing initiatives under way. Adjustments for co-morbidities were made using regression analysis. Number of cases is obtained from the Hospital Discharge Abstract Database. Average spending is estimated as the product of average cost by the number of cases. We then decompose spending growth into price growth and quantity growth. Preliminary results suggest that for specific conditions, rising number of cases, rather than rising cost per case, was mainly responsible for the spending growth; which is in line with some international studies.

0040 Risk Equalisation and Competition in Voluntary Health Insurance Markets Armstrong J. (Voluntary Health Insurance Board, Dublin, Ireland) Community rating has long been a cornerstone of the Irish private health insurance market. In 1996, following under the Third Non-Life Insurance Directive, a risk equalisation system was introduced in government regulation. Its aim is to prevent cream skimming and thus protect the community rating system that operates while still enabling a competitive health insurance market.number of modifications to the 1996 scheme have been made. However, effectively the same basis for equalisation operates now as originally. This is different than the basis for the mechanisms that operate in other countries such as Belgium, Germany, Israel, the Netherlands and Switzerland. However, no transfer have, of yet, been made between insurers and the central fund due to both legal and implementation issues.first purpose of this study is to review the history, structure and likely effectiveness of the risk equalisation system in Ireland. The second objective is to consider, first, the likely consequences for the competitive health insurance market of the absence of transfers and, second, the consequences should transfers commence. It will conclude that, with this absence, there is clear evidence that market outcomes consistent with cream-skimming have arisen and that this has led to higher premiums for some insured persons than would otherwise have been required. Furthermore, it will conclude that economically effective competition can only occur in the Irish health insurance market with risk equalisation.

0212 Community Rating in Competitive Health Insurance Markets Armstrong J. (Voluntary Health Insurance Board, Dublin, Ireland) Community rating is increasingly used as a tool to ensure that health insurance is affordable to everyone regardless to their risk profile. However, it is a term that is not always easily understood by policymakers in the context of a competitive health insurance market with multiple benefit packages. In this article it is argued that there are, in fact, many different methods that could be used to determine the appropriate community rated premium in such markets. Each method leads to differences in respect of premium flexibility, equity and market stability. Furthermore, the feasibility of introducing community rating in markets where risk rating already applies is reviewed. Finally, the regulatory elements that are necessary under the various forms of community rating are examined. Of particular relevance will be risk equalisation and the different approaches that could be used under each form of community

rating will be considered.paper will build on existing work in this area. Cost data is used from an insurer operating in the Irish health insurance market to determine the appropriate community rated premium levels that should be charged under the possible alternate community rating definitions. It shows that using the same data having the alternate definitions may have a considerable effect of the cost of the contributions for a contributing entity – be it the State or member. This can lead to a significant opportunity cost for the contributor.is concluded that while there are many different methods of applying community rating each one has different consequences is terms of equity and efficiency including, in particular, minimising the risk of adverse selection.rating is increasingly used as a tool to ensure that health insurance is affordable to everyone regardless to their risk profile. However, it is a term that is not always easily understood by policymakers in the context of a competitive health insurance market with multiple benefit packages. In this article it is argued that there are, in fact, many different methods that could be used to determine the appropriate community rated premium in such markets. Each method leads to differences in respect of premium flexibility, equity and market stability. Furthermore, the feasibility of introducing community rating in markets where risk rating already applies is reviewed. Finally, the regulatory elements that are necessary under the various forms of community rating are examined. Of particular relevance will be risk equalisation and the different approaches that could be used under each form of community rating will be considered. paper will build on existing work in this area. Cost data is used from an insurer operating in the Irish health insurance market to determine the appropriate community rated premium levels that should be charged under the possible alternate community rating definitions. It shows that using the same data having the alternate definitions may have a considerable effect of the cost of the contributions for a contributing entity – be it the State or not. This can lead to a significant opportunity cost for the contributor. It is concluded that while there are many different methods of applying community rating each one has different consequences is terms of equity and efficiency including, in particular, minimising the risk of adverse selection.

0476 Are the old wealthy healthier? Askildsen J. E. (Department of Economics, University of Bergen, Norway) Grasdal A. Usually inequalities in health status are associated with income and other socio economic variables. For elderly people, in particular retired receiving pensions, income dispersion is likely to be much less than for younger cohorts. On the other hand, wealth status may vary more among seniors. It is not obvious whether differences in wealth are caused by health condition, or whether wealthy people are in a better position to maintain good health. In this paper we explore the associations between wealth of older people compared to younger people, and their self reported health status, using 6 waves of the Norwegian Standard of Living Survey (SLS). Data from SLS are matched with data from the income and wealth registers at Statistics Nor way, providing information on income, transfers including pensions, and financial assets for all individuals in the survey. We use a fixed effect model to estimate the effect of lagged health on net financial assets of three age groups, 25–45, 46–67 and 68+.

Eur J Health Econom Suppl 1 · 2006

| S11

0166

0303

The cost-effectiveness of telemedicine for pre-clinical traffic accident emergency rescue

Critical review of the economic evaluation studies in Korea

Auerbach H. (Institute of Pathology, Berlin, Germany) Schreyögg J.

Bae EY. (Health Insurance Review Agency, Seoul, Korea/south/)

Objective: Since the European Commission set a goal of reducing the number of road accident deaths across the EU by 2010 to one half, the use of telemedicine for pre-clinical traffic accident emergency rescue is very high on the agenda. This study assesses the cost-effectiveness of two equipment versions of a telemedical device for pre-clinical traffic accident emergency rescues in Germany compared at baseline assumptions. Methods: Cost-effectiveness analysis with multi-way sensitivity-analysis is performed from a societal perspective. Costs resulting from personal injury, wasted journeys and erroneous dispatching of rescue services are based on data from the German Federal Highway Research Institute. The effects (life years gained) are based on probabilities for reducing therapy-free inter vals and improvements in first-aid provided by laypersons derived from two logistic regression models, based on data from the GIDAS database. Results: The cost-effectiveness ratio of the version “Automatic Accident Alert” (€ 247,977 per life year gained) would be slightly higher than of the full equipment version (€ 239,524 per life year gained). Decreasing costs of devices would disproportionately reduce total costs and rapid market penetration would largely increase the cost-effectiveness ratio of both devices. Conclusion: The net costs per life year gained in the application of the two versions of the telemedical device are very high. The implementation of the devices seems only realistic as part of a larger European coordinated initiative.

0323 The effects of reference pricing on ex-factory prices of Rx drugs in German social health insurance: A panel data approach Augurzky B. (RWI Essen, Germany) Göhlmann, Greß, Wasem This paper examines effects of reference pricing for prescription drugs in German social health insurance based on econometric panel data methods. We analyse the effect on ex-factory prices and on market entries as well as exits. Moreover, we investigate whether manufacturers adapt prices of their products not subject to reference pricing as a consequence of changes in reference prices of their products subject to reference pricing. We use a large panel data set of nearly all German prescription drugs on a monthly basis between January 2002 and July 2005. Altogether, the data comprise more than 1.5 million observations. They provide information on ex-factory prices, reference prices, manufacturers, market shares of the products, type of prescription drug, and market entries and exits. Our results show that there is no full price adjustment A 1%-change in reference prices leads to a 0.2%-change in market prices. Price adjustment is fast, it only takes two months. Furthermore, the first introduction of a reference price reduces market prices of the affected products by approximately 6%. Finally, we observe a significant time effect which is positive in the market without reference prices and negative in that with reference prices.

S12 |

Eur J Health Econom Suppl 1 · 2006

Objectives: This study aims to review the economic evaluation studies undertaken in Korea according to draft guidelines of Health Insurance Review Agency (HIRA), Korea Methods: All databases related to health service research in Korea were searched, and any CEA or CUA studies published by September 1. 2004 were screened. As a result, 34 studies were finally identified and reviewed. Results: Most of studies reviewed complied HIRA guidelines in general, but some points need to be addressed to assure the quality of information. First, many studies were based on the data sources of weak evidence like case series studies. Some studies (8 out of 34) estimated the effect from the meta-analysis of published studies, but none of them used the locally collected clinical data. Uncertainty also was not considered sufficiently. 16 out of 34 studies did not conduct sensitivity analysis, and most researchers did not pay attention to the uncertainty of quality weight. Conclusions: Economic evaluation studies undertaken in Korea have not aimed to be used in decision-making process until now, which seems to be the reason why many studies have not presented enough information needed in reimbursement decision. HIRA guidelines are expected to improve the quality of economic evaluation studies.

0587 The Impact of Payroll Taxes for Funding Health Care in New EU Member States Baeza C. (World Bank) Langenbrunner J. Most new member states rely upon high payroll taxes of 10-18 percent for funding health services. At the same time, countries face double digit unemployment and significant informal economies. The paper looks at impacts of this labor tax on health revenues and the economy as a whole, and considers other options for funding of health care services. Three countries are utilized as case studies: Hungary, Poland, and Slovakia.

0346 NICE ways of making health care decisions Baker R. (University of Newcastle upon Tyne, UK) Donaldson C., Shackley P., Robinson A., Smith R., Mason H. National Institute for Health and Clinical Excellence (NICE) issue guidance for the rest of the National Health Service in England and Wales about the priority to be given to new, and existing, health technologies. To date the focus of NICE economic evaluations has been calculation of the cost per quality adjusted life year (QALY) produced by health interventions. However, there may be considerations other than just maximising the number of QALYs that the public value and, further, these may differ from the considerations of health policy makers. This study therefore elicited the views of a sample of the general public and a sample of policy makers faced with making prioritisation decisions, using Q methodology. This methodology combines qualitative and quantitative techniques to investigate subjective opinion. A set of statements expressing views about priority setting was generated using focus groups. Members of the general public and health service policy makers then completed ‘card sorts’ (known as ‘Q sorts’) of these statements, ranking them from ‘most agree’ to ‘most disagree’. A form of factor analysis is used to identify common and differentiating patterns in respondents’ Q sorts. This paper reports the findings of this analysis and the implications for health care priority setting.

0011 Informal Payment in Hungary Health Care System Balázs P. (Department of Public Health, Semmelweis University, Budapest, Hungary) Hungary has a state run mandatory social security system with universal coverage for the health care expenses. Income levels of doctors contracted to the social health insurance are traditionally very low compared to other professions in this country, or related to the proportions of corresponding income groups in the developed industrial countries. Nine out of ten primary care doctors are now in the private sector, but the ownership does not play any important role, for practically all services, let alone the dental ones, are reimbursed by the single public fund of social health insurance. In the specialised care public institutions with salaried doctors represent an overwhelming majority of the outpatient and inpatient sector. private financing paid mostly in cash and without any private insurance coverage came about in dentistry and cosmetic surgery. Special comfort services in hospitals may be ordered also by the patients on social health insurance. But, some services, typically the home visits of family doctors, are covered only on the provider’s decision justified by the severe ill state of the patient. Besides the legalized waiting lists, there are shortages in supply and workforce bringing about the invisible waiting lists registered only in the pockets of public employee doctors, who represent their so-called own patients in the public hospitals. Choosing the doctor with all subsequent wishes must be paid for, but these are legally non-existing services in the public facilities.the four decades of socialist economy, in order to mask this private business (services mentioned above inclusive the optional home visits) Hungary developed an exceptional system. The patients were only thankful for a service beyond all legal definitions and paid “thanksgiving” money to the doctor. On this terminology, the ritual flair of transaction is and was the ideological backbone to justify this income since the middle of the 20th century. the fall of Communism, the problem and its terminology remained with us in Hungary. But, there is no need any more for informal bargaining and metaphoric language between the government and the medical profession. The real market economy does not tolerate an unregulated mixup of public and private financing. Solving this problem, Hungary must find the appropriate measures to convert the private purchasing power trapped now in the “thanksgiving” ideology without doing any harm to the public interests of the health care system.

0067 Dentists’ migration in the European Union and economic area (part 2) Hungarian dentist’s patterns of migration Balázs P. (Semmelweis University, Budapest, Hungary) Within the global migration of physicians, internal migration in the European Union and Economic Area (EU/EEA) is a specific problem. In this developed part of the world, rela-tively small differences may generate considerable regional shifts in both the numbers and professional training of the health care workforce. Moreover, the latest EU-enlargement in May 2004 opened the doors to a one-way migration of physicians out of the former so-called socialist Central and Eastern European (CEE) countries. In the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia, there are many factors influencing the movement of workforce toward the old member states. Socioeconomically, it has been due to a significantly lower level of domestic income. For socio-cultural rea-sons, the majority of doctors from the CEE appear to have sought employment in the UK or in the Germanspeaking countries. Unfortunately, for the time being there are no scien-tific and statistically verifiable, complex evaluations at EU/EEA-level of this migration. Obtaining adequate information requires a specific methodology and interactive data bases.order to propose a feasible future model, the authors have restricted their analysis to dentists, a rela-

tively small group of the health care community. Moreover, in dentistry, as opposed to the mainly publicly financed other branches of health service, there is a consid-erable contribution from private financing. This fact may enhance the socio-economic driving forces behind professional migration. The present Hungarian study, and its U.K. counterpart, are the first to shed light on dental migration at national and EU/EEAlevels. In 2004, Hungary had 10,2 million inhabitants and 5670 dentists, of whom 2487 worked in the public financed primary dental care. The remaining 2738 (included university staff mem-bers who run private practices) worked in the private sector. Proportion of immigrant dentists was found to be relatively high (7,81%). came mainly from large Hungarian ethnic minorities in two neighbouring countries (Rumania and Ukraine). Based on professional registration, the authors analysed migration, inland migrational attitudes of domestic provider and remigration of immigrants. Because of a lack of reliable data, the loss of domestic dentists’ to the workforce was calculated on a cohort-analysis of dentists who graduated from 1970 to 2005. The resulting data show, that there has been a relatively low international emmigration by Hungarian dentists. Internal migration of dentists has also been very low. Its driving factor appears to have been a con-siderable demand for private dental care in the Western part of the country.authors conclude that in future in order to monitor the migration of health care workers (and subsequently plan rofessional training and service capacities within the EU/EEA) it will be necessary to establish a compatible national and EU-wide system of registration and notification. The dental workforce would be an ideal group in which to pilot this system.

0410 Expected longevity and smoking in Italy Balia S. (University of York, UK) This paper investigates longevity, health, smoking behaviour and socioeconomic characteristics among the elderly using Italian data from the early (2004) release of the Survey of Health, Ageing and Retirement in Europe (SHARE). Individual lifetime utility is maximized at some optimal demand for health, where future utility and hence the decisionmaking process depend upon the probability of survival. Health status and the risk of mortality are affected by lifestyle and health investment decisions such as smoking, and subjective measures of life expectancy have been shown to be powerful predictors of actual mortality. The SHARE data provides a numerical measure for this, which is internally consistent and appropriately covarying with risk factors and other socio-economic variables. We use subjective sur vival probability as a proxy for mortality in a structural equations model, where the survival probability is explained by past individual characteristics, health status, health-related behaviours and unobservable characteristics. This motivates a recursive model for subjective mortality, self-assessed health and smoking duration that includes both observable and unobservable factors. Unobservable individual heterogeneity is considered by estimating a mixture model via the EM algorithm, which allows division of the underlying population according to different classes and estimation of class membership probabilities. Smoking duration is shown to have a strong impact on the probability of sur vival and unobservable frailty explains the heterogeneity in the population.

0337 Exchanging knowledge – using economic principles in priority setting practice Bate A. (Centre for Health Services Research, University of Newcastle upon Tyne) Donaldson C., Murtagh M., Ruta D. Healthcare resources are scarce and, thus, health organisations have to make choices about how best to meet objectives. Economics, as the ‘sciEur J Health Econom Suppl 1 · 2006

| S13

ence of scarcity’, founded on the principle of opportunity cost, should provide the theory and solutions to help these organisations undertake this difficult task. In the UK, the use of economics to inform priority setting is evident at the national level with bodies such as the National Institute for Health and Clinical Excellence (NICE) using health economic information to inform their decision-making. However, economics is less well developed in the context of local decision-making, and researchers are questioning the usefulness of promoting health economic methods, especially cost-effectiveness and cost-utility analysis, for priority setting at this level. In this presentation we discuss an alternative approach to developing the use of economics within local healthcare organisations. Using participatory action research (PAR) we worked with a local UK NHS commissioning organisation in the North-East of England to integrate economic principles into the current decisionmaking management process to create a more systematic framework for prioritisation. Observation, documentation of meetings, and focus groups were used to identify the barriers and facilitators to adoption of the framework; assess how the framework is interpreted and reconstructed by its users; and evaluate the usefulness of the framework from their perspectives. We present the results from one case study and examine the potential for the further development of economic systematic frameworks by identifying how they can be informed by, and inform ‘real-world’ decision-making.

0611 Development of a nation-wide quality indicator system in Hungary Belicza E. (National Health Insurance Fund, Budapest, Hungary) Takács E., Boncz I. The Hungar ian National Health Insurance Fund Administration (NHIFA) launched a program to evaluate and assess the quality of Hungarian health care providers in 2002. In Hungary health care is financed through social insurance fund, which is handled by the NHIFA as a central organ of public administration. For the indicator development the NHIFA’s working group uses the administrative data, which are reported by health care providers contracted with NHIFA. The NHIFA is the only health insurance company of the compulsory social health insurance in Hungary. The administrative data base contains among others the insurance identification number, sex, date of birth, ZIP code, date of treatment, diagnoses and interventions related to treated patients on every patient visit. The quality indicator program aims at the accountability of health care services, the encouragement and support of quality improvement projects and the quality improvement of administrative data. The indicators are calculated at national and hospital level. The providers are grouped and compared according to their patient mixes. At the initial stage of the indicator program the health care providers are not named but indicated with code number. Nevertheless, the providers know their code number, so they are able to compare their own results to other providers. According to our experience the following advantages and disadvantages of the use of administrative data can be determined. As the data base is available, it is cheap compared to other possible data collection alternatives. It is directly accessible, includes all hospital cases and makes the continuous development of the indicator system possible. We can connect the data about the provisions of a given patient even among more providers. Further advantage is that we can take some risk factors into account due to detailed data. There is also an opportunity to analyse the trend of indicator values as detailed data have been reported to NHIFA for more than 10 years. One of the biggest disadvantages is that the primary data collection objective, that is to get information for financing, makes the data biased. The lack of some necessary risk factors limits the possible analyses. Unfortunately, we have less data for process indicator than for the less interpretable outcome indicators. The publication of the first indicators triggered an intensive media reaction, which resulted in a more active participation of the expert organisa-

S14 |

Eur J Health Econom Suppl 1 · 2006

tions. They realised they should suggest some refinement of the screening criteria of involving cases and hospital grouping by a different methodology. One lesson for us is that we should put more emphasis on communication with professional organisations to make the program acceptable for them.

0178 An Alternative Approach to Estimating Medical Costs: The Case of Bariatric Surgery Banerjee R. (Mayo Foundation, Rochester, USA) Shah N. Rationale: Traditional cost accounting data are not always a good measure of resource costs for economic evaluations in health care because the fixed component of total costs is large. Since price is likely to be a major factor determining the demand for discretionary medical procedures that require high out-of-pocket payments, we evaluate firm behavior using standard economic theory. Objectives: We estimate a cost function for Bariatric surgery without using traditional cost data. Methodology: We assume the market for Bariatric surgery is imperfectly competitive and that medical institutions are profit maximizers. We estimate the cost function for Bariatric surgery using data on price, quantity and institutional features with data from the National Inpatient Sample for 2003. We derive our measure of price from the reported expected payer source. Results: Preliminary analysis reveals a mildly positively sloped marginal cost curve. This is robust to a variety of linear specifications. We find an inverse marginal cost elasticity of 0.14. Conclusions: Rising marginal costs may be evidence of capacity constraints. Typically, U-shaped marginal costs are considered evidence of large fixed costs or positive externalities. These are likely to be observed in time-series data for an individual firm or in a panel of firms. Our ongoing analysis uses panel data from 1998-2003 to further study these factors.

0511 The effects of ownership structure changes in the Hungarian health provider sector Baranyi L. (Ministry of Health, Budapest, Hungary) Mihalicza P., Lelkes I., The effect of the ownership structure of a health provider on the health care delivery has been discussed several times. There is no theoretical or empirical consensus in the literature whether the public or the private ownership is better for providing health care services. It is a quite new phenomenon in Hungary that the local governments – which are the owners of health care facility mostly – assign a private company to provide health care services. The company may use the public facility and will be responsible for delivering health care in the territory of the local government. This solution is not widely used today and only short term experiences are available we are going to try to evaluate them. The aim is not to make a clear statement which form of ownership is superior but to compare them according to the methodologies developed in the literature. Our expectation is that several useful lessons will be learnt from the Hungarian case.

0469 Competition among Health Plans: A Two-Sided Approach Bardey D. (University of Rosario (Bogota) and Gremaq University of Toulouse 1, Bogota, Colombia) Rochet JC. In this paper, we use a two-sided approach to modelize competition between Health Plans. Indeed, Health Plans compete on policy holders’ market but also on physicians’ market in order to attract them in their networks of care. Our paper reveals some interesting networks’ effects that occur in a competitive health insurance market. Health Plans providing a high diversity of physicians attract, in average, policy holders’ characterized by a higher risk. According to the mode of remuneration used, our two-sided approach shows how the Health Plans supplying a higher diversity of physicians can surprisingly benefit from this risk segmentation effect to obtain more important discounts and realize higher profits in equilibrium.

0128 Clinical, and Socio-Demografic Variables as Predictors of Services Usage and Costs in the Mental Health Sector Barrella A. (University of Rome Tor Vergata, Rome, Italy) Polacsek A., Ricci R., Spandonaro F. In the 80s the Italian Mental Health sector has witnessed the closure of Mental Home and reorganization of psychiatric services into a Mental Health Department (MHD) including community services, short and long term care, and hospitals psychiatric service.sector is currently characterized by Lack of services standardizationHigh presence of chronic patients, with occasional acute episodes requiring care with variable intensityA payment system financing DRG for hospital units and providing rates for community services, both criticized in literature.studies demonstrate that socio-demographic factors and diagnosis are important in ser vices usage prediction. They also emphasize that costs variation blurs the relation between patient attributes and ser vices usage, hindering the development of a system of perspective financing based on needs estimation.This project aims to define a model of Prospective Financing System for the MHDs, by classifying patients through the HoNOS scale, integrated by diagnosis and socio-demographic factors. this purpose, a 3-years research has been carried out. The survey, which collects data on 2.600 patients, has applied both a multiple regression and an analysis of multiple correspondences, to characterize patients’ profiles grouped according to common characteristics.

0484 Moral hazard and the demand for health services: a matching estimator approach Barros P. (Universidade Nova de Lisboa, Portugal) Machado M., Galdeano A.S. In this paper we estimate the impact of health insurance coverage beyond National Health Insurance on the demand for several health ser vices. Traditionally, the literature has tried to deal with the endogeneity of the private (extra) insurance decision by finding instrumental variables. It is hard to think, however, of any variable that a priori would be a good instrument and, therefore, we take a different approach. We concentrate on the most common health insurance plan in the Portuguese Health Survey, (ADSE), which is given to all civil servants and their dependants. We argue that this insurance is exogenous for most people i.e. not correlated with their health status. Under this identifying assumption we estimate the impact of having ADSE coverage on three different health ser vices using a matching estimator technique.

The measures of demand for health services are number of visits, number of blood and urine tests, and the probability of visiting a dentist. Preliminary results show large positive effects of ADSE for number of visits and tests among the young (18 to 30 years old) but only for tests are these effects statistically significantly different from zero. The magnitude of the effects represent 21.8 and 30 percent of the average number of visits and tests for the young. On the contrary, we find no evidence of moral hazard on the probability of visiting a dentist. Finally, we argue that there is evidence of a positive cumulative effect of ADSE on health over the years.

0092 Contractual design and public-private parternships for hospitals Barros P. (Universidade Nova de Lisboa, Portugal) Giralt X.M. The Portuguese Government has recently announced the launching of public privateprogramme to build new hospitals. Such partnerships have, the designation of PFI – Private Finance Initiatives – been already usedthe UK. The novelty of the Portuguese PFI lies in that both infrastructureand the clinical activities management will be awarded to a private. This makes a significant departure from PFIs in other countries, whereonly the infrastructure construction is given to the private agent. A careful economic analysis of the properties of this optionshed light on potential problems to be expected, and provide guidelines fordesign of public-private partnerships in general. It is also clear that itcoordination issues that do not arise in the traditional public-private, and therefore one needs to check whether usual prescriptions fromtheory do still hold.Portuguese PFI will force the creation of two entities, one in charge ofconstruction, the other with responsibility of clinical operations.shareholder structure of the winning parties in each contract is expected toto a considerable extent. The contract for infrastructure constructionhave a duration of 30 years, while the contract for clinical services managementbe of 10 years, renewable twice. A constraint present in the Portuguese programme, by design, is that the same core consortium will take both contracts in a first phase, though independent juridic entities will be in charge of each contract (despite the existence of a substantial overlap in shareholders identities), though not necessarily identical.implementation of this public-private partnership program demands acontract design, and the definition of what is included in the contract ofentity and what remains to be included in the contract to to the other one. Some activities are clearlyone side or the other. However, there are others that can conceivably be includedany of the two private parties. These are, essentially, the soft facilities, likeservices, catering, and the like.definition of boundaries may have a significant impact on theprivate economic agents are willing to take, and it constitutes part ofcontract design problem for the Government. The quality of these servicesbe easily observed by the entity in charge of clinical activities, though itwell be non-contractible in the sense that it cannot figure in the contractsby the Government.quality of these soft facilities may not be contractible in all relevant, the contractual design must provide the proper incentives. One instrumentto the Government is the choice of the entity that will managefacilities.simplify matters, we take a crude view of soft facilities management. Thefacilities can be included in the contract of clinical activities management. Alternatively, they can be included in thestructure institution, being directly paid by the Government. In the first option, the clinical activities management has to contract with the soft facilities provider its input. Another option would be the clinical activities management to provide directly the soft facilities services, a situation we term vertical integration.systems are linear on costs, which conforms well with observed practice. Moreover, contracts for infrastructure construction (and maintenance) and for clinicalmanagement must specify in full all economic transactions between theand the Government.focus on the external effects of having two contracts, one for soft facilities services provisionthe other for cliniEur J Health Econom Suppl 1 · 2006

| S15

cal services management. For that purpose, wefrom the usual asymmetric information problems. Thus, our approach isto the literature on sufficient vertical restraints than on optimal contracts.complements the information-based contracting analysis.turns out that first-best choices can be achieved under each regime (notsurprising, as we assumed away the usual informational problems), withdefinition of the payment rule. particular, we want to address here the optimal contract design under the assumption that quality decisions in soft facilities can be observed by the entity in charge of clinical activities but not by the Government. Clinical activities provide their own input, which combined with infra-structure and soft facilities quality, produces treatment to patients.can look at this issue from several perspectives. The more general one is to take it as a direct application of asset specificity theories, within the theory of the firm (Hart, 1993), and ask whether, or not, there is a motive for integration of soft facilities into the clinical activities management entity. At a general level, this takes us to the discussion of which party makes the more decisive specific-investment and which one is the more sensitive to the awarding of control rights.this line, it seems reasonable to argue that clinical activities should lead the process of vertical integration, as its marginal contribution and sensitivity of asset-specific investment is presumably greater than that of soft-facilities management. straightforward argument, however, is only part of the story as the existence of a payer to both entities providing services and the design of the payment schedules introduce some further structure into the model, and complementarities between inputs may play a decisive role.recent model of Jelovac and Macho-Stadler (2002) looks precisely at the issue of centralization and decentralization in the context of health care provision. Their primary example uses hospitals and physicians as the relevant agents. Decentralization in their setting means that an entity contracted by the payer will also set the contract to a second entity. Centralization means the payer offers directly contracts to both physicians and hospitals.terms of our problem, the agents will be the managers of clinical activities and the managers of soft facilities. The Government may set the contracts to both (centralization) or allow for the manager of clinical activities to set the contract for the soft-facilities manager (decentralization). According to the results of Jelovac and Macho-Stadler (2002), and assuming the contribution of clinical activities to the health of patients to be more important than that of soft facilities, we fall again in an argument in favor of including soft-facilities management within the clinical activities contract, and then let the managers of clinical activities hire and organize the soft facilities input. However, the setting of of Jelovac and Macho-Stadler (2002) leaves out a couple of elements that we believe to be also relevant to include first, the application of linear payment schedules, and, second, allowing for patients welfare to be a concern to providers as well, which takes us to a version of the model of Ellis and McGuire (1986).

0091 An assessment of the empirical validity of the EQ-5D, SF-6D and EuroQol VAS in a sample of the general population Barton G. (University of Nottingham, United Kingdom) Sach TH., Avery AJ., Jenkinson C., Doherty M., Muir KR. Objective: To assess the empirical validity of three health related quality of life (HRQL) measures: the EQ-5D, SF-6D and EuroQol visual analogue scale (VAS). Methods: Information on six personal characteristics (age, gender, ethnicity, smoking status, body mass index (BMI), occupational skill level), ten health conditions (back pain, hip pain, knee pain, heart disease, stroke, asthma, cancer, diabetes, rheumatoid arthritis and osteoarthritis) and HRQL was requested from 2770 patients aged ≥45 years in one general practice. Linear regression analysis was used to assess the relationship between each of the three HRQL measures (EQ-5D, SF-6D, VAS) and sixteen explanatory variables (the six personal characteristics and ten health conditions).

S16 |

Eur J Health Econom Suppl 1 · 2006

Results: Increasing age, worse BMI, lower occupational skill level, back pain, hip pain, knee pain, heart disease, rheumatoid arthritis, and osteoarthritis were associated with a significant loss in HRQL (on the EQ-5D, SF-6D, and VAS). As were having smoked regularly and diabetes (EQ5D and VAS), stroke (EQ-5D and SF-6D), asthma (SF-6D and VAS) and cancer (VAS). Conclusion: The EQ-5D, SF-6D and VAS were able to detect the independent loss in HRQL associated with many different personal characteristics and health conditions.

0246 Application of local search methods in cost-effectiveness analysis Barton P. (Health Economics Facility of Birmingham, United Kingdom) Jobanputra P., Bryan S., Burls A. Rheumatoid arthritis (RA) is a chronic condition for which a large number of disease-modifying anti-rheumatic drugs (DMARDs) are available. Typically DMARDs will be stopped after a time for various reasons. Thus the appropriate long-term strategy requires the use of a sequence of DMARDs. Any possible sequence is in principle a candidate for cost-effectiveness analysis. With 11 commonly used DMARDs to consider, there are a total of nearly 40 million possible sequences of DMARDs, and over 100 million sequences if subsets are to be considered as well. It is clearly not possible to test all these sequences. of DMARDs are compared using the Birmingham Rheumatoid Arthritis Model (BRAM). The BRAM allows realistic distributions to be used for the time spent on any DMARD, and the effects of a DMARD may depend on a patient’s previous history. The model works by generating a large number of virtual patient histories from which population mean costs and QALYs are estimated. This work illustrates the use of local search methods from operational research, including simulated annealing and genetic algorithms, to find the optimum sequence in a computationally feasible way. The stochastic natures of the model and search algorithms are made to work together.

0499 Age-related public expenditure projections for the EU 25 Member States Bartozs P. (European Commission, DG Economic and Financial Affairs, Labour Markets, Taxation and Quality of Public Finances, Brussels, Belgium) Part of the organised session by the European Commission: Policy choices for Europe- ageing and sustainability: “Through the Ageing Working Group of the Economic Policy Committee data has been gathered from all EU Member States and new projections are calculated for the increases in public expenditures related to demographic changes. Different scenarios have been calculated and sensitivity analysis has been conducted on different assumptions. The projections suggest that most of the increase in public spending due to purely demographic effects would be related to pensions, health care and long-term care. There are variations between the Member States and between the impacts of the different components of the projections”

0389 The Extent and the Depth of Drug Coverage in Canada Basu K. (Microsimulation Modelling and Data Analysis Division, Applied Research and Analysis Directorate, Health Canada, Ottawa, Canada) Gupta A. Objectives: study 1) the extent and patterns of prescription drug coverage in Canada and 2) the socio-economic analysis of out-of-pocket drug expenses.

Data: We synthesized data from diverse sources such as Survey of Household Spending, Survey of Labour and Income Dynamics, and Canadian Community Health Survey. The resultant micro database contains individual/family drug coverage information along with a diversity of socioeconomic characteristics. Methodology: The extent of drug coverage is estimated by using a logistic regression model with drug coverage (yes/no) as dependent variable and individual age, sex, income, rural/urban, education, marital status, source of income, employment status (full-time/part-time), and provincial dummies as independent variables. Out-of-pocket drug expense is estimated using a tobit model. This model uses out-of-pocket drug expenditure as a proportion of income as a dependent variable and age, sex, rural/urban, and the types of coverage as independent variables. Results: Compared to Ontario, households in all other provinces have higher out-of-pocket drug expenses. Households in Prince Edward Island, Quebec, Saskatchewan, and Alberta have relatively high burden. Males have lower burden than females, presumably, because females have relatively higher drug consumption. Compared to households not covered by a drug plan, households covered by private plans have a lower relative burden while those covered by public non-premium plans have a higher relative burden.

0140 Time trends and spatial spill-over effects in the Spanish public pharmaceutical expenditures Bech M. (University of Southern Denmark, Odense C, Denmark) Lauridsen J., López F., Sánchez M.M. Pharmaceutical expenditure in Spain is mainly funded by the Spanish National Health System, however, regulation of the Spanish drug market has provided very sparse cost-containment incentives for both consumers and providers. Pharmaceutical expenditure as a share of total health care expenditure has steadily increased from 18% in the mid 1990s to above 23% in the late 1990s. The focus of this study is to determine the dynamics across provinces and time causing pharmaceutical expenditure to increase. A regression model for per capita public pharmaceutical expenditure is analyzed, based on small-area data from the fifty Spanish provinces, observed annually for the period 1996-2003. Adjustment for residual correlation across time periods and residual variance instability is shown to be essential for proper inference regarding effects of exogenous determinants. Presence of parametric instability is demonstrated and found to be mainly captured using increasing or decreasing linear time trends in the coefficients. Finally, it is revealed that control for endogenous spatial spillover is highly important in order to obtain proper conclusions regarding the effects of determinants of pharmaceutical expenditure. The study thus adds to previous knowledge and modelling practice by demonstrating the fallacy of simpler traditional approaches alike OLS and further stresses the necessity of, not only adjusting for each of the above features, but integrating them into a unified framework when analyses are based on small-area health care data.

puzzling as the fear of overregulation is one of the major issues in the current political discussions in Germany and Switzerland. The predominant question is how to identify the optimal risk adjustment-regulation and how to prevent possible overregulation. paper presents a simple, theoretical model, based on insurer’s trade off between cost containment activities and risk selection in order to deal with the regulation problem in a positive manner. Over-, under-, und optimal regulation can be described. Although the model is quite general, it is possible to derive practical and unambiguous policy recommendations conditional on the given situation in the health insurer market. This implies the necessity for monitoring the market. description of a statistical framework needed to monitor market activities is described and illustrated with empirical examples from Switzerland. The paper also compares the derived policy implications with the actual discussion in Germany, Switzerland and the Netherlands. Without the need to refer to normative differences the recommendations of the model correspond with the different political debates in the three countries. This shows that the model produces practical and applicable guidelines for policy makers, confronted with the issue of risk adjustment regulation.

0024 Cost Sharing in Health Insurance Becker K. (University of Zurich, SOI, Zurich, Switzerland) Zweifel P. Health insurance is potentially subject to risk selection, i.e. adverse selection on the part of consumers and cream skimming on the part of insurers. Adverse selection models predict that competitive health insurers can eschew high-risk individuals by offering contracts with low deductibles or copayment rates, while attracting low-risk individuals with higher copayments, resulting in a separating equilibrium. This contribution seeks to determine whether in competitive Swiss social health insurance policies with deductibles in excess of the legal minimum do indeed serve as an instrument of risk selection. In a discrete choice experiment, effected in 2003, some 1,000 individuals were given the hypothetical choice of alternative insurance contracts that differed both in terms of deductibles and copayments and in benefits covered. Results suggest that individuals without medical ser vices during the past six months, tended to select a policy with a high deductible. Compensation demanded for voluntarily accepting an increase in the annual deductible also varies with socioeconomic characteristics and increases with the current level of deductible, as predicted by theory and constituting evidence in favor of the risk selection hypothesis. The experiment allows to compute necessary premium reductions and provides guidance for the pricing policy of insurers when offering differentiated products.

0109 Cardiovascular Technology variability of use in Spain during the period 2000-2004 at the Spanish NHS Belaza J. (Madrid, Spain)

0119 A General Model of Insurer Behavior to derive optimal Risk Adjustment Policy in a Competitive Social Health Insurance Market Beck K. (CSS Insurance, Lucerne, Switzerland) There is plenty of literature about risk adjustment in a competitive social health insurance market focusing more and more on the actuarial aspects of the problem – for example optimizing the R2 of the formula, optimal estimating techniques, availability of data etc.: saying comparably little about the problem of optimal regulation. This is particularly

Intro duction and objectives: In Spain there are 17 Autonomous Regions that have the responsibility of Health ser vices providing and funding through devolution of a % of the income tax as well as through the use of specific taxes on alcohol, etc…The role of the Central NHS is basically one of co-ordination, approval of new drugs and health devices, price fixing for pharmaceuticals as well as recommendations on health policy through different mechanisms such as what is called an integral health plan for an specific disease such as CVs or cancer by setting the standards or care to be applied and resources and practices to be incorporated at the National level by each Autonomy. Being Spain a country that in the last ten years 1993-2003 dedicated a fixed percentage of the GNP to public medicine at the expense only of Eur J Health Econom Suppl 1 · 2006

| S17

the increase of the private insurance, we could ask ourselves if the difference of 1-2% of the GNP invested in the market would not affect the degree of technology diffusion at the level of the Autonomies as also the health expenditure per capita shows, big variations between them. In a study done by FENIN in Spain reviewing the 2004 health budgets for several Autonomies there was an average 6% dedicated to health technologies (including all kind of products, implants, medical devices and heavy equipment investment) what in the EUCOMED 2003 profile was estimated in 75€ per capita : 25% below UK and Italy, 50% below France and 70% below Germany. This preliminary analysis should make as think on the possible existence of variations at the level of use of high technologies, where the degree of expertise and resources needed is big and therefore requires an investment in facilities and human resources as well as expertise. Due to the fact that CV diseases are the main reason of mortality in Europe and at the OEDC by representing between 30–50% of the total mortality, we have decided to study the variability of use of three CV technologies used in the treatment of these diseases: ICP (intra-coronary procedures to treat blocked arteries.), ICD (implantable defibrillators to treat malignant arrythmias that happen in patients mainly after an MI) and CRT (cardiac resynchronization devices to treat Heart Failure patients to improve his quality and decrease its mortality). Material and methods: We have reviewed the data on the use and implant of three technologies (ICP.ICD and CRT) in the period 2000– 2004 by accumulating the data from all hospitals belonging to each Autonomous region and dividing this value by the population in order to get an indicator per million of habitants. To measure the variability we have calculated the VR (variability ratio or ratio between the higher and lower values) and the Non Weighed Coefficient of variation (CV, calculated as the ratio between the standard deviation and the mean values). The results for the year 2004 are the following: ICP (1.97, 0.23), ICD (4.5, 0.33), CRT(10.6, 0.64) and including the analysis for the implant of pacemakers a technology older than 40 years PM (1.97, 0.23).This means that the most stable technology is ICP and that the variability is higher for the ICD and specially for the CRT (5 years life technology). In order to study the situation on the year 2004 we used an econometric model including offer variables (Health expenditure per capita, Gross Regional product per capita, No. of centers per million) and demand variables (% of population over 40+ and over 65+ years old mortality, incidence and prevalence of diseases) Results: The best explanatory variables were the Gross Regional Product and the No. of centers per million at each region (offer variables) while the demand variables related with the burden of disease had a very low explanatory power. In the case of HF and the use of CRT the model included just the No. of centers and the % of population of 40+years although in this case the coefficient was negative, what can be explained by the fact that this therapy is diffusing earlier in the more rich regions that are not necessarily those with higher prevalence of HF. The total variability explained is almost 65%, F = 15, p=0.00042 and both variables being quite significant with a p <0.01. A more detailed analysis will be presented with models found for each technology. Conclusion: The variability is lower for the ICP and higher for ICDs and CRT. The diffusion of rather young technologies is facilitated by the richness of each region as the experience and facilities can be more easily acquired that in the others. For all technologies, even for very stable ones such as PM, there exist a medium to high variability of use that creates for the patient a situation of lack of equity at the SNS level. In Spain the level of use of these three technologies is well below the EU-15 average even taking into account the lower incidence of Ischemic Diseases in our country. An econometric modelling for 2004 reinforces the predominant role of the Region richness as an explanation of the variability. This technologies are also under -prescribed compared with similar European countries, according to current clinical guidelines, but what it is worst, with the existence of an unacceptable level of inequity (2:1, 4.5 : 1 or 11 : 1) according to where the patient lives.

S18 |

Eur J Health Econom Suppl 1 · 2006

0528 Hungarian cost-utility analysis of anastrozole vs tamoxifen in postmenopausal women with early breast cancer (EBC) Benedict A. (MEDTAP Institute at UBC, Budapest, Hungary) Objectives: In the ATAC trial, anastrozole produced significantly longer disease-free survival and time to recurrence compared with tamoxifen in postmenopausal women with hormone receptor-positive (HR+) EBCafter 5 years of treatment. A cost-utility analysis of anastrozole compared with tamoxifen was undertaken in Hungarian setting for reimbursement submission. Methods: A Markov model based on patient-level trial data was used to project 5-year outcomes from the ATAC trial to 25 years. Resource utilisation data were obtained from Hungarian databases of the National Insurance Fund. Unit costs (in 2004HUF) were obtained from Hungarian DRGs and outpatient fee schedules. Utility scores from a UK study in postmenopausal EBC women were incorporated. Costs and benefits were discounted at 5%. Incremental cost effectiveness ratios (ICERs), 95% CIs, and acceptability curves were calculated. Results: The ICER for anastrozole compared with tamoxifen at 25 years was HUF 4.77 mn/QALY (CI: HUF 2.38-15.1 mn). Results were sensitive to the cost of metastatic treatment and the assumptions about the duration of treatment benefit. Conclusions: Compared to a cost-effectiveness threshold of ¬20,000, often quoted in countries of the EU zone, anastrozole provides a costeffective alternative to generic tamoxifen for primary adjuvant treatment of postmenopausal women with HR+ EBC.

0207 National medical guidelines and priority setting in Sweden Berggren F. (National Board of Health and Welfare, Stockholm, Sweden) Kärvinge C. The aim of the health care sector is to provide good health on equal terms for the entire population. The work with national guidelines provide evidence based ranking by prioritising interventions linked to a specific condition. An ethical platform consisting of human value, need and solidarity and cost-efficiency provide the foundation of the priority setting. Three groups work together with a core team to provide a guideline that ranks interventions for specified indications from 1 to 10. The groups consist of a medical scientific group, a health economics group and a priority-setting group. These groups provide the core group with evidence-based documentation where the core group delivers a priority setting league table. Implementation is evaluated through studies of clinical practice. Guidelines for stroke, asthma and COPD, heart diseases and venous thromboembolism have been published. Guidelines for depression, dementia and cancer (breast cancer, colorectal cancer and prostate cancer) are under way. This form of national guidelines provides the possibility to draw conclusion concerning allocation and reallocation of resources and have been well received by the county councils that provides the health care. The heart disease and stroke guidelines are the first set of guidelines that have baseline investigations on clinical practice completed.

0396 Admission rates or length of stay: What is more important for cost-containment policies? Bernal-Delgado E. (Institute for Heath Sciences, Aragon, Spain) Peiró-Moreno S., Ridao-López M. Background: From a population-based perspective, hospital expenditure could be considered as a function of admission rates, hospitalization

costs and cost of adverse events. Cost-containment policies focus pretty much on hospitalization costs (DRGs as standard), but small area variation studies suggest admission rates containment as a better option. Methods Population: We selected 2002 hernia repair and cholecistectomy hospital admissions performed in 95 healthcare areas (from 10 Spanish regions). Main outcome: procedure-specific expenditure per inhabitant and year; Predicting variables: procedure-specific admission rates per 10,000 inhabitants, length of stay (LOS), percentage of procedures using ambulatory surgery, laparoscopic technique, percentage of emergency readmissions 30 days after discharge. Analysis: Pearson coefficients to determine significant correlations, and multiple lineal regression were used. Results: Admission rates explained 84% (hernia repair) and 62% (cholecistectomy) of expenditure variance. No other variables were significant in hernia. However, cholecistectomy expenditures were also explained by LOS (14%), readmissions (8%) and percentage of ambulatory surgery (5%). When multiple regression was used, admission rates, LOS and readmissions were significant, explaining about 97% of expenditure variance, both in hernia repair and cholecistectomy. Conclusions: Admission rates were found the most important factor explaining expenditure per habitant. Cost-containment policies should be reoriented and focused on admission rates management instead of LOS management.

0023 How does access to capital impact hospital financial and operating performance? Bernet P. (Florida Atlantic University, Florida, US) Hsuan F. Nonprofit hospitals in the U.S. raise capital primarily through bond issues. On one hand, this capital can be invested in new plant and equipment that can improve operating efficiency, which may further be reflected in improved financial results. On the other hand, the additional burden of debt repayment may force the hospital to make cuts that ultimately hinder performance. This study explores this relationship through an analysis of all not-for-profit hospital bond issues between 1993 and 2002. Key measures of financial and operating performance are measured both before and after each bond issue to determine if ratios improve as a result of the new infusion of capital. This study has the added benefit of being able to account for self-selection bias, as the same performance datum is collected for hospitals that choose to not issue debt. Results will also indicate those hospital characteristics that improve the odds for success after debt has been issued. This study may better inform facility funding decisions to help insure that investments are made where they will be most effective.

0248 Choice of parametric distributions in decision analytic models and expected value of perfect information Bischof M. (Institute for Clinical Epidemiology, University Hospital Basel, Switzerland) Sendi P. Background: Probabilistic sensitivity analysis (PSA) allows us to address second order uncertainty in decision analytic models. This usually entails the choice of probability distributions for relevant input parameters in a cost-effectiveness model. It is unclear how the use of different alternative parametric distributions may influence the results. We examined the impact of using different distributions for major model parameters on the expected value of perfect information (EVPI), which is a means of summarizing uncertainty in a monetary way. Methods: A probabilistic Markov model for the analysis of the costeffectiveness of preventive treatments for osteoporosis was used and dif-

ferent parametric distributions were applied to those groups of parameters with the highest expected value of sample information (EVSI) (treatment effect: lognormal, normal; quality of life: beta, gamma, truncated normal; costs: gamma, normal). The overall EVPI was calculated for different combinations of input distributions. Results: The EVPI may vary by more than 50% depending on the choice of input distribution in our model. Conclusion: The choice of parametric distributions can substantially affect overall uncertainty as reflected by the EVPI. Presenting the results for a variety of different parametric distributions can help to address this issue.

0360 Productivity in Psychiatric Out-patient Treatment Björngaar J. (SINTEF Health Research, Trondheim, Norway) Magnussen J. There has been a demand for a productivity increase in Norwegian outpatient clinics. This demand has however been criticized for not considering possible negative side-effects, such as reduced quality in treatment. paper raises three research questions. To what extent are differences in productivity between therapists a function of individual characteristics, such as formal qualifications and experience? To what extent are differences in productivity caused by characteristics of the psychiatric team, such as team function, size and professional composition. Are productivity differences associated with differences in average levels of patients’ experiences in the teams? Data were collected from two separate sources. Data on time spent on different activities in all Norwegian adult outpatient clinics and teams were collected by registration forms for a period of two weeks in September 2004. The material consists of 90 outpatient clinics, from 180 teams and consists of 1700 therapists. In addition we have data on experiences/satisfaction from 6 677 patients who received treatment in out-patient clinics and teams in September 2004 is measures using data envelopment analysis. Inputs are measured as number of hours spent on patient related activity by type of profession and experience. Outputs are measured as number of concultations and an index of quality which is contructed from the patient satisfaction data.

0171 Economic Evaluation of Valsartan in Patients with Chronic Heart Failure: Results From the Val-HeFT Study Adapted to the Netherlands Cornelis B. (University of Groningen, Department of Social Pharmacy, Pharmacoepidemiology and Pharmacotherapy, Groningen University Institute for Drug Exploration (GUIDE), Groningen, the Netherlands) Groot M., Jasmina J.I., Boom G., Postma M.J. Background: The Val-HeFT-study was a multinational randomized trial, in which a total of 5010 patients with NYHA class II to IV heart failure were assigned to valsartan or placebo (added to background therapy). Valsartan reduced heart failure hospitalizations significantly. Objectives: To conduct an evaluation on the economic impact of valsartan based on Val-HeFT including subgroup analyses following valsartan’s recent registration for CHF in Europe. Methods: Resource-use was based on data collected during Val-HeFT. These were multiplied by Dutch cost estimates and categorized in costs for hospitalizations, inpatient and outpatient services, ambulance transportation, death outside the hospital, and outpatient cardiovascular medications (1999 price-level; 3% discount rate). Results: Mean patients’ follow-up was 23 months and costs for hospitalizations were €617 lower among valsartan patients. Mean total costs for valsartan and controls were €8,810 and €8,441 respectively, resulting in incremental costs of €368. In patients receiving an ACE-inhibitor but Eur J Health Econom Suppl 1 · 2006

| S19

no beta-blocker these incremental costs were even lower, €171. There was an overall reduction in costs of €1,311 in patients not receiving an ACE-inhibitor at baseline. Conclusions: Valsartan provides clinical benefits at modest costs in the Netherlands. Subgroup analyses show that valsartan therapy is cost-saving and increases survival for patients not receiving ACE-inhibitors.

0380 How much are we spending to treat diseases? Bohigas L. (Barcelona, Spain) Gisbert R., Brosa F., De la Puente ML Healthcare public budgets tend to focus in resources (i.e. personnel, goods and ser vices) or ser vices (hospital, primary care, etc.) but they do not explain which diseases the public healthcare systems try to prevent or treat. This paper presents the results of analyzing in which diseases the public Catalan healthcare budget is spent. The methodology used is the one of disease costing, but taking in account the whole public budget. The disease unit is the ICD9 main categories. The results help to illuminate policy makers and planners of Catalunya about the priorities implied in the budgetary system.

0313 Introduction of health economics into decision making on public health interventions Boncz I. (National Health Insurance Fund Administration (OEP), Budapest, Hungary) Sebestyén A., Betlehem J., Oláh A., Dózsa Cs., Gulácsi L. Purpose of this policy paper is to give an overview on the introduction of health economics into decision making on organized nationwide screening programmes. Methods: We analysed the available evidences in the Hungarian setting on the decision making process on breast, cervical and colorectal screening. Results: After international scientific evidences were critically reviewed, pilot screening programmes were introduced in Hungary in all the 3 cancer type. Health economics (cost-effectiveness) analysis was conducted for breast, cervical and colorectal cancer and the results measured by the cost per life years saved were acceptable from financial point of view. Budget-impact analysis was prepared also. Finally, in nationwide organized screening programmes women between the age of 45-65 are invited by personal letter for mammography breast screening with 2 years screening inter val; women between the ages of 25-65 are invited by a personal letter for cervical screening, with a 3 years screening interval. The decision on turning the colorectal cancer screening pilot into a nationwide programme is still pending; currently persons between 45-65 are invited with a 2 years screening interval. Conclusions: The implementation of nationwide organized cancer screening programmes was based on careful evaluation of medical evidences and health economics analysis.

0514 Impact of illegal payments on the health care facility structure in Hungary Bondar E. (Ministry of Health, Budapest, Hungary) Baranyi L. Despite the totalitarian power structure of the society at large, the “planned character” of health systems of our region the development of the facility network was out of control of official policy makers due to a hidden and uncontrollable informal incentive penetrating almost all segments of the delivery structure. This is the “gratuity” payment. This fenomenon can be discussed by different (ethical, justice etc) points of

S20 |

Eur J Health Econom Suppl 1 · 2006

view, but it is important to analyse it from the point of view of impact on the delivery structure. The parts of health system (professions, type of care, geographical, etc) where patients were willing and able to be “grateful” are typically better developed (staffed, equipped), than other segments. Where shortage experienced in the number of grateful patients, the staff is smaller than required and providers are not keen on developing services. This illegal incentive is recently complemented by an other type of illegal payment, the promotion of producers and distributors. By and large gratuity is about discrimination, the promotion is an incentive to widen the patients cured regardless of their social position. Both impacts can be estimated in order to make a policy recommendation.

0036 Are health insurers able to channel patients towards preferred providers? An empirical study Boonen L. (Erasmus University, Rotterdam, Netherlands) Schut FT. Aim: This paper investigates the effects of various incentive mechanisms employed by health insurers to channel consumers towards preferred pharmacies. Method: Using two datasets on individual choice of pharmacy from two different Dutch health insurers (N≈45,466), we examine the impact of financial and non-financial incentives on provider choice for different types of consumers. We estimate a logit model of pharmacy choice in which the various incentives to visit a preferred pharmacy, distance (as a proxy for time price) and consumer characteristics are the most important explanatory variables. Results: Consumer choice of pharmacy appears to be quite sensitive to the financial incentives, while the non-financial incentives appear to be a less effective channelling device. The main driver of consumer choice, however, appears to be the distance from the home address towards the pharmacy. Conclusion: The results show that, even though relatively few consumers switched to a preferred pharmacy, (financial) incentives have a significant effect on consumer choice of pharmacy. Distance appears to play a dominant role in the choice of pharmacy by consumers. Particularly in non-urban regions, this may limit the opportunities for health insurers to channel their subscribers to preferred pharmacies.

0498 Striving after the general extension of the transparent coverage policy system – transformation of coverage system over medical devices Borcsek B. (Ministry of Health, Budapest, Hungary) Dózsa Cs., Boncz I., Nagy J., Daubner M., Rupnik E., Ilku L, Lakatosné Kovács A. Background: The rapid succession of health technologies presses financers to evaluate constantly which health technologies are to be subsidized, which aren’t. In order to manage this difficulty in Hungary the National Health Insurance Fund Administration(NHIFA) and the Ministry of Health(MOH) have developed a complex coverage policy system. Objectives: The main objectives are realizing cost-effective health care through economic evaluations and HTAs, making decisions in the interests of the patients, and maximizing the social benefits. Methodology: At first the coverage policy system of pharmaceuticals has been developed, because in consequence of the EU accession Hungary had to adapt the transparency directive in the reimbursement policy. The future aim was the general extension of the transparent coverage policy system for the medical devices and medical procedures. Findings: The MOH and the NHIFA have been very keen on extending the predictable, transparent reimbursement process over other health

technologies as medical devices, diagnostic and surgical technologies applying the experiences about the current system on the pharmaceutical reimbursement because only a comprehensive coverage policy can ensure that cost-effectiveness be a major aspect of the decision making process on whether something should be funded or not. The legal framework and the procedure – which are under development – are based on the methods and experiences of the pharmaceutical coverage system but different approaches have been taken into consideration because of the specific nature of these types of technologies. We would like to demonstrate the experiences of the codification process, the change-over stage and the further aims presenting some considerable cases(eg. DES) on this topic.

0352 Adverse selection and risk equalization in the italian insurance markets Borgia P. (CEIS SANITA, Rome, Italy) Doglia M. Foreword problem to financial cover health expenses represents a subject of central importance not only for the distribution of resources within the public health sector, but also for the offer of services stemming from it. The private health sector contributes to the increase of opportunities for treatment and it often represents a more flexible alternative to the satisfaction of individual preferences., databank and methodologyaim of the paper is to analyse the economic health risk of an insured sample. This has been obtained by a multivariate analysis and regression on data collection coming from Bank of Italy, National Institute for Statistics, Department of Health and Italian Association of Insurers. In particular the consumption behaviours, the income balances and their demographic and professional characteristics have been observed. health expenses are mainly sustained directly by consumers while only a residual part of it is guaranteed through a voluntary health insurance. insurance contributes to an improvement of access opportunities in treatment services. However its development is still limited.study shows as adverse selection represents one of the main obstacles for a future integrative health insurance development and suggests some interventions for risk equalization.

0079

ence was not statistically significant (mean difference –€114, 95% CI –€1003; €933). Cost-effectiveness planes indicated that there were no significant differences in cost-effectiveness between the two groups. Conclusion: This disease management program for major depression in elderly primary care patients had no significant effects on clinical outcomes, costs and cost-effectiveness. Therefore, continuing usual care is recommended.

0556 Fourth hurdle in Hungary – latest developments and future challenges Brandtmuller A. (Center for Public Affairs Studies, Budapest, Hungary) Karpati K., Boncz I., Gulacsi L. Rationale: The fourth hurdle is a live question in the Hungarian health care system. The pursuit of a more prudent health ser vice provision, budget constraints and the new regulation introducing the fourth hurdle in the pharmaceutical market encourage a new way of thinking. The Hungarian health economic guideline was developed on the ground of international counterparts and was issued in the official gazette of the Ministry of Health in 2001. After joining the EU and with the promulgation of the new legislation its recommendations are supposed to be followed in health economic analyses. Objectives: The study aimed to investigate the recent pharmaceutical coverage decisions and the opportunities for development some years after the issue of the Hungarian health economic guideline. Methods: We investigated the dossier of 25 innovative drugs submitted recently, to see if basic information are available for decision-makers. Results: There are encouraging developments in regulation and institutionalization: a new HTA instituiton and decision processes were established. However, coherency and transparency of coverage decisions, and the use of economic evidences are still to be improved. Analysis of 25 dossiers showed that decisions are primarily grounded on clinical effectiveness and importance for public health. The size of the target population, budget impact and Hungarian cost-effectiveness evidence are rarely available, however international clinical evidences and HTAs are presented. The biggest challenges are to eliminate the considerable shortage of Hungarian data, and to promote the transferability of international analyses.

0314

Cost-effectiveness of a disease management program for major depression in elderly primary care patients

Reference pricing of pharmaceuticals

Bosmans J. (EMGO-Instituut, Amsterdam, The Netherlands) Bruijne M., Hout H., Marwijk H., Beekman A., Bouter L., Stalman W., Tulder M.

Brekke K.R. (Norwegian School of Economics and Business Administration, Department of Economics, Bergen, Norway) Königbauer I., Straume O.R.

Background: Major depression is a common disorder in older adults and is associated with increased health care costs. Depression often remains unrecognized and undertreated in older adults, especially in primary care. Objective: To evaluate the cost-effectiveness of a disease management program for major depression in elderly primary care patients compared with usual care. Design: Economic evaluation alongside a cluster randomized controlled trial. Methods: Clinical outcome measures were severity of depression, recovery from depression and quality of life. Resource use was measured over a 12 month period using interviews and was subsequently valued using cost prices. Confidence intervals around cost differences and the uncertainty surrounding the cost-effectiveness estimates were estimated using bootstrapping. Results: There were no significant differences in clinical outcomes between the intervention and usual care group. Mean total costs were €1784 in the intervention and €1898 in the usual care group. The differ-

We consider a therapeutic market with potentially three pharmaceutical firms. Two of the firms offer horizontally differentiated brand-name drugs. One of the brand-name drugs is a new treatment under patent protection that will be introduced, if the profits are sufficient to cover the entry costs. The other brand-name drug has already lost its patent and faces competition from a third firm offering a generic version perceived to be of lower quality. This model allows us to compare generic reference pricing (GRP), therapeutic reference pricing (TRP), and no reference pricing (NRP). We show that competition is strongest under TRP, resulting in the lowest drug prices (and medical expenditures). However, TRP also provides the lowest profits to the patent-holding firm, making entry of the new drug treatment least likely. Surprisingly, we find that GRP distorts drug choices most, exposing patients to higher health risks.

Eur J Health Econom Suppl 1 · 2006

| S21

0125 The shortage of organs Breyer F. (University of Konstanz, Germany) Kliemt H. In Germany, as in other European countries, there is a severe shortage of donor organs. In particular with kidneys, the annual number of new entries to the waiting list exceeds the number of transplantations by more than 1,000, and hundreds of patients on the waiting list die every year. Furthermore, while 2 out of 3 Germans declare their “principal willingness” to donate their organs after death, only 12 per cent possess a donor card. The paper is organized in three parts. Part 1 provides estimates of the annual demand and the potential supply of donor organs. Part 2 identifies reasons why the actual supply remains far below potential supply. Emphasis is placed on insufficient reimbursement for hospitals and lacking incentives for potential organ donors. Part 3 is devoted to an analysis of possible remedies. The controversial literature on a number of radical reforms (such as the presumed consent model, reciprocity in organ allocation and payments to living kidney donors) are discussed in detail. The authors conclude that the organ shortage is due not to a law of nature but to inappropriate social institutions.

0074 Patients’ preferences for health care system reforms in Hungary Akkazieva B. (University of Heidelberg, Germany) Gulacsi L., Brandtmuller A., Pentek M., Bridges J. Recently, there is a growing focus on patients’ opinions/preferences to develop a better evidence base for health care policies. The objective of the study is to understand patient preferences for health care reforms in Hungary as a means of informing the future health care policy. A conjoint analysis was designed and administered to the study population within rheumatology out-patient centre in Flór Ferenc County Hospital, Hungary. Attributes and attribute levels were developed on the basis of key informant interviews and literature review. A mixed effects linear probability model was estimated, holding all respondent characteristics constant and correcting for clustering. Conjoint Analysis questionnaires were distributed by a physician to the patients in a clinic and additionally mailed by post. Regression results (R2=56.8%) indicated that patients preferred a health system that was not to be cost constrained (P=0.003), one that was based on solidarity (P=<0.001) and one where patients were empowered (P=0.024). Further, they would choose a system with no choice over provider to avoid paying a co-payment (P=0.005). This study demonstrates that patients have clear preference over health care system’s policy so that the stated preference methods, such as conjoint analysis, could be used to develop evidence based health care policy.

0562 Cost-effectiveness assessment of infliximab therapy in Hungary – adaptation of a Swedish rheumatoid arthritis study Brodszky V. (Corvinus University of Budapest, Hungary) Májer I., Péntek M., Kobelt G., Gulácsi L. Objectives: Several studies have shown the cost-effectiveness of infliximab, an effective but costly new biological drug in rheumatoid arthritis (RA). Hungary is facing the difficulties of introducing the “fourth hurdle” once cost-effectiveness analysis is a demand before reimbursement but resources are limited for health technology assessments. The aim of our study was to assess the cost-effectiveness of infliximab in RA in Hungary using country-specific economic and health status data and clinical results of a Swedish follow up study. Our analysis should serve first

S22 |

Eur J Health Econom Suppl 1 · 2006

experimental results of the methodological issues emerging with the adaptation of international cost-effectiveness models into a new European Union member state. Materials and methods: The analysis was based on our previous costof-illness study involving 255 RA patients in Hungary, which has established the costs, disease activity and health status on different disease severity levels. Patients who are candidates for biologic therapy and correspond to the sample of the Swedish study were selected. We estimated the utility gain of 1 year infliximab treatment in the Hungarian patients using the results of the clinical study, with sensitivity analysis for three different assumptions. To assess the marginal costs of infliximab therapy we considered the transitions of patients between different disability levels and the consequent cost-savings, involving treatment costs of infliximab therapy in Hungary. The time horizon of the analysis was 1 year. Both direct and fric tion cost methods and human-capital approach were used. Results: 43 patients were eligible for the criteria. The mean utility gain (QALY) was 0,338-0,372 according to the various assumptions. Costsaving after 1 year infliximab therapy was 155 EUR (only direct costs), 380 EUR (friction cost method) and 544 EUR (human capital approach). The incremental cost-effectiveness was estimated as 25833–189570 EUR/QALY Conclusion: Our results offer basic preliminary data on the cost-effectiveness of infliximab in Hungarian RA target population. Though our study has limitations, our results reflect that it is within the generally accepted threshold also in Hungary. Our study confirmed that collaboration between health technology assessment units and harmonization of the methods used might speed up and give a positive spin to the comprehensive implementation of “fourth hurdle” in the EU.

0243 A dollar is a dollar is a dollar Brouwer W. (Department of Health Policy & Management University Rotterdam, Netherlands) Exel NJA., Baltussen R., Rutten FFH. It is normally stated that an economic evaluation should take the societal perspective and that this implies the incorporation of all costs and effects, regardless of where these occur. However, this broad perspective may be in conflict with the narrower perspective of the health care decision makers we are trying to aid. We argue that not all costs have to be considered equally important for health care decision making and that there is a discrepancy between the economically preferred societal perspective and the aim of aiding health care decision makers. This is related to the theory of local optimality. Three reasons why some costs may be considered more important for health care decision makers than others are (i) relevance, (ii) equity and (iii) responsibility. suggest that it may be useful to adopt a two-perspective approach as a standard, presenting one cost-effectiveness ratio following a strict health care perspective and one following the common societal perspective. The health care perspective may assist the health care policy maker better in achieving health care goals, while the societal perspective indicates whether the local optimality of the narrow health care perspective is also in line with societal optimality. More research on actual decisions should provide more insight into the relative weights attached to different types of costs.

0445 Providing Informal care, what does it mean? W.Brouwer (Department of Health Policy & Management University Rotterdam, Netherlands) In this presentation, we provide an overview of all relevant concepts related to informal care. The following issues will be presented, both

theoretically as empirically: appropriate measurement of informal care tasks performed, measurement of time invested, opportunities sacrificed due to providing informal care, measurement of the psychological burden of caregiving, identifying and measuring the process utility of providing informal care, the quality of life of the caregiver. The authors present many empirical results for informal caregivers providing care for various patient groups.

0142 The Empirical Relationship between Community Social Capital and the Demand for Cigarettes Brown T. (University of California at Berkeley, USA) Scheffler R.M., Sukyong S., Reed M. We show that community social capital is inversely and strongly related to the number of cigarettes that smokers consume. We do not find community social capital to be related to the decision to participate in smoking for adults. Using a new validated measure of community social capital, the Petris Social Capital Index and three years (1998–2000) of U.S. data on 38,807 adults, we estimate a two-part demand model incorporating the following controls community-level fixed effects, price including excise taxes, family income, a smuggling indicator, nonsmoking regulations, education, marital status, sex, age, and race/ethnicity.

0032 Macroeconomic Conditions, Mortality and Health Buchmueller T. (University of California, Irvine, USA) Grignon M., Jusot F., Perronnin M. In contrast to the conventional belief that health deteriorates during economic downturns, a growing body of research suggests that mortality and morbidity decline when the economy temporarily weakens. Since most studies in this area use data from the U.S., it is not clear whether this relationship holds in other countries or is an artifact of the unique features of the U.S. health care system and social safety net. This paper extends the literature by using aggregate and individual-level data from France to investigate the relationship between macroeconomic conditions and health outcomes. The aggregate analysis is based on data on mortality and unemployment rates from each of France’s 95 départementes for the 1982-2004 period. In addition to examining the relationship between overall mortality and unemployment, we conduct separate analyses by source of mortality. The individual-level data come from the Enquête sur la Santé et la Protection Sociale (ESPS), a survey of the French population that has been conducted biennially since 1988. The ESPS provides detailed information on health status, health behaviors and the utilization of medical care. We relate several health outcomes to local economic conditions, controlling for individual demographic and socioeconomic characteristics, general time trends and area fixed effects.

0174 Health financing in high income countries: lessons for countries in transition Busse R. (Department for Health Care Management, Berlin University of Technology, Berlin, Germany) Schreyögg J. Objectives: The aim of this study is to review evidence on health financing arrangements in high income countries over the last 30 years and to identify and evaluate key policy lessons from this experience for low and middle income countries. Methodology: The presented review is based on a conceptual framework of health financing “functions”. Countries have been grouped for

comparative evaluative purposes into three groups, based on the main mechanism of financing health care. Decisions on coverage and benefit entitlements as well as various functions of health care financing (i.e. collecting, pooling and purchasing) of 25 high income countries are described and analyzed. Results: The pooling function is most essential in order to provide coverage to as many individuals as possible, thus reducing their financial risk being the principal aim of public health financing. Private health insurance, medical savings accounts and other forms of private resource collection can only be supplementary models for increasing universal coverage. The evolution of health financing schemes towards universal coverage was pretty similar in the included countries. Next to the precondition of economic growth, the most essential lessons are to initiate pilots for voluntary health insurance, to foster the ability to administrate, to ensure political commitment to expand population coverage, to combine expansion of population coverage with risk-pooling and to ensure evaluation of products and services at each stage.

0441 What is in the health care basket, who decides and what criteria are taken into account? Reinhard B. (Department for Health Care Management, Berlin University of Technology, Germany) Objectives: The objective is to identify and analyse a framework for health baskets, the taxonomy of benefit catalogues for health care services and the criteria for the in- or exclusion of benefits in nine EU member states (Denmark, France, Germany, Hungary, Italy, the Netherlands, Poland, Spain and United Kingdom). Meth ods: An open ques tion naire was de vel oped, fol low ing the framework of functional categories of the OECD’s “System of Health Accounts”. Researchers in each country were asked to identify the decision-making processes on the benefit basket as a whole in their country and to provide a detailed description of the existing benefit catalogues for each functional category, the actors involved in decision-making and the decision criteria. Results: It turns out that the explicitness of benefit catalogues varies largely between the countries. In the absence of explicitly defined benefit catalogues, in- and outpatient remuneration schemes have de-facto the character of de-facto benefit catalogues. The criteria for the in- or exclusion into benefit catalogues are often not transparent and (cost)effectiveness is only applied for certain sectors. Conclusion: An EU-wide harmonisation of benefit baskets does not seem realistic in short or medium term as the variation of criteria and the taxonomies of benefit catalogues are large but not insurmountable. There may be scope for a European core basket.

0225 Testing linear restrictions in the analysis of data from discrete choice experiments Cairns J. (Health Services Research Unit, Public Health and Policy, London School of Hygiene and Tropical Medicine, London, UK) A common method of analysing discrete choice data is by use of difference models, where the difference in attribute levels between two options are entered as independent variables. The implicit assumption is usually that there is a linear relationship between differences in attribute levels and the strength of preference for one option over another. In cases where the attribute is not measured continuously one analytical option is to measure the differences in levels. This imposes the restriction that the difference between level i and level i+1 has the same significance as the difference between level i+1 and level i+2. This might be a reasonable approximation in some circumstances but particEur J Health Econom Suppl 1 · 2006

| S23

ularly when the attribute is measured ordinally rather than on an interval scale this should be tested rather than assumed. Even for differences measured on an inter val scale the assumption of linearity may not be valid and should be tested. This paper shows how the restriction can be tested and illustrates the implications of alternative assumptions using two datasets, one on preferences over dental restorations, and the other concerning assessments of health risk by divers and other offshore workers.

0173 Medical tourism and its implications for the single market Carrera P. (University of Heidelberg – Medical School, Heidelberg, Germany) Bridges J. Faced with long waiting lists and the high cost of private treatment, patients from developed countries are seeking treatment abroad. This has led to the rise of medical tourism as a growth industry for developing countries and, within the EU, among new EU member states not least because of EU regulations and lower prices. To be sure however, even Germany and France are increasingly providing medical services to European and international patients. Unfortunately, given the paucity of formal literature and conceptual imprecision, better understanding of medical tourism is hindered. We offer a systematic literature review of the term “tourism” in the context of health care using the Medline database. Based on the review, we defined medical tourism as the organized travel outside one’s natural health care jurisdiction for the restoration of the individual’s well-being in mind and body. Building on this definition, we provide an overview of the medical tourism market in Europe, mapping out the direction of trade between healthcare systems particularly the main players in the provision of medical care. Issues such as the push factors driving patients to seek healthcare outside of their systems and the impact of the rise of medical tourism are likewise explored.

0308 The ‘ACE’ Approach to Priority Setting in Australia: Background and Progress To Date Carter R. (University of Melbourne, Australia) Vos T. While the importance and need for priority setting in the health sector is clearly established, the central question of how priority setting is to be achieved remains contested. In this talk the genesis and application of an innovative Australian approach to priority setting (called “ACE” – Assessing Cost Effectiveness) is presented. The starting point for ACE was the development of a checklist on what constitutes an “ideal” approach to priority setting. The checklist incorporates guidance from economic theory, from empirical experience, from ethics and social justice and from the more pragmatic needs of decisionmakers. The “ACE Approach” is innovative in its close attention to technical rigour (in the economic and epidemiological analyses), balanced by a concern for “due process” and inclusion of broader considerations through a second-stage filter analysis. The filters include ‘equity’; ‘acceptability to stakeholders’; and ‘feasibility of implementation’. The ACE approach has now been trialed in four studies in Australia (in cancer; mental health; heart disease; and obesity prevention) funded by government or through the competitive grant process. A further major study has commenced on the prevention of non-communicable disease. In a companion talk, the ACE: Obesity study will be presented in more detail.

S24 |

Eur J Health Econom Suppl 1 · 2006

0359 Explaining income-related inequalities in health care utilisation in England, Scotland and Wales Castelli A. (Centre for Health Economics, York, UK) Decentralisation of the health care sector has been a popular policy reform in the last decade. The effects of this reform are neither thoroughly investigated nor completely understood. On the one hand, decentralisation of the health care sector can jeopardise equity if individuals who are “equal” in every respect but the jurisdiction in which they happen to live are treated differently than they would be in a more centralised structure. On the other hand, however, it may be beneficial to local jurisdictions as these will be able to tailor the production and provision of health care goods and services to local needs and preferences. First, we investigate how the decentralisation of the health care sector affects equity in health care utilisation within a decentralised system, using the constituent countries of Great Britain as a case study. Data are taken from the General Household Survey from 1995/96. The utilisation of primary and secondary care is proxied by GP visits, outpatient visits and inpatient stays. Secondly, we analyse the determinants of income-related inequalities in health care use observed in England, Scotland and Wales before the devolution reform and how these have been affected by the reform itself through a decomposition approach.

0085 A new approach to measuring health system outputs and productivity Adrianna C. (Centre for Health Economics, York, UKnited Kingdom) Dawson D., Gravelle H., O’Mahony M., Street A., Weale M., Jacobs R., Martin S., Stevens P., Stokes L. The productivity of the health system should capture the valuable things that the health system produces. This is not straightforward because of difficulties in defining health care outputs and of attaching values to the outputs. We propose two new methods for measuring output growth. An ‘ideal’ value weighted output index has two fundamental features. First, the value attached to each output reflects its relative contribution to health outcomes and, second, the value of other important characteristics of health care, such as the process of care delivery, can be readily incorporated. Comprehensive calculation of this index is currently unfeasible because of a lack of health outcome data. In the meantime an ‘interim’ cost weighted output index is proposed, which incorporates quality adjustments. We calculate the ‘ideal’ index for a limited set of ‘specimen’ activities, for which data on health outcomes are available, and the ‘interim’ index for the English NHS from 1999/99 to 2003/04. This interim index is combined with a measure of input growth to assess total factor productivity growth over the period.

0409 The determinants of regional health expenditure in a decentralized system Cavalieri M. (University of Catania, Department of Economics and Quantitative Methods (DEMQ), Catania, Italy) Guccio C. In latest years there has been a growing interest on factors explaining health expenditure at a regional level. The existing literature on the determinants of health expenditure has followed two different approaches. Pioneering studies have focused on cross country health expenditure analyses but were not able to appropriately deal with the problem of country hetereogeneity. Most recently, few papers have restricted the analysis to single countries with either a federal system (e.g. Canada and Switzerland) or multiple higly autonomous jurisdictions (e.g. Spain). In

the case of Italy, even though the health system was initially designed as a centralized one, the government has progressively undertaken a decentralization process up to the point of modifying the Constitution in order to appoint regions with exclusive health care responsibilities. In this paper, we analize data on regional per capita health care expenditure (in real terms) for the period 1990-2003. The objective is to determine the effects of different variables and to evaluate the resulting implications for the design of a financial perequation model to be used in a future federal system.

0518 How Equitable is the Financing of Health Care? The Case of Malaysia Chai P.Y. (University of Nottingham, UK) Whynes D.K., Sach T.H. Introduction: Equity is one of the criterions for assessing the performance of health financing systems. As a tax-financed country, the Malaysian government contribute substantially towards financing the health care services. Even though both public and private sectors contribute to the system, ultimately, households bear the financial costs of health care services. Objective: This study aims to comprehensively assess the equity of health care financing in Malaysia. It focuses on households’ payment for health care, in order to determine the distribution of financial burden across households with different ability to pay. Methods: The data used was the Household Expenditure Sur vey Malaysia 1998/99. Four approaches are employed, i. The proportion approach, ii. The tabulation approach, iii. Concentration curves and iv. The Kakwani’s progressivity index. Results: Results showed that Malaysia’s health financing system was slightly progressive, with a Kakwani’s progressivity index of 0.0897. Such results are the offsetting effect from, i. Mildly progressive out-ofpocket payments (0.0093), ii. Progressive private insurance (0.2293), iii. Progressive contributions to EPF and SOCSO (0.1194), iv. Regressive indirect taxes (-0.1863), and v. Progressive direct taxes (0.3827). Conclusions: The Malaysian experience has shown that the parallel existence of public and private health care services and individual’s health seeking behaviour produce a mildly progressive financing system.

0081 Fiscal sustainability and the health financing systems of South-Eastern Europe

cal sustainability concerns, (v) pharmaceutical expenditures are a key cost driver. Based on these findings, the study presents policy options for reforms.

0428 Development Pharmacy-based Risk Adjustment Model Chang R.E. (Graduate Institute of Health Care Organization Administration, College of Public Health, National Taiwan University, Taipei, Taiwan) Lai C.L., Hsieh M.F. Diagnosis-based adjusters and prescribed drugs adjusters have attracted the research attention for their rich clinical message. Reviewing medical expense structure in Taiwan, one-third of outpatient medical expense is spent on prescription drugs, and chronic disease prescriptions account for the main portion in drug expenses. This study intends to refine diagnostic risk adjusters with prescription information to improve the predictability of risk adjustment models in Taiwan. With detailed enrollment data, medical expense data, and diagnostic and prescription data of contracted institution in 2000 and 2001, the predictabilities of various risk adjusters were analyzed. The results show that more clinical information improves the predictability. Using prescription information to refine diagnosis-based risk adjusters by splitting it into refined cost groups can improve the risk adjustment model not only in r-squares but also predictive ratio of specific chronic disease group, which can be of help to reduce risk selection.

0586 Fiscal Sustainability in Health in New EU Member States Mukesh C. (World Bank) While GDP growth in new member states is robust, the health sectors continue to roll up debt. Against this backdrop, countries are also facing aging populations, growth in medical technologies, and rising consumer expectations. The paper examines trends and looks at impacts of various options to make expenditures more sustainable.

0421 Higher Health state and health consumption for 50+ VPH insured in Europe?

Chakraborty S. (The World Bank, Washington DC, USA) Boskovic T.

Chevrou-Séverac H. (IEMS (Institute of Health Economics and Management), University of Lausanne, Switzerland) Lufkin T., Huguenin J.

The countries of South-eastern Europe (SEE) (Bosnia and Herzegovina, Croatia, FYR Macedonia and Serbia and Montenegro) are looking towards EU accession. Fiscal discipline and expenditure management are key issues. Public spending on health is already high in these countries. The SEE countries face the prospect of aging populations and growing demand for better quality health ser vices. These factors are likely to fuel the growth of public spending on health. This means that SEE countries have to adopt health financing strategies that create fiscal sustainability while promoting quality and access. The study analyzes the underlying determinants of the patterns of health financing and expenditures in SEE countries and implications for reforms. The study is based on an analysis of secondary data (WHO Health for All database, World Development Indicators, government data). The findings are: (i) there is little scope for increasing public spending on health (ii) public resources are not focused on cost-effective health interventions and public goods and merit goods, (iii) health care institutions have low productivity, (iv) in a context where unemployment is high and there is tax evasion, depending on payroll taxes is creating fis-

Background: In almost all European countries health coverage or health insurance is compulsory for almost all people, irrespective of the type of health insurance scheme (Beveridgian, Bismarckian or private compulsory). As defined by OECD (2004), VPHI could be of three categories: duplicate, supplementary or complementary. In the SHARE countries, the VPHI offers are directly linked to the health baseline system. Aim: We focus on the impact on health status of having a VPHI, compared to the baseline insurance scheme among people aged 50 years old or more, from the Survey on Health, Aging and Retirement in Europe (SHARE).. Given the diversity of health systems described above and given that the SHARE dataset includes questions on the types of health care (HC) insured by VPHI, it is possible to specify, by country, the different VPHI coverage mostly chosen by respondents. We then analyse the impact of having a VPHI on individual health status in each country. We expect that this effect transits also through higher health care consumption. Do differences in health care consumption between people having a VPHI and people not having one exist? Do these differences have an impact on health status? Eur J Health Econom Suppl 1 · 2006

| S25

However, in the case of VPHI, we should first take adverse selection into account. It is depends on expected health status and HC consumption. Moreover, health states and health consumption are linked, and having a VPHI could affect health care utilization and health status. Knowing that health care consumption and health status could also be simultaneously determined and/or endogenous, we analyse the impact of VPHI on health status thanks to a switching model with simultaneous equations. Methodology: We tackle the adverse selection issue with a Roy switching model (Maddala, 1983), with endogenous switching. The choice of having a VPHI is generally made at younger ages. Henceforth, this decision depends on past expected health state and/or health consumption, information which is not observed today. This issue could be circumvented in the case of longitudinal data. Instead of this information, we only observe a realization of this latent variable in the 1st wave of the SHARE data: the fact of being privately insured today. The VPHI participation criterion will determine if the observed outcome is the one in the regime with VPHI, or in the regime without. Notice that, in our specification, the outcome is composed of two variables: HC consumption and health status. Moreover, self-perceived health is also a latent variable. The response depends on a personal appraisal of one’s own health which is not observable. Henceforth, the Roy model is fully appropriate. Firstly, the VPHI participation equation is estimated by a probit model. Secondly, we estimate two simultaneous equation systems. Both systems are composed of one linear equation for health care consumption and one probit specification for self-perceived health (SPH). The estimation of each system is made through full information maximum likelihood (FIML) estimation (Huguenin, 2004). This estimation method allows to test firstly for endogeneity between variables and, secondly, for simultaneity. Preliminary results: Up to now, we have focused on three countries: Austria, Denmark and Italy. The first one has a Bismarckian health system and the last two have a Beveridgian one. Generally, because of the cross-sectional feature of the SHARE dataset, estimating adverse selection effect is challenging. Moreover, in the case of VPHI, insurers can deny or limit coverage because of pre-existing health conditions. In Denmark, people are mostly privately insured for direct and wider access to GP and specialists. First of all, in Denmark, there is no simultaneity between HC consumption and SPH. Secondly, SPH doesn’t have any impact on HC consumption for VPH insured and non VPH insured. HC consumption is mainly determined by objective health measures. Contrary to our expectation, HC consumption affects SPH only for non VPH insured people. In Austria, people are mainly privately insured for extended or full hospital coverage. This is confirmed by our probit estimation of VPHI participation. Here again, we do not detect simultaneity between HC consumption and SPH. In both regimes, HC consumption is highly significant in explaining SPH. It is also the case in both regimes for the impact of objective health measures on HC consumption and SPH. Here again, SPH does not explain HC consumption. In Italy, the reasons to take a VPHI are more mixed. It can be for duplicate care, supplementary or complementary coverage. Moreover, people can have an individual or a family contract, and be insured through their employer or not. Furthermore the OECD figures for private insurance are much higher than in the SHARE sample. It seems plausible that insurers refuse to insure elderly people due to risk selection. In both regimes, SPH is not affected by HC consumption. For non privately insured people, socio-economic status and chronic diseases influence HC consumption. For privately insured 50+ Italians, our specification is less powerful in explaining HC consumption. Our model probably suffers from the fact that we were forced to pool the different reasons for taking a VPHI. Extensions: At this point, we are in the process of pursuing the analysis for the remaining SHARE countries. Moreover, we plan to extend the model to take HC quality and access into account. We also wish to analyse the European countries conjunctly, taking into account the differences in baseline health system and also in private health insurance.

S26 |

Eur J Health Econom Suppl 1 · 2006

Such a global approach is relevant in trying to determine if there are common features across Europe.

0547 Is there a need for centers of excellence in Greece? Christodoulou I. (Medical School, Democritus University of Thrace, Thessaloniki, Greece) Pogonidis Chr., Xenodoxidou E., Drimala AIM of this study is to examine the opinion of physicians about the need for Centers of Medical Excellence in Greece. Material and methods: A questionnaire was distributed in 85 physicians, 35 specialists and 50 trainees. Answers were taken from 9 different hospitals in Greece. Results: Many physicians (69%) believe that surgical operations of advanced difficulty can be executed in any Hospital since “good name of surgeons and of Surgical Departments” depends on them (25%). In a percentage of 33%, doctors consider that therapy in a Center of Excellence is often not needed, while the rest believe that only too old patients and patients with advanced malignancies might lose the chance of advanced medical therapy.Current therapeutic interventions held in Centers of Excellence are mostly “overtherapy” according to the 65%. In total, physicians are not opposite to the idea of national institutions of Excellence (100%).Almost all trainees but only half of the specialists wish to work in a Center of Excellence, even circumstantially. Conclusions: Centers of Excellence offer advanced health ser vices to the public but have a special role in the National System of Health. If National Health Institutes of Excellence are established in Greece, they would attract many physicians, especially the younger ones.

0070 Economic Resources and Child Health Chung W. (National University of Singapore, Singapore) We use data from the Indonesian Family Life Sur vey (IFLS) to examine the relationship between child health and household economic resources, and find that economic resources as measured by equivalized expenditure have a statistically significant positive effect on child health. We make additional use of the data (where extensive data on children’s nutrition, household sanitation, the utilization of medical care, and family health status are available) to assess the mechanisms through which economic resources may affect child health. We find evidence that economic resources have a sizable and significant effect on these potential intermediary factors, and that they partly explain the protective role of economic resources in child health.

0412 Explanatory factors of the inequalities in the utilisation of health care in Spain Clavero A. (University of Málaga. Applied Economics (Statistics and Econometrics), Málaga, Espanol) González-Alvarez, Luz M. This investigation tries to decompose the degree of inequality in the health care utilisation into the contributions of some explanatory factors by Oaxaca approach. Due to the different characteristics of the visits to general practitioner and to specialists is required a separate analysis for each medical care.European Community Household Panel (ECHP) has been the selected information source because it contains information about the utilisation of the aforesaid ser vices and about the general health situation and socio-economic characteristics. Other advantages of the ECHP are the monitoring of a given sample of indi-

viduals over time, and thus permit to measure the impact of the previous use on current use. Oaxaca decomposition focuses on the gap in the utilisation between two groups public and public-private insurance coverage. This gap can be attributable to differences in the explanatory factors and to differences in the effects of each factor on the health care utilisation of the two groups. The results suggest that there are not greater gaps in GP visits, but the gap in the probability of a visit to specialist is explained by the differences in income and education and the elasticity of utilisation respect to the previous use.

0004 Pharmaceutical expenditure, Health Care Expenditure and GDP Clemente J. (Zaragoza University, Zaragoza, Spain) Marcuello C., Montañes A. ABSTRACT expenditure is one of the main components of Health Care Expenditure. In 1998, pharmaceutical expenditure made up about 13% of total Health Care Expenditure in Sweden and 11% in the United States. The evolution of pharmaceutical expenditure in relation to Gross Domestic Product has markedly increased since the 1980, although some countries have recently managed to reverse this tendency. The evolution of Health Care Expenditure has been analysed in numerous studies and results have confirmed the existence of a long-term relationship between Health Care Expenditure and Gross Domestic Product (GDP). Nevertheless, this relationship is not a stable one as there is no commonly applied international norm and differing income-elasticities by groups of countries can be observed (Clemente et al. 2004). In this study, we analyse the long-term relationship between aggregate Health Care Expenditure, Gross Domestic Product and the effect of pharmaceutical expenditure on that relationship.

0311 Do peers lifestyles influence obesity in adolescents? Clímaco M.I. (ISCAC-Coimbra Polytechnic Institute, Coimbra, Portugal) Barros P.P., Lourenço Ó. Framework: Obesity is an important problem of public health, which is often initiated by unhealthy behaviours during adolescence. Food habits among young people are generally influenced by their peers, the so-called “peer pressure”, as well as other hazardousconsumptions such as alcohol, tobacco and drugs. In this work, obesity is viewed as a household produced good, and self-image and social interactions are likely to play a key role in determining adolescent weight. Objectives: The main objective of this paper is to identify the risk factors which may explain obesity/overweight among the group of adolescents, according to the Body Mass-for-age, a medical criterion alternative to the BMI adapted to adolescents. Moreover, it aims to draw attention to the dimension of the problem in a risk group composed of potential future obese adults, with all the associated health costs. Particular attention is paid to the potential influence of the peer pressure effect on obesity, since consumption decisions are affected by age peers. Methodology and data: A recursive simultaneous equation model, involving two binary choice variables, has been developed, incorporating the importance of peer effects and to control its potential endogenous nature. Control variables include age, gender, school performance, family cohesion, friendship cohesion, social and economical family level, food habits, sports practice, and consumption of addictive substances. The data used are from a Portuguese survey on the health of adolescents made in 2002. The universe of this survey is composed of students in the 6th, 8th and 10th years of secondary education, with ages between 11 and 25 years old. Conclusions: The key finding is that peer effects play a significant role in the probability of obesity. Dieting and corporal self-image are also important determinants of obesity.

0538 Casemix in Belgium Closon M.C. (University of Belgium) APRDRGs are actually used for the financing of the day of stay for all acute hospitals (45% of hospital financing) in Belgium and are progressively used in the financing of the medical services and drugs. Prospective payment is based on mean length of stay or lump sum by admission according to APRDRGs and levels of severity. A research has been made on the impact of the implementation of a prospective payment of the days of stay in 1995 in Belgium on a data base including all acute hospitals from 1991 to 1998.In summary, there were a small decrease in length of stay, a large increase in severity, a sharp increase in readmission, an increase in growth of total expenditures, a slowdown in the decline of mortality rates, a risk of selection and of “by pass” of the reform. It can favour adverse selection and impair accessibility to care if some factors (social, severity...) are not taken into account. It should be made clear that such reform needs to be as global as possible to avoid cost shifting and adverse selection. As it is often the case in the health care sector, introduction of efficiency incentives has to be hardly public regulated in order not to produce adverse effect. It is impossible to take into account all factors having an impact on the treatment and the costs of each patient. This is the reason why APRDRGs are not a good tool for a financing by patient or admission but is only a good tool for the calibration of a global budget for the hospital. This allows a better share of risk across patients and across APRDRGs. Mix financing system (partial prospective and partial retrospective) is another way to share risk between financing authorities and hospitals: retrospective special payment for outliers, financing for specific functions (emergency care, research, teaching...), exclusion of some pathologies or expenditures from the prospective payment system (psychiatry, new drugs...) can be used. In Belgium, thanks to the fee for services system, it is also possible to share by act or by drugs the risk between financing authorities and hospital. It is already done for clinical biology and will be done for drugs, prospective payment according to APDRG can be applied only on 50% or 70% of the fees. The complementary percentage continues to be financed by act or drugs. This system allows maintaining the information system linked to resource really used that can be very useful to evaluate quality of care, cost shifting.

0439 The Effects of Individual and Child Illness on Work Absence Decisions Coomer N. (North Carolina State University, Raleigh, USA) This paper examines the effects of personal and child illness on absence decisions of parents. Absenteeism is shown in the literature to have significant effects for both firms and households. Losses in work time lead to costs in the form of decreases in productivity, increases in labor costs and decreases in household income. This research explores the absenteeism decisions of employed parents in an effort to better understand behavior that may contribute to absenteeism. The theoretical framework models the decision to miss work when either the individual is ill, the child is ill or both are ill. The availability of child care and sick leave provisions impact the individual’s absence decision. Structural estimation will allow for the introduction of new policies relating to absenteeism and their evaluation.

Eur J Health Econom Suppl 1 · 2006

| S27

0045

0039

The Effect of Taxes and Bans on Passive Smoking

Prudence and health care risks

Cornaglia F. (University College London, UK) Adda J.

Courbage C. (The Geneva Association, Geneva, Switzerland)

This paper evaluates the effect of excise taxes and bans on smoking in public places on the exposure to tobacco smoke of non-smokers. We directly quantify passive smoking using data on cotinine levels- a metabolite of nicotine- measured in non smokers. Exploiting state and time variation across US states, we reach two important conclusions. First, excise taxes have a significant effect on passive smoking. Second, smoking bans have on average no effects on non smokers, because we show that bans can have contrasting effects. In particular, while bans in public transportation or in schools decrease the exposure of non smokers, bans in recreational public places can in fact per versely increase their exposure by displacing smokers to private places where they contaminate non smokers.

0449 A comparative analysis of pricing and reimbursement for in-patent drugs in 7 EU countries Cornago D. (Mario Negri Institute, Milan, Italy) Compadri P., Garattini L. This comparative study analyses pricing and reimbursement policies for in-patent drugs in seven European countries (Belgium, France, Germany, Italy, the Netherlands, Spain, and the United Kingdom), conducted according to a common methodology in each country. General institutional framework, market access, pricing, reimbursement and cost containment measures adopted by each domestic market were considered. The analysis implied two phases (i) reviewing the literature on the subject in national and international journals, and (ii) interviewing a selected expert panel of key players in each country (composed of at least one civil servant, one pharmaceutical manager, and one health economist). The analysis showed a general trend to introduce short-term measures to contain pharmaceutical expenditures, which seem to be hard to control in almost all countries analysed. However, we still found a wide heterogeneity in the general approaches adopted by each country for pricing and reimbursement procedures, probably because of domestic and historical issues.

Prudence is a common assumption in risk and insurance economics. Prudence can be defined either through a precautionary saving motive (Kimball, 1990) or in terms of preference towards risks (Eeckhoudt and Schlesinger, 2005). Prudence has been established in4

0123 Factors influencing the evolution of new drugs prices between 2000 and 2005 in Portugal Crisostomo S. (Centre for Pharmacoepidemiologic Research, Lisboa, Portugal) Zilda M. Objectives: To characterize drugs with prices first approved in 2000; to analyse prices evolution till 2005 and to investigate its relation with some characteristics of the price regulation mechanisms in place in Portugal. Methods: Data on new drugs (reimbursed and non-generic) and their prices were collected from official sources. Due to skewed distribution of relative price changes nonparametric statistics were used. To investigate the effect of other variables (and its interactions) on relative price changes the general linear model (GLM) was used. Results: For the 528 drugs under analysis, lower relative price changes, between 2000 and 2005, were found for drugs with provisional price approvals, drugs included in the reimbursement system, drugs with more recent reimbursement approval and drugs reimbursed under special reimbursement schemes (all p<0.05). The model accounts for 77.5% of the variance in the relative price changes, with price approval (provisional/definitive) having the greatest influence (p<0.05). Conclusion: Current price regulation mechanisms seem to have contributed to control drugs prices. In particular, provisional prices show a significantly greater reduction compared to non-provisional prices, suggesting current mechanisms work to correct provisional prices in the post-approval period, according to the corresponding prices in the reference countries (Spain, France and Italy).

0241 Health Economics & Health Policy

0406

Cromwell J. (Research Triangle Institute, Waltham, Massachusetts, USA)

Generic market penetration and cross-country competition in EU-5, Canada and the United States pharmaceuticals market

The U.S. Congress enacted a moratorium on physician-owned hospitals specializing in cardiac or minor surgery on the grounds of unfair competition. Specialty hospital growth exploded in the U.S. in response to payer overpayments by the federal government. The moratorium enacted by Congress, however, was found to be an inefficient, untargeted solution by unnecessarily limiting competition. Arguments against physician ownership were found to be anti-competitive and relied on criticisms previously lodged against for-profit acute hospitals; namely, avoiding severely ill patients and the uninsured. Physician ownership percentages were generally small and not drivers of lucrative referrals. Congress rescinded the anti-competitive moratorium in favor of more fundamental price corrections based, in part, on our recommendations. The U.S. government is aggressively implementing many pay-for-performance demonstrations that tie payments to health care improvements. Many different payment schemes, definitions of outcomes, and evaluation designs are being tested. Health economists, like ourselves, have been designing experiments and evaluating provider and patient responses to new financial incentives tied to payments. Our study reviews the strategies, strengths, and weaknesses behind the different approaches. Payfor-performance is the Holy Grail of both health economists and policy makers, but may be technically impossible to achieve.

Costa Font J. (London School of Economics, Health and Social Care, London, UK) Kanavos P. Little evidence is known on the determinants of generics penetration in countries subject to significant market regulation. This paper empirically examines the extent the determinants of generic entry in different European and North American countries using IMS data from 20002005. On the other hand, we scrutinise the extent to which generics penetration explains cross-country price differences in different countries. Interestingly we find that country size and country specific fixed effects explain most significant factors in explaining market penetration rather than cross-country price differences.

S28 |

Eur J Health Econom Suppl 1 · 2006

0325 Survival Models For Cost Data Dario G. (Department of Public Health and Microbiology, University of Turin, Torino, Italy) Pagano E., Petrinco M., Desideri A., Bigi R., Gregori D. Survival models fail in accounting for non proportionality in the costaccumulation process. In case of no censoring, the log-gamma GLM model seems to be a robust alternative (Basu, et al. 2004). We propose the additive Aalen model as a valid alternative, allowing to obtain a reasonable fit of the cost-accumulation process, maintaining the additivity of the effects, and obtaining additional information as compared to other models.Gamma regression with log link and the Aalen additive model were fitted to a series of Monte Carlo data simulations (four cost distributions: two Lognormal and two Gamma; seven sample sizes; 1000 Monte Carlo experiments). Then, the Aalen model has been applied to the data of the COSTAMI trial.bias for the Aalen model is about 10% for the log-normal generated data and almost negligible for the Gamma case, showing a conser vative coverage also for small samples. the COSTAMI case, the Aalen model allowed the evaluation of each covariate effect also at extreme observed costs.a non-censoring framework the Aalen additive model performs well and can be a reasonable alternative to the standard Gamma regression models. Still, it offers additional information about the relationships between costs and specific covariates.

0383 Measuring relatives’ QALY-weights for usage in health economic evaluations Davidson T. (Linköping University -Department of Health and Society, Linköping, Sweden) Krevers B., Levin L.Å. Aims of the study: The objective of this study was to estimate the value of the effects on relatives due to a family member’s illness for usage in health economic evaluations. Methods: Data used in this study is from the Swedish part of the EUROFAMCARE study, in where relatives caring for an elder at least 4 h per week have been inter viewed in six European countries. We introduce a new term, R-QALY-weights, which is the influence on the relatives’ QALY-weights due to a family member’s health state. We created RQALY-weights in two ways. First by subtracting a population based QALY-weight from the QALY-weight for the relative, created from EQ5D. In the second method of creating R-QALY-weights we asked the respondents to assess their answers in EQ-5D assuming that the elder’s health was so good that he or she would not be in need of care. Results: The results indicate that R-QALY-weights occur when the respondents assessed their own R-QALY-weights but not for the whole sample when we compared with the general population based weights. However, also in the latter case, R-QALY-weights could be found in two subgroups: women below 80 years old and caregivers whose situation is described as burdensome.

0583 Health in Argentina: diagnosis and risks in relation to MDOs? Puente C., (Ministry of Health, Argentina) The following up and the analysis of the Millenium Develoment Objectives in a complete perspective has directed our country to institutionalize the various actions and the isolated initiatives under the control of an official institution. This innovated task has officially started in 2004, by the named Social Policy Coordination National Council, as

the institutional coordinator which includes a team work formed by technicians of all Ministries: Education, Technology and Science; Social Development; Production and Economy; Federal Planning; Ser vices and Public Investment; Environment and Health; Work, Employment and Social Security; and the Census and Statistics National Institute. Specifically, the followin up of the more directed objectives related to Health is under the responsability of the Ministry of Environment and Health. The objective of this presentation refers to how the following up of MDOs in health field in Argentina is working and what the main problems which should be dealt to achieve the goals established are. Argentina has incorporated in its international commitment with the MDOs certain objectives, goals and additional indicators to adapt to the country`s necessity and reality. In terms of health, some goals and indicators were included to the following up of the present objectives. This presentation shows the present results of some indicators, the changes of these objectives and goals in actions and programs of concrete execution, and the challenges for the future in achieving the MDOs.

0233 Is a catch-up hepatitis B vaccination campaign in high-risk adults cost-effective? De Wit G.A. (Nat. Inst. of Public Health, Centre for Prevention and Health Services Research, Bilthoven, The Netherlands) Kretzschmar M., Mangen M.-J.J. Background: Current hepatitis B (HBV) prevention policies in the Netherlands include screening of pregnant women and vaccination of children from endemic populations. In addition, a 4 year vaccination campaign in high-risk adults (sexual risk groups and hard-drug users) was initiated in 2002. Aim: To determine the cost-effectiveness of this catch-up vaccination campaign in high-risk adults at national level. Methods: A dynamic model to estimate future incidence of HBV in the population was developed. This model describes horizontal, vertical and sexual transmission of HBV. Also, an economic model estimating direct health care costs, life years lost and quality of life consequences of HBV infection was constructed. Actual campaign data on number of persons reached, compliance and program costs, including outreach activities, were used. Results: It is estimated that this 4-year vaccination campaign results in 3000–5000 infections prevented in the Netherlands. Cost per person receiving at least one vaccination are € 170. In the base-case analysis, cost per lifeyear gained and cost per QALY are € 9,400 and € 7,600, respectively. Conclusion: This 4 year catch-up campaign was cost-effective at a national level. The cost-effectiveness of continuation of the programme depends heavily on the amount and cost of (more expensive) outreach activities.

0330 Impact of pharmaceutical firms strategies on French physicians: antidepressants prescription Desquins B. (IEMS, Lausanne,) Protopopescu C., Decollogny A., Holly A., Rochaix L. Rationalesome countries such as France, physicians’ information regarding the efficacy of new drugs is mainly provided by pharmaceutical companies who invest heavily in promotional campaigns. There is a growing concern that biased information (not true) diffused by these campaigns affect GPs’ choice of drug. To analyse such potential biases, we focus on the prescription of antidepressants for two reasons. The first is related to the fact that in this particular market, firms have engaged in heavy product differentiation strategies, while the actual therapeutic efficiency of these various antidepressants is judged comparable by medical experts. The second is related to the important uncertainty that surrounds depression diagnosis and how it may affect physicians’ choice of drug.this paper, Eur J Health Econom Suppl 1 · 2006

| S29

we examine the impact of the pharmaceutical firms’ promotional campaigns and GPs’ risk aversion on antidepressants’ initialisations. use an attribute model and entropy index to evaluate these impacts.data covers antidepressant initialisations by 500 GPs practising in the PACA region in France.and Conclusionsto this paper, GP’s aren’t patients’ perfect agents since they can’t dissociate the true of biased information. This obser vation is particularly true for physicians who have risk aversion. This paper highlights also the impact of promotional spending engaged by drug companies on French GPs antidepressants’ prescription.

0209 Modelling waiting times for elective surgery Dimakou S. (City University, London, United Kingdom) Parkin D., Devlin N., Appleby J. Waiting is an important means by which health care is rationed in the UK NHS. Waiting times for elective surgery are a key policy and political concern. The principal policy response has been to introduce waiting time targets against which performance is measured and rewarded.waiting times fall, interest has grown in questions such as how have incentives associated with targets affected managerial and clinical behaviour? How are targets met: and how have they affected the distributions of waiting times? Are they met by increased throughput or by improved management practice? Are waiting times cut down as a whole, or do long waits replace short waits? And more importantly, what factors determine first the decision to place a patient on list for surgery and second the threshold of patients on a waiting list at any given time? We will investigate these questions using time-to-event analysis applied to NHS waiting time data.joining a waiting list is not a stochastic event, this paper develops a model of waiting lists where they are regarded as under the control of hospitals and clinicians and may be manipulated to respond to internal and external supply conditions including hospital capacity and performance requirements.

0038 Health Financing in Brazil in the context of macro-economic adjustment Domínguez Ugá M.A. (Escola Nacional de Saude Pública, FIOCRU, Rio de Janeiro, Brazil) This paper analyses health care financing in Brazil considering two different contexts)The constitutional reform of 1988, which created the new Brazilian Social Security System (which includes Health Sector) and structured a financing system according to its principles;) The context of macroeconomic adjustment, initiated in the nineties, which introduced many distortions to the original financing model, and resulted in the need to create a specific tax to finance the Brazilian health system. final section of this paper deals with equity aspects of health financing in Brazil, considering its distribution between public and private sector, the composition of public financing (considering the direct and indirect taxes) and the distribution of its burden between the different income classes.presentation results from a research developed by the author at the National School of Public Health (Escola Nacional de Saúde Pública / FIOCRUZ), concluded on March 2005.

0117 Health care to irregular immigrants in the Italian NHS Donia Sofio A. (University of Rome Tor Vergata, Roma, Italy) De Gaetano A. Irregular immigration is a wide range phenomenon especially in a country like Italy that has always been both a longed for goal for job seekers

S30 |

Eur J Health Econom Suppl 1 · 2006

and a transit ground to other countries.if is impossible to monitor a phenomenon which is out of control by definition, a comparison between the values related to financing health care for irregular immigrants and the number of regularizations following up in these years induces some considerations. It appears that between the year prior to each measure and the following year, the need of resources for financing provision of health care to immigrants varies remarkably owing to the decrease in illegal immigrants. On the contrary, observing the amount of funding over several years a constant trend is evident. This evidence leads us to deduce that the above funding is not established on the basis of real data necessary to adequate planning and allocation of resources. The result emerging from this analysis is that an effective information system continually updated becomes a necessary instrument to help politicians to allocate resources properly, such as to grant health protection to this special category of foreigners, while on the other hand preventing waste of money due to partial knowledge of the phenomenon “illegal immigrants”.

0127 Does Cost-Sharing Affect Compliance? Dor A. (Case Western Reserve University, Cleveland, Ohio, USA) Encinosa W. This paper departs from previous work on insurance and utilization of prescription drugs in two important respects. First, borrowing for the medical literature, we focus on patient compliance (adherence to a prescription drug regimen) rather than expenditures on prescription drugs, more commonly found in economic studies. This allows us to identify unambiguously behavioral responses to cost sharing (non-compliance), thereby avoiding the difficulty of separating price and quantity in expenditure measures. Second, rather than considering only levels of cost sharing, we go further by considering theoretically and empirically incentives in two main cost-sharing regimes Flat copayment, namely a fixed dollar amount paid out-of-pocket, and “variable” copayment, as percent of full price (we refer to this as coinsurance). We develop a simple model that differentiates between the two cost-sharing regimes, using properties of the mean-variance utility function. Comparative statics show that cost-sharing reduces compliance. More importantly, we show that patient compliance should be lower under coinsurance than under copayments due to uncertainty in cost-sharing. In addition, we derive explicit functional forms for the compliance decision that differ between the two regimes and can be estimated separately, using a large private insurance claims database, we track patients over a one year period (2000) to see whether they complied continuously, intermittently, or not at all to a prescribed regimen of medications. We focus on the example of diabetes (Type II), a common chronic condition that requires permanent and continuous use of insulin-enhancing medications, and disease leads to severe complications when inappropriately treated.estimation sample consists of 20,494 individuals in the copayment regime, and 6,563 individual in the coinsurance regimes. Together, these individuals belonged to 11 firms, with a total of 30 pharmacy benefit plans. Within each insurance regime we estimate ordered logit models for full compliance, partial compliance, and non-compliance with firm, plan, and drug fixed effects elasticities cannot be obtained directly from such discrete-choice models, we performed a number of simulations to compare cost-sharing effects in the two models. Results are consistent with predications generated by the theoretical model. In the coinsurance model, an increase in the coinsurance rate from 20% to 75% resulted in the share of persons who never comply to increase by 9.9%, and reduced the share of fully compliant persons by 24.6%. In the copayment model, a comparable increase in the copayment from $6 to $10 resulted in a 6.2% increase in the share of non-compliers, and a concomitant 9% reduction in the share of full compliers. Similar results hold when the level of cost-sharing is held constant across regimes.non-compliance reduces expenditures on prescription drugs it may also lead to increas-

es in indirect medical costs due to avertable complications (spillover effects). Using available aggregate estimates of the cost of diabetic complications, we calculate that the net effect of a $6-$10 increase in copayment would have the direct effect of reducing national drug spending for diabetes by about $250 million. (We will also present preliminary estimates of indirect costs obtained from linking our data to hospital records for patients in the sample). For now, the results suggest that both private payers and public payers may be able to improve compliance and reduce overall medical costs by switching from coinsurance to copayments in prescription drug plans. Interestingly, the soon-to-beimplemented Medicare prescription drug benefit is based on coinsurance (excluding the so-called doughnut hole), whereas ctive copayment method, suggesting room for budget-neutral improvements. In comparison, flat copays are the nom in most EU countries.

0269 A Cross-National Study of Incentives and Financing in Technology-Intensive Modalities of Care Dor A. (Case Western Reserve University, Cleveland, Ohio, USA) Pauly M. Chronic Kidney Disease (CKD) is a debilitating medical condition leading to organ failure, requiring intensive and costly treatments such as dialysis and transplantation. Due to factors such as the aging of the population and rising obesity rates there has been a alarming increase in the in the prevalence and incidence of CKD in all high-income countries. While payment incentives to providers have gone through a phase of experimentation and change in recent years, it is not clear how different approaches in various countries impacted outcomes and the delivery of care. To address this, we conducted the International Study of HealthCare Organization and Financing (ISHCOF).Under ISHCOF, a series of surveys were administered to country investigators in all DOPPS (Dialysis Outcomes and Practice Patterns Study) countries in 2004 and 2005 on various aspects of the CKD financing. Countries included are Australia, Belgium, Canada, France, Germany, Italy, Japan, New Zealand, Spain, Sweden, U.K, and USA.preliminary analysis suggests a remarkable degree of similarity between ISHCOF countries in the organization of CKD programs, with a mix of private and public providers, social insurance, and low copays for patients. In all countries government is the ultimate payer (including the U.S., under Medicare), and in several countries CKD is administered separately from the rest of the system. However, there are significant variations in the way payments to providers are determined, ranging from fee-for-service at governmentset prices, to negotiated prices under global budgets, to full capitation. Significant variations exist in the cost and total payment of hemodialysis (the main modality to treat CKD) and in physician incomes, particularly for Nephrologists (CKD specialists) even within the European zone. sample sizes do not allow for multivariate analysis, bivariate analysis suggests that nephrologist incomes and the cost of dialysis are dominant factors in explaining CKD expenditures per capita (R2 = 0.11 and 0.18 respectively); primary-care physician incomes and the share of patients with transplantation have no significant effect. Other factors such as “intensity” of dialysis (dosage) are fairly constant across countries. Payments and costs for dialysis payments are moderately correlated with payments matching costs in the U.S., Germany, and Belgium, exceeding costs in Spain and New Zealand, and below costs in Italy and Australia. We do not characterize the pricing in the latter countries as ‘inefficient’ since it can be shown that optimal social prices for maintenance ser vices must incorporate the net gain from averted complications. like per capita expenditures, mortality rates from CKD vary substantially across countries, ranging from 19% in the U.S., to 8% in Japan, with the median found in Italy at 11%. We are currently in the process of obtaining age-adjusted and risk-adjusted mortality rates per country, which will allow us to examine the impact of payment rules system ‘generosity’ and other system variables on measurable outcomes. These sets of results will be added to our presentation.

0371 The influence of supplementary health insurance on switching behaviour Dormont B. (Université Paris-Dauphine, Paris, France) Geoffard P-Y., Lamiraud K. This paper focuses on the switching behaviour of sickness fund enrollees in Switzerland. Though the regulation promotes competition between funds, there are large premium variations within cantons. This reveals some inertia among consumers. We investigate one possible barrier to switching behaviour the influence of supplementary insurance on the choice for basic insurance. Though there is a clear-cut regulatory separation between basic and supplementary insurance, it turns out that most enrollees take up basic and supplementary insurance with the same fund. We assume that insurers can risk adjust indirectly if they somehow tie the conditions of a supplementary contract to the possession of a basic contract at the same fund.develop a theoretical model that captures some important elements of the market. The focus is on the interaction between holding a supplementary contract and choosing a health plan. We then estimate a simultaneous equation model of this joint decision. We use adminstrative information and survey data (n=2152 enrollees) on health plan choice. We show that holding a supplementary contract decreases the propensity to switch. Our results suggest that the regulation of the supplementary health insurance market should integrate the effects it may have on the basic insurance market.

0555 Turning the uniform DRG base-rate into a volumecontrol contract in the Hungarian DRG system Dózsa Cs. (Ministry of Health, Budapest, Hungary) Nagy J., Sebestyén A., Kövi R.., Borcsek B., Boncz I. The aim of this policy paper is to analyse the process of the development of uniform DRG base-rate and turning it into a volume-control contract. Data and methods: The Hungarian DRG system was introduced in Hungary in 1993 replacing the former global budget approach of hospital financing. Our data derive from the financial database of the Hungarian National Health Insurance Fund Administration, covering the period 1993-2004. Results: From 1993 until March 1997 the Hungarian hospitals were classified into different groups (county-, city hospitals, universities, national medical institutes), and each group had its own DRG base-rate. Between March 1997 and January 1998 the DRG base-rate of different hospital groups had been closing up and in February 1998 we achieved a nationwide uniform base-rate. In addition to the DRG system, a volume-control contract has been added in 2004 as a ceiling for hospital performance. This volume-control resulted in a floating DRG base-rate, although not national but hospital level. Conclusions: The activity based financing of hospitals through the DRG system resulted in higher performance therefore higher health insurance expenditures. The volume-control contract provides an artificial ceiling over increased hospital performance.

0495 EFPIA HTA Key Principles: An Academic Perspective Drummond M. (Centre for Health Economics, University of York, UK) This presentation focuses on three of the key principles. First, should evaluations allow new data to be considered? The data available at product launch are seldom sufficient for a thorough assessment of cost-effectiveness. Therefore, systems need to be put in place to gather additional data post-launch and to consider these in decisions about the appropriate use of health technologies. Secondly, is flexibility required in handling the uncertainty around new technologies? Given that little will be Eur J Health Econom Suppl 1 · 2006

| S31

known about the full therapeutic value of new products until there is experience of real-world use, manufacturers and payers need to develop a flexible approach to handling this uncertainty. As new data become available, changes should be made in the reimbursement of the technology in a timely and efficient fashion. The existence of such a system may make the initial reimbursement decision less protracted, since both payers and manufacturers would know that changes could be made later. Thirdly, should assessments take place at the national level? Because a number of factors, likely to affect cost-effectiveness, vary from one jurisdiction to another, the most appropriate place for health technology assessments is the national, rather than international, level. In most countries sufficient resources should be available to conduct evaluations to a reasonably high standard.

0550 Evidence based policy making – Value of statistical life Dziworski W. (European Commission, DG Health and Consumer Protection, Unit 2, Strategy and analysis, Brussels, Belgium) The presentation will be a part of the organised session by the European Commission: Policy choices for Europe- ageing and sustainability. “The adoption of a European Commission policy proposal is mainly a political decision that belongs to the College of Commissioners, not to officials or technical experts. Ideally however, such a decision should be based in every case on sound analysis fed by the best data available. Impact assessment (IA) methodology developed by the European Commission is a key tool in this respect. Implementation of Better Regulation principles like IA methodology, through provision of a careful and comprehensive analysis of likely social, economic and environmental impacts, both direct and indirect, contributes to meeting the specific commitments of the Lisbon and Sustainable Development Strategies. It improves the quality of policy proposals and above all explains why an action is necessary and that the proposed response is an appropriate choice or, conversely, demonstrates why no action at EU level should be taken. DG Health and Consumer Protection has identified that it can best contribute to the needs of healthier citizens and more confident consumers by producing proposals whose beneficial impacts can clearly be explained and justified. Nonetheless, evidence based policy making in health area requires however high involvement of external experts, especially economists. Development of tools useful in policy making, like Value of Statistical Life or specific Health Impact Assessment methodologies is however essential. Thus the economic analysis could be lastingly incorporated into policy making to the benefit of EU citizens.”

0068 Dentists Migration patterns in the European Union and Economic Area (part 1) Overview and immigration of dentists to the United Kingdom Eaton K. (Eastman Dental Institute, University College London, United Kingdom) Balázs P. One of the key principles of the European Union is that all workers, including health workers, who are citizens of any of the 29 member states of European Union and Economic Area (EU/EEA) and who obtained their trade/professional qualification, may work anywhere within the EU/EEA. It can therefore be suggested that workforce planning should become a Pan-EU/EEA issue as an over or under supply is both socially and economically undesirable. At present, as far as health care workers are concerned, there appear to be few reliable data. As an example of the problem, this presentation and the one that follows will highlight the current deficiencies in the data for numbers of dentists in the EU/EEA. Data on immigration patterns of dentists to the United Kingdom (UK)and den-

S32 |

Eur J Health Econom Suppl 1 · 2006

tists’ patterns of migration in Hungary will be presented and discussed. In 2004, over 360,000 dentists were registered in the 29 EU/EEA member states. However, registration procedures vary from member state to member state and in many member states dentists remain on the register after they have retired or emigrated. In a recent survey of EU/EEA registration bodies, 11 were unable to provide any data for the numbers of dentists who had immigrated to their states. Of the EU/EEA member states who reported data, the greatest number of immigrant dentists were to be found in the UK (7236 out of a total of 33,314 registered dentists). An analysis of immigration patterns revealed that 4606 dentists were immigrants to the UK from other EU/EEA member states and that 2630 came from other non-EU/EEA countries. The greatest numbers came from South Africa (1474) and Sweden (1454). In three of the last seven years the number of new registrants from outside the UK has been greater than the number graduating from UK dental schools. Over the past 15 years there have been discernable trends in the patterns of immigration of dentists to the U K, which can be associated with economic and political factors in the countries of origin of the dentists concerned. In 2005, the greatest number of immigrant dentists to the UK came from Poland and India. The latter are required to pass the International Qualifying Examination before they can be registered for practice in the UK. It can be concluded that although the immigration of dentists to the UK has benefited this country’s economy, there may be moral objections to recruiting dentists (and other health care workers) from economically less well developed countries.

0007 Individual Socioeconomic Effects on the Demand for Health Care Services in European Union Countries Economou A. (University of Macedonia, Economics Dept., Thessaloniki, Greece) Nikolaou A.., Theodossiou I. The purpose of this study is to shed light on the individual socio-economic status (SES) and demographic determinants of the demand for health care in a cross-comparison study of nine E.U. countries. The existing literature, both at the theoretical and the empirical level, verifies the existence of a SES gradient on the demand for health care. The present study gives special emphasis on the effects of individual employment characteristics on alternative indicators of health care demand, an issue that requires further attention by applied research. The evidence supports the existence of a SES gradient based on employment characteristics on the demand for health care, differentiated with respect to the type of health care examined and the institutional and environmental settings of the countries utilised in the study.

0098 Cost effectiveness of intravenous thrombolysis with rt-PA within 3 hours window following acute ischemic stroke Ehlers L. (Aarhus University Hospital, HTA Unit, Aarhus, Denmark) Andersen G., Clausen LB., Bech M., Kjoslashlby M. Introduction: The aim of this study was to assess the cost-effectiveness of treatment of acute ischemic stroke within 3 hours window using intravenous thrombolysis with rt-PA compared to conservative treatment. Material and methods: A health economic model was designed to calculate the marginal cost-effectiveness ratios for time spans of 1, 2 and 30 years. Effect data were extracted from a meta analysis of six largescale randomised and placebo-controlled studies of thrombolytic therapy with rt-PA. Cost data were extracted from a pilot study of thrombolysis treatment at the Aarhus Hospital and from published literature. Results: The calculated cost-effectiveness ratio after the first year was US$/QALY 46 460. After the second year computation of the cost-effec-

tiveness ratio showed that thrombolysis was both less expensive and better (dominance), but the result was sensitive to changes in several variables. In the long term (30 years), thrombolysis was a dominant strategy given the model premises. Discussion: The short term results showed that thrombolysis costs exceeded traditional cost thresholds recommended upon introduction of new medical technologies. The long-term computations showed potentially large-scale health economic cost savings, but these calculations are uncertain. The analysis concludes that the empirical data needed for longitudinal computation are currently not available.

0283 Pharmaceutical innovation, ageing population and health care expenditures Eisen R. (Johann Wolfgang Goethe-University, School of Economics and Business and Institute of European Health Policy and Social Law, Frankfurt am Main, Germany) Ilgin Y. Pharmaceutical innovation, ageing population and health care expenditures should an ageing population be a problem? Sometimes it is argued that ageing population inherently causes high future health care expenditures then controlling health care expenditures appears reasonable. Several studies forecasting health care spending predict continuous increasing expenses due to at least three different factors ageing, technical progress, and increasing prices. Here we look in particular at pharmaceutical innovations. Whether the effect of pharmaceutical innovations are cost driving or cost lowering to health care expenditures will be analysed and evaluated with regard to their current impact. Arguing as for total health expenditures that the corresponding proximity to death is integrant not the age then this proximity would be the crucial element for rising expenses. Still in older age cohorts there are assumable more people in their last life span. Therefore in this period it is important to determine whether more and more treatments or pharmaceuticals are applied as a vainer attempt to avert death which in consequence causes disproportionate high pharmaceutical and/or health care expenditures or whether the expectancy of life is increasing by medical progress or by a more healthy way of living. The attempt to sketch the situation indicates that there must be a kind of multiplicative and interdependent relationship between influencing factors such as pharmaceutical expenditures, medical progress, ageing structure or health-related absenteeism. The interdependent relationship could be founded on the Sisyphus Syndrome, arguing that medical progress causes an increased longevity and this increase in turn leads to an amplified allocation of resources of health care and therefore rising expenditures. Or it could be founded on the multimorbidity of older patients. The successful medical combat of one disease would increase longevity without making them healthier but inducing the next disease to be medicated.

0076 Treatment of diabetes is essential for preserving quality of life and maintaining productivity Emneus M. (Novo Nordisk, Bagsvaeligrd, Denmark) Bjoslashrk S., Christiansen T., Green A. A socio economic assessment of diabetes in Denmark (year 2001) is presented. No treatment of diabetes leads to serious disabilities and premature death. Based on epidemiological evidence, a demographic model for diabetes in the years 1900-2002 has been established, providing estimates of the number of patient-years lived with diabetes as currently (CS); in a worst scenario (WS) with no treatment options; improved scenario (IMS) with improved treatment; ideal scenario (IS) where diabetes is free of excessive risk of complications and mortality. Current costs have been estimated from public registers and official sta-

tistics, and clinically based estimates are structured in healthcare and non-healthcare resources, patients’ time and informal caregivers’ time. Effects have been quantified as patient-years, quality-adjusted patientyears and income by working including household production. Currently (2001) more than 153.000 people live with diabetes. IMS showed 16.000 life-years gained compared to current treatment with the gain in productive life-years larger than the number of life years. 20-35% of the current total costs is related to health care; the rest is cost of nursing of people with irreversible complications. Improved treatment could reduce costs to 60% of current level due lower non health care cost and less people with sever complications.

0453 Optimal prevention and saving: individual decisions and public policy Etner J. (GAINS, University of Maine and EUREQua, Paris) Jeleva M. The aim of this paper is to analyze optimal decisions of prevention and saving of individuals facing a long term health risk involving expenses not covered by insurance policies. In a first part we study the interactions between the two decisions (saving and prevention) and their determinants. In the second part of the paper, we assume that prevention generates positive externalities and determine the optimal public policy instruments in this context. More precisely, we consider individuals whose utility function depends both on consumption and health and who face a health risk. In case of disease, individuals face expenses that involve on one side a minimal treatment costs, covered by insurance and on the other side additional costs (higher quality treatment, home care etc.) not covered by insurance and thus paid by the individual himself. To manage health risk, individuals have two possibilities: invest in primary prevention, in order to reduce the probability of the disease and/or save, in order to face the costs of the disease that are not covered by insurance. We prove that the impact of wealth and interest rate on prevention and saving decisions strongly depends on the perceived substitutability between health and wealth. When prevention generates positive externalities (as vaccine for infectious diseases for instance) as expected, we prove that individuals’ optimal levels of prevention are lower than the socially optimal ones. This gives room for public intervention in order to restore optimality. The instrument that we analyze consists in subsidizing prevention expenses and financing these subsidies by a tax on saving benefits.

0089 Risk Based Guaranteed Renewability in Competitive Social Health Insurance Eugster P. (Socioeconomic Institute, University of Zurich, Switzerland) Zweifel P. Without major change in the premium, guaranteed renewability (GR), i.e. the unilateral commitment of an insurer to extend the life of a contract, is an important feature of competitive health insurance that has received not much attention in the literature. As an exception, Pauly, Kuhnreuther and Hirth (1995) have provided a theoretical analysis of GR. Their work forms the point of departure for the present contribution, which seeks to determine the degree to which GR is jeopardized in a competitive social health insurance system that imposes uniform premiums. This danger is reflected by the discrepancy between a riskbased premium and the regulated contribution. Using data from one of the largest Swiss social health insurers individual risk-based shadow premiums are calculated on the standard contract with minimal deductible. This allows for estimation of an upper bound of a GR-compatible premium. These premiums are then compared with those estimated for the U.S. individual insurance market by Pauly and Herring (2003). Eur J Health Econom Suppl 1 · 2006

| S33

0260 Savings in the Hungarian managed care pilot programme between 1999-2004 Falusi Zs. (National Health Insurance Fund Administration (OEP), Department of Health Policy, Budapest, Hungary) Boncz I., Nagy B., Sebestyén A., Kóti T., Dózsa Cs. Background and aim: In 1999 a new managed care pilot programme was introduced in Hungary. Nowadays 16 organizers (Managed Care Organizations, MCOs) take the responsibility of managing health care for a population of 2,300,000, which is 23 % of the total Hungarian population. The aim of the study to calculate the savings realized in the managed care pilot programme. Data and methods: Data derive from the financial database of the Hungarian National Health Insurance Fund Administration, covering the period 1999-2004. We calculated the savings as the difference between the risk adjusted capitation of MCOs and their expenditure. Results: The total savings of the programme were as follows: 3,56 % (1999), 10,40 % (2000), 6,54 (2001), 8,73 % (2002), 3,42 % (2003), 4,01 % (2004). The saving of the individual MCOs varied between 0-17,94 %. The saving measured by the % of expenditures was the highest in chronic and acute in-patient care, medical devices’ reimbursement, while measuring by nominal monetary value it was the highest in acute in-patient care, drug budget and out-patient care. Conclusions: Although the Hungarian managed care pilot programme led to a cost saving, there is a hard debate on the further development of this pilot programme.

0481 Value of Treatment – Results of an Event Study on the Formation of Alliances in the European Biotechnology Industry Farag H. (European Business School, Oestrich-Winkel, Germany) Prior research has often used the biotechnology industry as setting for analysing collaborative activity. However, most researchers have restrictively focussed their work on general factors as explanatory variables for the value created through alliance formation, including alliance direction, firm resource endowments, and functional areas of collaboration. In the context of the European biotechnology industry, the present study extends this line of research by explicitly considering industry-specific determinants of value creation. Specifically, it addresses the value associated with different indications (i.e. areas of treatment), technologies used, as well as stages of clinical development. The present research relies on advanced event-study methodologies accounting for the industry-level drivers of biotechnology firms’ stock prices. In line with current econometric practices, cross-sectional analyses use random-effects panel data models. The findings document substantial overall returns to alliance activity. Collaborative benefits significantly vary depending on the underlying technologies and targeted medical areas of treatment, with highprofile, high-sales indications and advanced stage projects experiencing the highest returns. Furthermore, the analysis identifies two alternative strategies of collaborative value creation: Firms may either work towards ‘high value’ transactions or pursue multiple smaller agreements.

0071 The economic consequences of good quality of laboratory tests Fauli S. (The Norwegian Medical Association, Oslo, Norway) Thue G. Objective: To develop a method for evaluation of the economic consequences of quality improving measures of laboratory analyses.

S34 |

Eur J Health Econom Suppl 1 · 2006

Design: Helicobacter Pylori (H. pylori) tests are used to illustrate a method. A decision model is designed to compare expected costs and outcomes of three H. pylori tests with different analytical quality. Information from our data and medical literature are used to determine probabilities of clinical events. A cost-effectiveness analysis is done from a societal perspective. Finally we use a probabilistic sensitivity analysis (PSA) to model the decision uncertainty. Main results: In a 100 days period, tests with high accurate quality (high sensitivity and specificity) dominate tests of less accuracy. The cost of the quality improvement programme is minor compared to the potential of cost-reductions. Of the two tests with similar accuracy the test with highest sensitivity was significantly more cost effective than the test with highest specificity when the willingness to pay for each healthy day was more than €18.4. Conclusions: We have developed a method that can be used to evaluate the economic consequences of quality improving measures of laboratory analyses. This study demonstrates that even small improvements in quality may be cost-effective.

0510 The economic consequences of ill health in South-eastern Europe Favaro D. (University of Padova Department of Economics, Padova, Italy) Suhrcke M. The purpose of the present study is to provide empirical evidence on the impact of adult ill health in South-eastern Europe (SEE). We assess quantitatively the economic impact of ill health in SEE both from a macroeconomic and microeconomic perspective. On the macro level we project the likely impact of improved health on future economic growth based on a growth regression framework, and conditional on various, plausible future mortality rate scenarios. On the micro level, we evaluate the effect of ill-health on occupational status, productivity, earnings, hours of work and early retirement we estimate several econometric models, controlling for the endogeneity of different measures of health, using primarily the World Bank‘s LSMS datasets (esp. from Albania, Bosnia & Herzegovina, Bulgaria, Kosovo). Results strongly confirm the negative impact of ill health on both the individual and the macro level. Depending on the country, ill health seems to have a negative effect on the probability of being employed, on the wage rate, on wages and on hours of work. On the macro level we find that reducing the adult disease burden in a sustained manner would produce substantial economic benefits for the economy.

0381 Costs Analysis of Breast Cancer Screening in Italy Federici A. (Agency for Public Health (ASP) Lazio, Rome, Italy) Borgia P., Guasticchi G., Ricciardi A., Palazzo F., Ratti M., Mennini F.S. The present work analyses the costs of Breast Cancer screening programme to define a new financial system based on real costs.the screening program into different processes (levels) and activities (costs centers) related with disease management, the Center of Technology Assessment in Public Health (CTAph) has collected data in five Local Health Units (LHU) of Lazio Region. The whole total costs are respectively € 1,557,183 for LHU 1 which screened 26,080 women (39,7% of involved population) € 1,998,142 for LHU 2 which screened 39,485 women (62,1% of involved population), € 691,811 for LHU 3 which screened 9,919 women (14,5% of involved population), € 345,595 for LHU 4 which screened 6,839 women (55,5% of involved population) and € 745,289 for LHU 5 which screened 15,453 women (63,7% of involved population).the new fee system the CTAph proposes to finance the effective activities of each level. It requires to upgrate the only one fee per invited woman with three different fees based on screened persons in

each level. In this way the productivity of each costs center is properly remunerated; as well it’s possible to reduce the waste of the financial resources.

0373 Cost-effectiveness of interventions to reduce tobacco smoking. An application of the RIVM Chronic Disease Model Feenstra T. (Institute for Public Health and the Environment, RIVM, Bilthoven, Netherlands) Baal P., Hoogenveen R., Vijgen S., Bemelmans W. Smoking is related to substantial morbidity and mortality. This study assessed the cost-effectiveness of eight interventions to enhance smoking cessation in adults, namely tobacco taxes, mass media campaigns, minimal counseling, structured GP-support, telephone counseling, nicotine replacement therapy with minimal and intensive counseling, and bupropion with intensive counseling. To estimate costs per (quality adjusted) life year gained, a dynamic population model, the RIVM chronic disease model, was used to project health gains and effects on health care costs. Sensitivity analyses were performed for variations in costs, effects, time horizon, program size and discount rates. Tax increases resulted in costs per QALY around €5000. The costs per QALY for a mass media campaigns was below €10,000 for a broad range of effects. Costs per QALY for individual cessation support varied from about €7000 for GP support to €21,000 for telephone counseling. In conclusion, most interventions were cost-effective compared to current practice. Evaluating all inter ventions with the same model improved comparability. However, comparison of inter ventions is difficult, because taxes and campaigns are combined with individual support. Taking that into account, taxes seem to provide most value for money, with tax revenues probably more than making up for possible costs of regulation.

0255 The Effects of Optional Deductibles in German Social Health Insurance Felder S. (Institute of Social Medicine and Health Economics, University of Magdeburg, Germany) Werblow A. A German sickness fund offers 240 € per year to the clients if they pay the first 300 € of their health care bills, except for doctor visits where a flat rate of 20 € applies. The paper studies the effects of this deductible on health care demand by comparing the 12,000 participants with a control sample. We use insurer’s claims data covering inpatient care, prescription drugs and ambulatory care. The data extends to three years including two years before and the year when the program started. We apply two methods, a matching and a parametric approach which model the choice of the deductible, the probability of positive expenses, as well as the demand for health care services conditional on having positive demand. study shows that individuals rationally opt for or against the deductible. Controlling for the factors that explain the choice of the program, the number of general practitioner consultations decreases by 24 percent and that of specialists by 42 percent. Moreover, the outlays for prescription drugs and inpatient care also significantly decrease. Prevention activities, not subject to the deductible, however, remain constant. The qualitative results do not depend on whether we use a non-parametric or a parametric approach.

0479 High Costs for Children with Juvenile Idiopathic Arthritis (JIA) Feldman D. (Université de Montreal, Ecole de réadaptation (School of rehabilitation), Montreal, Canada) Bernatsky S., Duffy C., St. Pierre Y., Malleson P., Feldman D., Gibbon M., Oritz-Avarez O., Clarke A. Juvenile Idiopathic Arthritis (JIA) is a potentially devastating chronic pediatric disease. OBJECTIVE: To describe direct health care costs in JIA, compared to pediatric controls. METHODS: Consecutive clinic attendees (N=155) with JIA were enrolled from two pediatric centres along with outpatient clinic and community controls without JIA (N=181). Data on direct health care costs were obtained at 3-month inter vals, using The Cost Assessment Questionnaire. Unit costs for health services were obtained from federal, provincial, hospital, lab, and professional association sources. Average annualized direct health care costs were calculated in 2002 Canadian dollars and compared for JIA subjects vs. controls. Results: The average age at enrollment was similar in the JIA sample (10 years, SD 4.3) vs. controls (10.5 years, SD 4.0). The total difference in annualized average direct health care costs for JIA vs. controls was $ 1,261 (95% confidence interval: $606, $1,916). Regarding cost components, JIA subjects had substantially higher costs related to medication use: difference of $682 (95% confidence inter val: $460, $904). JIA subjects also had higher costs related to visits to health care professionals and diagnostic tests (imaging/lab). Conclusions: The economic impact of JIA is substantial. Work is in progress to describe the total economic impact of JIA, including indirect costs, and establish clinical/social determinants of cost in JIA.

0048 Cost-effectiveness analyses of a rotavirus vaccination program in Brazil Ferraz M. (Sao Paulo Center for Health Economics, Sao Paulo, Brazil) Soarez P.C., Ciconelli R. M., Ferraz M.B. Background: Rotavirus is the leading cause of severe acute gastroenteritis in children < 5 years of age worldwide. It is estimated that, each year, 350,000–600,000 children die of rotavirus infection, 2 million are hospitalized and 25 million require an outpatient visit. A new rotavirus vaccine was recently licensed, and Brazil was the first country to implement a universal vaccination program. Objectives: The aim of this study was to estimate the cost-effectiveness of a universal rotavirus immunization program in an infant population in Brazil. Methodology: A decision tree used sources of local rotavirus disease incidence, health care expenditure, vaccine coverage, efficacy, and price. A sensitivity analysis using case scenarios toward and against vaccination was performed in some variables. Future costs were discounted to present value. Estimates were made from the societal perspective. Results: A universal rotavirus vaccination program could prevent 14,362 domiciliary cases, 11,127 outpatient cases and 1,258 hospitalization cases. It also could avert approximately, 26,746 diarrhea episodes, and 78 deaths among these children. The cost per case and death averted would be $1,487 and $512,635, respectively. And a QALY gained would cost $10,614. Conclusions: A rotavirus vaccination program is cost-effective depending on the vaccine price, mortality and morbidity incidences. Disclosure information: This research was carried out on behalf of CPES. CPES is a not for profit organization associated to the Federal University of Sao Paulo. The authors are researchers of CPES and have no conflict of interest with regard to this project.

Eur J Health Econom Suppl 1 · 2006

| S35

0293 How consistent are health utility values? Ferreira P. (Faculty of Economics, University of Coimbra, Portugal) Ferreira L.N. The use of preference-based generic instruments to measure Health Related Quality of Life of a general population or of individuals suffering from a specific disease has been increasing. However, there are several discrepancies in terms of utility results between instruments.study seeks to compare SF-6D and EQ-5D across subjects with cataracts and to investigate the differences in agreement between them. The main objective is to explore the eventual reasons for divergences found and to explore their implications.between EQ-5D and SF-6D health state classifications is assessed by correlation coefficients. The distribution of individuals reporting the lowest or highest health state in each instrument is analysed in terms of their report in the alternative instrument. Econometric models were estimated to examine the nature of the relationship between both indices. is evidence for floor effects in SF-6D and ceiling effects in EQ-5D. Comparisons of mean indices found that SF6D values exceed EQ-5D values. Differences in valuation methods and in scoring algorithms contribute to the main differences. The interpretation of the constant term and the interaction terms generate important effects on the indices.overcome their weaknesses, revisions of one or both instruments in their descriptive systems or on their scoring algorithm are necessary.

0310 An attempt to reduce the floor effect in the health utility measurement Ferreira L. (University of Algarve, Faro, Portugal) Ferreira P.L. The SF-6D is a new preference-based measure of health derived from the SF-36. It has become widely used in economic evaluation, but it still has several limitations.study aims to investigate the problems found in this measure and to suggest some changes to it. This work also seeks to study the models underlying the SF-6D and to estimate new models for predicting health state utilities.sample of 249 health states defined by the SF-6D have been valued by a representative random sample of the general population, stratified by gender and age, using the Standard Gamble (SG). Additional 50 health states including extra levels on two dimensions of the SF-6D were valued, aiming to solve the floor effect assigned to SF-6D models were estimated on the relationship between the SF-6D and the SG values and analysed in terms of their coefficients, overall fit and the ability for predicting the SG values for all health states. The coefficients were found to be robust across model specification, both in the case of the SF-6D previously defined by Brazier and in the case of the SF-6D added with extra levels in two dimensions.preferencebased utility measure used seems to adequately predict the health states values of the general population. The results confirm the usefulness of additional levels on the SF-6D, improving its efficiency in measuring the health states utilities and in diminishing the floor effect.

0133 Estimating heterogeneity in patient preferences for a dermatology consultation Flynn T. (MRC HSRC, University of Bristol, UK) Louviere JJ., Peters TJ., Coast J. Greater patient choice is regarded as increasingly important in the delivery of health care. However, whilst an analytical framework exists to help inform population-level priority-setting in health care, there are little quantitative data about individual patient preferences for attributes

S36 |

Eur J Health Econom Suppl 1 · 2006

of ser vice provision and treatment. Best-worst scaling (BWS) studies not only provide quantitative estimates of patient preferences but they are particularly well-suited to estimating differences in preferences between patient subgroups defined by clinical or sociodemographic factors. BWS study was administered in a trial to value patients preferences for four aspects of a dermatology consultation waiting time, convenience of attending, expertise of doctor and degree of individualised care. Whilst waiting time was found to be the least important attribute, multinomial logit models indicated that it became more important when patients felt that aspects of their daily lives were affected by their skin condition. Furthermore, ‘widening’ of level scale values around the attribute importance estimates was found in more educated patients. The policy implications of these findings will be summarised and issues surrounding the compatibility of patient-centred care with national cost-effectiveness guidelines will be raised.

0509 The role of health economics analysis in decision making – an illustration from Sweden Forsberg B. (Dept of Public Health Sciences Karolinska Institutet, Medical Management Centre Karolinska Institutet, Dept. of Public Health Sciences Karolinska Institutet, Stockholm, Sweden) Medin E., Rehnberg C. This is a study of the decision making process around the purchase of high-cost technical equipment at a major university hospital in Sweden. The County Council and the hospital management were faced with different alternatives for providing the ser vices of a “Gamma Knife”: purchase the ser vice from a private provider on an item basis, purchase the equipment in the international market and operate it directly within the publicly owned hospital or purchase it locally for the same purpose. The cost-effectiveness of each of the options is presented. The evidence base in the decision making process was reviewed and key actors were inter viewed on the role of various pieces of information, in particular cost-effectiveness and financial analysis, in taking the decision on purchase and ser vice provision. The study shows that decisions on such huge investments are preceded by a complex process in which health economics makes an important contribution. Still, decisions are in the end influenced by many factors, health economics analysis being only one of them. The study suggests how the role of cost-effectiveness and financial analysis can be strengthened in health care decision making.

0035 Fatal and Nonfatal Motorcycle Injuries: Can Alcohol Policies Influence Rider Safety and Reduce Accidents? French M. (University of Miami, Coral Gables, Florida, USA) Numerous studies have established the effectiveness of alcohol policies (e.g., blood alcohol content for driving under the influence, DUI penalties and fines, minimum legal drinking age, beverage taxes) in reducing rates for automobile accidents, injuries, and fatalities. Only a few studies have analyzed the impact of alcohol policies on motorcycle accidents and none within the past 5 years. Given the growing popularity of motorcycle riding among all age groups and the inherent safety risks, a new and comprehensive investigation is warranted. The present study uses state-level longitudinal data from 1994 to 2003 to determine whether a variety of different alcohol and motorcycle safety policies have a significant impact on both fatal and nonfatal motorcycle injuries. Besides state-level policies, the analysis controls for numerous demographic and environmental factors in each state through the estimation of state fixed-effects models. The results lead to direct policy recommendations for how states can reduce motorcycle injuries and enhance overall motorcycle safety.

0542

0047

Images of Informal Payments for Health Care: Definitions, Theories and Practice

Sources of inefficiency among hospitals in Ireland

Gaal P. (Health Services Managment Training Centre, Semmelweis University, Budapest, Hungary) Objectives: Informal payments are known to be widespread in the postcommunist health care systems of Central and Eastern Europe and the countries of the former Soviet Union. However their role and nature remains contentious, with the debate characterised by much polemic. There is not even consensus on the definition of the phenomenon, although there is a general understanding of what is considered informal payments and what is not. This study steps back from the debate to examine the theoretical basis for understanding the persistence of informal payments, and confront this conceptual analysis with empirical evidence. Methodology: The study combines theoretical and empirical methods, including the review of the theoretical and empirical literature, conceptual analysis and case studies. In particular the study: – examines the variants of informal payments that have been reported in the literature, – analyses the proposed definitions of the phenomenon, points out their shortcomings and suggests a new definition based on the concept of entitlements and the additional nature of payments as the distinctive feature common to all types of informal payments, – considers the implications of this new definition for research and policy making, – identifies the motivation for informal payments as the key issue for health policy, and develops of new theoretical framework on the basis of the theory of government failure and Hirschman’s theory of “exit, voice, loyalty”, – uses a case study of private sector informal payments to understand the motivation for informal payments and to test the validity of the “inxit” theory of informal payments. Results: The wide variety of the different forms (or images) of informal payments including monetary transactions, in kind gifts (such as chocolates, spirits or flowers) and in kind contributions (such as drugs, clean linen, meals or nursing services) are not exclusively illegal, corrupt or informal and are not confined to the public sector either. The common to all types of informal payments is that they are given to health care providers in addition to any contribution that has been stipulated in the terms of entitlement. This raises the question: ‘Why do people pay more than what is required?’ Is it because they are genuinely grateful for being cured (donation hypothesis)? Is it because they are not aware of what they are entitled to? Or is it because they want to secure that they get the care they need (fee-for-service hypothesis)? To answer these questions we have explored the case of private sector informal payments in a small private for-profit hospital in Hungary, and it turned out that attempts by patients to pay physicians informally do exist in the private sector. Nevertheless, the deeper analysis of motivation for such payments revealed that they are rather attributable to the blurred boundaries between the public and the private sector than genuine gratitude. Conclusions: The key health policy question is the motivation for informal payments: whether it is gratitude or there is a pervasive failure of the health care system behind them. The theory of “inxit” considers informal payments as a potential reaction of patients and doctors to declining organizational and system performance, in addition to exit and voice described earlier by Hirschman. Although the existence of phenomenon in the private sector seems to invalidate this explanation, empirical evidence shows that even in these ‘extreme’ cases, genuine gratitude is probably rarely motivates informal payments. On the other hand the example of private sector informal payments highlights the complexity of the problem, policy makers face if they are to eradicate the phenomenon. Policies to tackle informal payments should find the delicate balance between exit and voice to optimise the chances of maintaining a good standard of public services.

Gannon B. (Economic and Social Research Institute, Dublin, Ireland) The analysis of efficiency in hospitals can make a major contribution to improving health services. The ultimate aim is to (1) identify poorly performing hospitals, (2) to understand why and (3) to address the underlying causes. Previous research suggests that there is substantial variation in efficiency among hospitals in Ireland, (Gannon, 2005). The aim of this paper is to explore the underlying causes. A two-stage model is applied to hospital level data between 1995 and 2004. Potential contributors to efficiency variation include hospital size, age distribution of patients and local environmental factors. Furthermore the issue of quality is addressed, and potential quality indicators include hospital specific mortality rates. In addition, the model includes results from a recent indepth hygiene audit on all hospitals. By identifying the underlying links between inefficiency and related factors, the results are useful in the context of ongoing change in the national health policy, B. (2005). “Testing for variation in technical efficiency among hospitals in Ireland”, The Economic and Social Review, Forthcoming December 2005.

0051 The Influence of Economic Incentives on Reported Disability Status Gannon B. (Economic and Social Research Institute, Dublin, Ireland) Self-reported disability status is often relied upon in labour force participation models, but this may be reported with error for economic or psychological reasons and can lead to a bias in the effect of disability on participation. In this paper, we explore the possibility that reported limitations in daily activities are misreported, in particular for those who define their labour force status as disabled, and assess if financial incentives influence this group to mis-report. The main questions we wish to address are (1) was there state dependent reporting error and did financial incentives play a role, and (2) did this change over the years 1995 to 2001? Using a generalised ordered logit model, we compute cleansed measures of disability that correspond to predicted responses individuals would have made if employed. The preliminary results indicate that the disabled group did over-report and the difference between actual and predicted probabilities fluctuated between 1995 and 2001. We discuss two particular institutional changes in Ireland that help to explain how economic incentives influenced reporting behaviour.

0448 A review of the pharmacoeconomic studies conducted in Italy. Garattini L. (Mario Negri Institute, Milan, Italy) Bassi L., Cornago D. To assess the state of pharmacoeconomics in Italy we reviewed the full pharmacoeconomic EEs done there and adopted common criteria of analysis to allow methodological comparison of the studies. The variables investigated can be grouped in three categories: general methods, costs, and consequences. We selected all the original studies published by Italian authors in national and international journals from January 1994 to December 2003. We reviewed 70 articles and broadly assessed 92 EEs since some articles contained multiple analyses. Only 20 EEs were classified as “ideal” according to the quality of the sources of clinical evidence and 21 EEs were classified as “useful to the NHS” according to the credibility of economic evidence: by combining the results of these two separated analyses, we found that only seven out of the 92 EEs appeared to be credible. To complete our critical evaluation, we analysed whether sponsorship might have somehow affected Eur J Health Econom Suppl 1 · 2006

| S37

the results and found that the vast majority of sponsored EEs showed positive results for the sponsored drugs. Our review on the EEs by Italian authors confirms the legitimacy of the public authorities’ scepticism towards pharmacoeconomic studies, whose results seem to be biased by flawed methods and sponsors’ interference on results.

0475 Child Obesity and Maternal Employment García E. (Foundation for Applied Economic Studies (FEDEA), Madrid, Spain) Increase of overweighted and obese children has been dramatic for the last decades of the XX century in developed countries. More than social acceptance or aesthetic considerations, obesity consequences on health and wellbeing are dangerous. Consequences of obesity have been largely dealt with in the literature. However, its causes still need to be cleared out. We center our work in some of the causes of obesity. Specifically, this paper investigates whether children are more or less likely to be overweight if their mothers works. During the last decades the evolution of women employment and over weight children in Spain has shown similar trends. In this sense, it is possible that female labor force participation has lead to a change in the traditional concept of childhood care and in family habits that may have affected children physical appearance and health. The main aim of this paper is to help determine if these parallel trends are just drawn by spurious correlation or by a causal relationship between the two variables. Whether or not we find causal effects has important implications for policy measures either in the labour market or fiscal ones. Using data from the National Health Survey for 2003, we relate mothers’ personal and professional situation with the degree of overweight and obesity of their children. Our results indicate that obesity ranges with the age and sex of children. Children of low-educated mothers and of overweight and obese mothers are also more likely to suffer from overweight or obesity. Finally, a child is more likely to be overweight or obese if his mother works, even if we control for genetic, environmental and socio-economic factors.

0268 Health Shocks, Employment and Income in the Spanish Labour Market García Gómez P. (Centre de Recerca en Economia i Salut (CRES) Pompeu Fabra, Barcelona, Spain) Nicolás A.L. This paper investigates the relationship between health shocks and labour outcomes in the Spanish population using the European Community Household Panel. In order to control for the non-experimental nature of the data we use matching techniques. Our results suggest that there is a significant effect running from health to the probability of employment and to labour income. Moreover, while we cannot investigate the influence of childhood events and other phenomena that trigger long run causal pathways from socio-economic status to health, we are able to find a significant reduction in the probability of reporting good health in individuals who transit out of employment in comparison with individuals who are otherwise identical in terms of reported health status at the time of the transition.

0062 Hybrid Risk Adjustment for pharmaceutical benefits Garcia-Goni M. (Universidad Complutense de Madrid, Pozuelo de Alarcón, Spain) Ibern P. Pharmaceutical benefits are increasingly under close monitoring since costs, quality, and appropriateness are crucial for health outcomes. While prospective risk adjustment has been shown to promote incentives for efficiency through controlling the costs, retrospective or concurrent formulae may obtain a higher predictive power and avoid incentives for selection. This paper follows a relatively new literature on risk adjustment in which a mixed formula – with both prospective and concurrent payments – seeks to maximize the efficiency incentives involved in the prospective risk adjustment and minimize the selection incentives through a concurrent risk adjustment for high risk individuals. We utilize individual data on pharmaceutical expenditures and health conditions -using Clinical Risk Groups (CRGs) and Diagnostic Cost Groups (DCGs) for morbidity adjustment- for the population belonging to an integrated healthcare delivery organization in Catalonia in years 2002 and 2003. Former results with the same database showed that predictive ratios predictive ratios and variance explained were in line with published research. The goal of this paper is to show the gain of a hybrid risk adjustment that can be used for resource allocation and management, focusing on a definition of regulatory mechanisms that can be applied in a setting with universal coverage and public financing.

0141 The impact of Australian government policy change on the private price of nursing home care Gargett S. (The University of Queensland, Wacol, Brisbane, Australia) This paper will analyse the impact that changes in Australian Government policy may have had, on the private price of nursing home care. The Australian Government has subsidised nursing home care since 1963, with residents also contributing to the cost of care. The Government’s policy as to amounts residents have had to contribute has varied, with new policies introduced in 1973, 1991, and 1997-98. Of specific interest is whether the most recent reforms, which required that residents contribute based on their assessed ability to pay, significantly altered the price residents paid. ‘intervention analysis approach’ will be used to analyse data on the price paid by residents. Estimates of the average annual net price paid per resident per occupied-care day, have been determined for 1960-61 onwards. After assessing the stationarity of the data, an autoregressive integrated moving average model (ARIMA) will be specified and estimated. Dummy variables are included to represent the policy interventions. The impact of the policies will be interpreted in the light of the coefficient values and the adequacy of the model. Given the predicted increase in expenditure on aged care services over coming decades, it is useful to evaluate the effects that policies, which may influence the demand for such services, may have had.

0059 The overall effect and significance of drug price elasticity among US seniors Gemmill M. (London School of Economics, London, UK) McGuire A. Despite the importance of prescription drugs for the elderly and increasing private and public expenditures on prescription medicines, few studies have estimated the elasticity of demand for prescription drugs among the elderly, and none have derived an estimate for the elderly as a whole in the United States. This paper draws upon the 1996-2002 Med-

S38 |

Eur J Health Econom Suppl 1 · 2006

ical Expenditure Panel Sur vey and uses panel data techniques to provide additional empirical evidence that fills the existing gap in the literature. The econometric techniques refer to models that treat data counts in the presence of missing observations in two distinct ways one model uses traditional methods of accounting for attrition and one fixed effects model simultaneously controls for rotating panels and unobserved heterogeneity. The results indicate that increased levels of cost sharing lead to a decrease in the demand for prescription drugs with pronounced elasticity estimates that differ from earlier studies. The estimates depend on the specification control employed but are generally calculated to be within a narrow range. In comparison to other studies that have calculated elasticities for the general population, our estimates are lower, possibly indicating that the elderly perceive fewer substitutes for prescription medicines than the general population.

0472 Reference Price on off-patent drugs: an analysis of the Italian and the Spanish experience Ghislandi S. (CESAV, Mario Negri Institute, Ranica BG, Italy) Junoy J.P., Garattini L. A comparative analysis of the Italian and the Spanish experience with Reference Price (RP) is performed. In Italy, the RP for off-patent products was first introduced in December 2001. In Spain, the RP was introduced in December 2000 and reformed in January 2004. As Italy and Spain present very similar situations in terms of pharmaceutical market structure (late introduction of the patent system, overcrowded market for copies, marginal role of generics, heavy price regulation), it is interesting to compare the outcomes of the two countries. Prices and RP trends are analysed for both countries using a sample of monthly observations. In Italy, the competitive pressure of generics is shown to play a primary role in the downward trend of both the RP and branded prices. In Spain, price reductions are also significant, but somehow delayed. The results confirm findings from the previous economic literature on RP.

0473 Optimal Pricing with Endogenous Reference Price Ghislandi S. (University of Oxford and CESAV, Italy) Here it is presented a model which tries to address some of the main problems related to the Reference Price (RP) literature. First, the RP is endogenous. Second, we improve the previous literature by introducing one branded product and n generic firms that are allowed either to compete or to collude among themselves. Third, the pricing game is modelled first as a one shot game, then as a two-stages game where in the first period firms set the RP, while in the second they compete taking the RP as exogenous. Finally, we will keep the analysis as general as possible, obtaining the results from a very broad set of realistic assumptions. Results are clear. RP mechanisms where the reimbursement level depends on the branded price are shown to be always less effective in keeping price down, independently from both the level of competitiveness among generics and the type of the game played. In case of a one shot game, no Pure Strategy Nash Equilibria exist when the RP is a function of the branded product. In the two stage game, collusive outcome are shown to be always worse than the competitive one. However, interestingly, first stage generics prices can be strategically lower under collusion than under perfect competition.

0291 Technological Characteristics and Cost Efficiency of Public Hospitals in Piedmont Piacenza M. (National Research Council (Ceris-CNR), Moncalieri, Italy) Alberto C., Turati G., Vannoni D. In this paper we evaluate cost efficiency and technological characteristics of public hospitals in Piedmont – a Region of the Northern Italy – over the period from 2000 to 2004. During the Nineties, different regulatory reforms aiming at controlling health expenditure affected the Italian National Health Service, ranging from the introduction of a new reimbursement scheme for hospitals (DRG-based) at the micro level, to the implementation of fiscal federalism at the macro level. We argue that all these legislative interventions rendered producers more prone to control costs, in order to be able to meet financial constraints. Our empirical findings support this hypothesis: average cost inefficiency, computed after estimating a Translog stochastic cost frontier by the Battese and Coelli (1992) approach, reduced remarkably during the observed years, while the adoption of more effective technologies shifted upward the minimum-cost frontier. We also find evidence of relevant scale economies not exploited by producers, which decrease with hospital size (number of beds) and output complexity (average DRG weight).

0357 Modeling the choice of medical specialty Gonzalez Lopez-Valcarcel B. (University of Las Palmas de GC, Spain) Vazquez-Polo FJ., Barber P. Mechanisms for matching supply and demand of medical residence positions vary among countries. From the micro perspective of young physicians who are deciding their future, medical specialization is determined by several attributes of the available positions. These attributes influence the individual utility function. They include characteristics of the hospital: “prestige”, technological level, case-mix complexity- as well as characteristics of the medical specialties – income and professional perspectives-. We estimate a choice model for the medical residence positions: what specific hospital and medical specialty has each candidate chosen- with micro data of about 15000 physicians from Spain in three consecutive years (2003–2005). The model shows that the geographical location of the hospital and its distance from the candidate’s residence is a major determinant of the physician choices. Income expectations in private medical practice have a significant influence in the choice of medical specialty, as well as the “technological level” of the specialties. Despite in theory there is a national market of physicians covering the whole country, our model suggest that this market is segmented regionally.

0162 Cost-effectiveness thresholds and pharmacoeconomic evaluation – limitations in clarity and guidance Gothe H. (IGES Institute for Healthcare and Social Research, Berlin, Germany) Seidlitz C., Mank J., Caeser M., Häussler B. Objectives: To review the literature for information about the impact of pharmacoeconomic data and cost-effectiveness thresholds on decision making processes in Australia, Canada, England, and Sweden. Methods: The websites of the institutions, responsible for the economic evaluation and recommendation of medicines, were searched for topically relevant terms such as reimbursement, pricing, threshold, pharmacoeconomics, formulary listing. Results: From the document search we obtained a total of n=7,717 hits for the selected search terms. In n=599 documents terms were comEur J Health Econom Suppl 1 · 2006

| S39

bined with other key words (i.e. “pharmaceutical”, “threshold”), content analysis showed that only n=38 documents were relevant for the study aims. None of the countries has published an official cost-effectiveness threshold, but it appears that an implicit threshold exists (e.g. 30,000 £ in UK). Concerning the recommendations of several medicines, publicly available information is heterogeneous between the four countries either with regard to the detailedness or availability. Conclusion: Extensive material and data can be identified but it does not necessarily provide reliable guidance. The interpretation of the findings may be difficult because of the heterogeneity of the available material. In order to guide submissions and to enhance assessment procedures authorities need to be more explicit about criteria for decision making, including cost-effectiveness thresholds.

0159 Cost-Effectiveness of Rosuvastatin in the Prevention of Ischemic Heart Disease in Portugal Gouveia Pinto C. (Technical University of Lisbon, Portugal) Carrageta M., Miguel L.S. Objective: To assess the cost-effectiveness of rosuvastatin compared to ator vastatin in the treatment of hypercholesterolemia and prevention of ischemic heart disease (IHD) in Portugal. Methods: A probabilistic Markov model was developed to analyse the costs and consequences of lifetime treatment with each statin. Results from head-to-head, randomised, double-blind trials evaluating low-density lipoprotein (LDL) changes and from a meta-analysis defining the relationship between LDL levels and fatal and non-fatal IHD events were combined. Incidence of myocardial infarction (MI) was derived from a 9-year Portuguese observational study. Resource use in the treatment of MI was estimated by a Delphi panel of 8 Portuguese cardiologists with at least 15 years of clinical practice. Costs were calculated in both the societal and the patients’ perspectives. Results: Rosuvastatin increases life expectancy in 5.63 days per patient. Moreover, it saves 1,051 € (622 €) per patient in the society’s (patients’) perspective. The incremental cost-ef fectiveness ratio is –68,107 € (–40,296 €) per life year saved according to the society’s (patients’) perspective. In the 10,000 simulations carried out rosuvastatin was always more efficacious, being cost saving in more than 9,800 cases for both perspectives. Conclusion: Rosuvastatin is a dominant alternative in the prevention of IHD in Portugal.

0415 Socio-economic determinants of national suicide rates Garvey E. (NUI Galway, Ireland) Kennelly B., Oea E. This paper analyses the socio-economic determinants of national suicide rates. Panel estimation techniques have been used with international data and national regional data to test the effects of a variety of social and economic variables on suicide rates. This paper tries to improve on the specifications used in previous models. We take account of the time series properties of the data in our empirical work and estimate both a long- and short-run model based on the results of these properties.a time series framework for panel data, we tested the main variables for (non-) stationarity and categorised the variables as being either I(1) or I(2). We estimated both a short term model with purely stationary variables as well as a long term cointegration relationship. In the short term model we found a positive relationship between the change in alcohol consumption and changes in male suicide rates and a negative relationship between acceleration of female labour force participation and change in male suicide rates.the long term model, we found that it was plausible that the four designated I(1) independent variables (Gini,

S40 |

Eur J Health Econom Suppl 1 · 2006

unemployment, GDP per/capita and alcohol consumption) were cointegrated with suicide rates. The results of the cointegrating model suggest that income inequality is an important determinant of suicide for men in OECD countries, as is alcohol consumption. Unemployment is an important determinant for female suicide rates and GDP per capita is important for both sexes.

0063 Region, health insurance premiums and risk adjustment: a five country comparison Gress S. (Department of Economics, Institute for Health Care Management, University of Duisburg-Essen, Germany) Beck K., Behrend C., Schokkaert E., Shmueli A., Vliet R., Voorde C., Wasem J. Research questions: This paper aims to answer the following questions: What are the consequences of forbidding or allowing regional premium differentiation by sickness funds? What are the effects of disregarding or capturing differences in regional health care expenses in a risk adjustment formula? In how far are regional premium differentiation by sickness funds and regional differentiation in a risk adjustment formula interdependent? What are the advantages and disadvantages of divergent regulatory regimes with regard to the questions above? Methods: The paper is based on the conceptual framework of socially “acceptable costs”, which is applied to the institutional settings and regulatory regimes of five countries (Belgium, Germany, Israel, the Netherlands and Switzerland). Results: The societal choice on what variables to include in a risk adjustment formula poses a fundamental trade-off between efficiency and equity. We found that this is especially true for the risk adjustment variable “region”. Our five-country comparison shows that the explicit or implicit choices on accounting for differences in regional health care expenses vary. The implications of these choices depend on institutional features of the regulatory regime, including the instruments available to sickness funds to manage care and the implementation or not of a health-based risk adjustment formula.

0592 Investing in Migration Health in Europe: For the Benefit of All Grondin D. (International Organization for Migration, Migration Health Services, Geneva, Switzerland) Investing in and managing migration health enables economic progress. Migrants in good health tend to enjoy an improved quality of life and raised life expectancy. Generally, they are more receptive to education and more productive, thus increasing national income and encouraging foreign investment. Enjoying physical, mental and social wellbeing means migrants can more easily seize the opportunities to participate positively in their communities. Integration policies that encourage migrants to strive for educational attainment, employment, and community involvement can further amplify these gains. A failure to recognize these benefits could lead to a higher level of disease among migrants, including mental and social ill being, reduced individual incomes, and slowed national output. Despite these facts, the management of migration health is still seen as expenditure rather than an investment. By investing in migration health nations and policy-makers worldwide can create a positive synergy that will magnify social and economic gains for all.

0240 Efficiency of procurement procedures for medical equipment Guccio C. (University of Reggio Calabria, Reggio Calabria, Italy) Pignataro G., Rizzo I. The paper investigates the efficiency of procurement for medical equipment. Differentiation of products in this market, due to doctors’ preferences for specific technologies, might reduce the number of suppliers and the potential positive impact of auction relative to negotiation. a large official dataset on technological equipment purchased by Italian hospitals in the period 1996-2003, we run an empirical analysis of the differential impact of procurement procedures on the purchasing price controlling for several factors like product differentiation, characteristics of purchasers, concentration of suppliers. The policy implications stemming from the paper point to stress whether incentives are needed for managers and doctors when auctions are not satisfactory means to ensure efficiency in procurement.

0299 Deficits, soft budget constraints and bailouts Hagen T.P. (Institute of Health Management and Health Economics of Oslo, Norway) Tjerbo T. Background: The Nor wegian hospital reform of 2002 transferred the ownership of the hospitals from the counties to the central state, and created five regional semi-autonomous health enterprises with the responsibility of governing the hospitals in their region. The hospital reform had elements of political centralization since the Parliament became the political body ultimately responsible for the decision making, but also of administrative decentralization as more decisions were expected to be made by non-political boards and directors at the regional and local levels. There were several arguments for reform. However, what triggered the reform were increased deficits at the county level in the last part of the 90s and demand for supplementary funds from the parliament to the hospital sector during the fiscal years. Question: Has cost control improved after the reform? Are there differences between the five regional health enterprises in budgeting procedures and deficits after the reform? How has the Parliament eventually responded to demand for supplementary funds after the reform? Theory: Classic contributions in the literature of fiscal federalism point towards decentralization as a mean for achieving allocation efficiency and cost control. Recent contributions do however indicate that the state-local institutional arrangements such as tax discretion and budgeting procedures affect both efficiency and cost control. Methods: Budget and accounting data, qualitative inter views. Results (preliminary): Regional health enterprises have strong incentives to increase production without caring for the monetary costs of the production increases. The regional health enterprises expect that a deficit will be balanced by intervention from the central state, simply because the political costs of not bailing out outweigh the costs of standing firm, at least in the short run.

0013 Who’s Going Broke? – Comparing Healthcare Costs in Ten OECD Countries Hagist C. (Freiburg University, Freiburg, Germany) Kotlikoff L.J. In the framework of this study, a demographic model is used to estimate the benefit level growth of public health care expenditure in ten OECD countries (Australia, Austria, Canada, Germany, Japan, Norway, Spain,

Sweden, the UK and the U.S.). We show that except for Japan, demography was not a major driver of public health care expenditure from 1970 to 2002. Furthermore, our estimation shows that the benefit level growth rates differ significantly between the ten observed countries. By applying different population projections for each country to our model we estimate the future fiscal burden of public health care systems. We conclude that unsustainable levels of public health care expenditure will occur in case of slow policy responses and in case that public health care expenditure grows as it did in the past three decades.

0242 Down (Under) in the Dumps: Incidence of Clinical Depression in Australia Haisken-DeNew J. (RWI Essen, Germany) In most western cultures, the incidence of clinical depression is a difficult social issue to deal with. As the causes for depression are often difficult to identify, an empirical analysis with rich family and job information will allow more insight into this complex issue. Using three waves of the Australian household panel sur vey HILDA, detailed information concerning not only overall life satisfaction but also depression is employed to identify the incidence and determinants of depression in Australia. The HILDA identifies symptoms of clinical depression separate from life satisfaction, allowing one to quantify the association between the two measures of well-being. In addition to standard controls, this paper examines triggers of depression, such as shocks to oneself or family members with respect to income, labour market status (firing, promotion), health (injury, death), and family status (i.e. separation, divorce, birth). Due to the panel nature of the micro data, a clear separation of individual unobserved heterogeneity and exogenous variables in the model can be made. This analysis identifies financial worsening, marital separation, death of spouse or child, being a victim of crime or violence as being the significant triggers for depression. Depression itself, as defined by being in the lowest 5% of the mental health distribution greatly reduces life satisfaction by as much as 0.5%-points on 0 to 10 scale. This is equivalent to negating any positive benefit of marriage on life satisfaction.

0447 Comparison of EQ-5D and SF-6D utilities in mental health patients Hakkaart L. (Institute for Medical Technology Assessment (iMTA), Rotterdam, the Netherlands) Lamers L.M., Bouwmans C.A.M., Straten A., Donker M.C.H. Abstract preference-based measures of health are available for use as an outcome measure in cost utility analysis. The aim of this study is to compare two such measures EQ-5D and SF-6D in mental health patients. data from a Dutch multi-centre randomised trial of 616 patients with mood and/or anxiety disorders were used. Mean and median EQ5D and SF-6D utilities were compared, both in the total sample and between severity subgroups based on quartiles of SCL-90 scores. Utilities were expected to decline with increased severity. EQ-5D and SF6D utilities differed significantly between patients of adjacent severity groups. Mean utilities increased from 0.51 at baseline to 0.68 at 1,5 years follow-up for EQ-5D and from 0.58 to 0.70 for SF-6D. For all severity subgroups the mean change in EQ-5D utilities as well as in SF-6D utilities was statistically significant. Standardised response means were higher for SF-6D utilities.concluded that both EQ-5D and SF-6D discriminated between severity subgroups and captured improvements in health over time. However, the use of EQ-5D resulted in larger health gains, which will be translated into lower cost-utility ratios, especially for the subgroup with the highest severity of mental health problems.

Eur J Health Econom Suppl 1 · 2006

| S41

0180 Aging, health expenditure, proximity of death and income in Finland Häkkinen U. (Stakes, CHESS, Helsinki, Finland) Martikainen P., Noro A., Nihtilä E., Peltola M. The aim of the paper is to evaluate how total health expenditure and its main components (somatic care, non acute inpatient care, long term care and prescribed medicines) are related to age, gender, household income and proximity of death. In addition, we will explore how accounting for age as well as proximity of death affects projections of health expenditure. a data source, we used a 40 % sample of the Finnish population aged 65 and over at the end of 1997 (N = 285 317). Followup for death, hospital and medication use was until the end of year 2002. With an individual level unique identification code we linked data from Statistics Finland Population Registration, Finnish Hospital Discharge Register, Finnish Death Register, registers of the Social Insurance Institution and data from the Finnish Hospital Benchmarking Project.each person in our sample we calculated health expenditure in 1998, which is used as the dependent variable. As independent variables we used age, gender, time to death, death (1, if an individual died prior to the end of the year 2002), and family income (divided by OECD consumption unit scale) and significant interactions of the variables. We estimated a two part model (logit + OLS) for health expenditure. A SUR (seemingly unrelated) estimation method was used in the second part for the main components of health care expenditure we analyzed. In addition, we estimated separate models for individual receiving long term care (LTC) services and non-LTC services. These results were compared with a naive estimation, which does not take into account the proximity of death.to death was an important determinant of health expenditure. LTC expenditure increases clearly with age among both the sur vivors and the deceased. However, its effect is much lower than that estimated using a naive model. On the other hand, the opposite is true for other components of health care. Especially in somatic healthcare and prescribed medicines health expenditure seems to decrease with respect to age. income is negatively related to LTC expenditure, but the association with non-LTC expenditure is negligible. Only in use of prescribed medicine we find a clear positive relationship between income and expenditure.also projected health expenditure for individuals aged over 65 to the year 2036. According to our results a model which accounts for proximity of death gives about 25 % lower expenditure estimates than a naive model.

0597 Hospital financing and efficiency in the Nordic countries Häkkinen U. (Stakes, CHESS, Helsinki, Finland) Kittelsen S., Linna M., Magnussen J., Medin E., Nikolajsen J., Rehnberg C., Søberg Roed A. Any empirical analysis of the effect of a financing regime for hospitals must have observations with and without such a regime in place, while at the same time accounting for other differences facing the hospitals. Even with observations before and after a reform in the financing system in a single country, other concurrent changes e.g. in technology, institutions or demand may also affect hospital behaviour. Our analysis uses the fact that the Nordic countries in may ways are broadly similar in its culture, living standards and institutions of the health care sector, but have had different financing regimes for hospitals and changes of the regimes at different points of time. In particular, DRG based financing arrangements are present in some but not all Nordic countries and with different intensities. Data are collected for harmonised definitions of inputs and outputs for a panel of about 200 hospitals in each year and Data envelopment analysis is used to calculate Farrell efficiency estimates and Malmquist productivity change indexes. We present

S42 |

Eur J Health Econom Suppl 1 · 2006

the structural similarities and differences of the Nordic systems over the time period 1999-2005, and use these as controls in the analysis of the effect of financing regimes on hospital performance.

0319 Mortality and income in Denmark Hansen K.S. (University of Aarhus, Institute of Public Health, Department of Health Services Research, Aarhus, Denmark) Enemark U., Foldspang A. A 10 percent sample has been drawn among the inhabitants of 4 counties of Denmark corresponding to 160,000 individuals. Using administrative registries, information for each individual with respect to different income measures and a wide range of personal characteristics including education, family status, type of job, employment status, consumption of various health ser vices and immigration status for the period 1990-2003 has been extracted. This data base will be utilised to explore the association between income and mortality using different regression analyses and controlling for relevant characteristics. In particular, a number of competing hypotheses concerning what is the most important income measure will be tested such as absolute income, relative income, income inequality or deprivation. Finally, the association between family income and child mortality controlled for various family characteristics will also be investigated.

0589 Privatization of Hospitals in ECA Countries Harding A. (World Bank) Much has been written about the role of the private sector during the transition to a market economy in ECA countries. In reality, although there has been some outsourcing of some services and other forms of private sector participation, most hospitals remain public or corporatized public enterprises. This paper will look at early experience in privatizing hospitals in the ECA region.

0522 Patient access to primary care during out of practice hours- Organisation and incentives in Europe Hartmann L. (University of Lille, France) Ulmann P., Rochaix L. This paper focuses on patient access to primary care during out of practice hours and for patients living in under-served geographical areas, on the basis of a European health care systems’ appraisal. We try to identify the conditions of allocative and productive efficiency, and horizontal equity of OOH ser vices’ delivery, in a framework where demand for these ser vices tends to increase (ageing, chronic diseases…) and supply to decrease (working time …). Relying on the experience of six European countries (France, Germany, Italy, United-Kingdom, Spain, Sweden), we study how access rules to health care and working conditions of the GPs’market have an impact on efficiency and equity of out-of-hours (OOH) services’ delivery. Moreover, we study the designing of OOH services (required or optional duties, subcontracting, delegating…), their relevance and their limits. In each country, the recent reforms of primary care seem to set up a common model of OOH services’ delivery – but with different degrees of development: the main features are the centralisation of calls towards a nurse’s triage (vertical organisation of care), the introduction of emergency primary health care centres and the development of tools for improving care continuity.

0235

0163

Economic evaluation of ingrown toenail surgery in Scotland

The causal effect of socioeconomic characteristics in health problems across Europe: a longitudinal analysis using the European Community Household Panel survey

Graham B. (NHS National Services Scotland, Edinburgh, Scotland) Dench H.C., Russell I.T., Thomson C.E. Aims calculate the cost effectiveness of surgery for ingrown toenails performed by podiatrists in the community setting compared to surgeons in the hospital setting, and the economic impact of moving all cases from their current setting to the more cost effective setting.full economic evaluation was carried out alongside a randomised controlled trial. Outcomes were measured between the two groups by means of a change in the Euroqol EQ-5D value between baseline and at a two month follow-up period. Costs were analysed from the viewpoint of the NHS in Scotland and the patient using 2001/02 values. for ingrown toenails was consistently more cost effective when carried out by podiatrists in the community setting than by surgeons in the hospital setting. If the 787 patients operated on as hospital patients in 2004/05 had been treated by podiatrists in the community setting, the analysis presented suggests that there were over €;854,000 worth of savings to NHS Scotland alone that could have been realised, whilst still delivering the ser vice to the patients. analysis provides compelling evidence that the potential savings are so great that these procedures should be transferred without delay to podiatrists to carry out in the community setting.

Quevedo C.H. (DERS, University of York, UK) Jones A.M., Rice N.

0460

The main aim of this paper focus on analysing the dynamics of the socioeconomic gradient in two binary indicators of health limitations across the European Union. For that purpose, we exploit the longitudinal nature of the European Community Household Panel (ECHP) dataset. We are interested in whether, and to what extent, socioeconomic characteristics as household income and education attainment affect health problems and how this varies across time and countries included in the ECHP, considering the differences between age and sex groups. This paper uses the European Community Household Panel Users’ Database (ECHP-UDB). The ECHP-UDB is a standardised annual longitudinal survey, which provides 8 waves (1994–2001) of comparable micro-data about living conditions in the European Union Member States (EU-15). Our analysis focus on two binary measures of health limitations, constructed from the answers to the question: “Are you hampered in your daily activities by any physical or mental health problem, illness or disability?”, included in the ECHP-UDB. The estimation of dynamic panel data models raises some empirical challenges that are addressed in our analysis: the existence of correlated individual effects, the initial condition problems and the existence and consequences of health-related attrition in the sample.

Comparing pharmaceuticals’ costs among countries: A Proposal of a Relative Price Index for Central America

0263

Hernandez-V J.A. (Univerisity of Costa Rica, San Pedro de Montes de Oca, San José) Xirinachs Y., Jara M. Comparing prices among different countries has always been source of many controversies as it is difficult to mix homogeneity with representativity when constructing the index and selecting the goods’ basket, especially when it comes to be pharmaceuticals. It is difficult to find out the best information to do both: use simple and conventional methods and reduce the sources of price index bias. The objective of this paper is to propose a methodological variant of a relative price index method that allows the use of a wide and representative sample of pharmaceuticals for different countries of the same region, in this case Central American countries. The proposed price index uses a sample of pharmaceuticals from a quarterly IMS survey during the period 1999–2002. The survey provides information on products sold in private drugstores in Guatemala, El Salvador, Honduras, Nicaragua, Costa Rica and Panama. 2089 different presentations of pharmaceuticals are picked from the 10510 pool to be pharmaceuticals of general use, sold at every time at any of the countries, and were weighed by their defined daily dosages (DDD). The proposed method allows the use of a wider pharmaceuticals’ sample from the pool and shows less dispersion of prices among the different countries than using a homogeneous sample. The survey shows Nicaragua and Honduras to be the cheaper countries, Guatemala the most expensive country and Costa Rica to follow an average tendency. It also shows how Panama holds and even reduces the pharmaceuticals prices, passing from being the most expensive country at the period’s beginning and almost at the average at the end while El Salvador passes from being around the median to be the second most expensive country of the region.

Direct and indirect costs of smoking in Germany Hessel F. (Institute for Health Care Management of Duisburg-Essen, Germany) Wegner C., Anker G., Wasem J. Objective and Methods: The health care costs and the costs of productivity loss due to smoking in Germany were estimated. Data on smoking behaviour were taken from German microcensus and combined with attributable risk fractions of the US cancer prevention trial and standard unit costs for Germany. Results: The annual direct medical costs were as 7.7 billion EUR mainly due to coronary heart disease (57%), respiratory diseases (27%) and cancer (16%). The annual indirect costs of smoking were estimated to be 4.3 billion EUR according to the friction cost method respectively 14.5 billion EUR using the human capital approach. Large differences between the two approaches are seen concerning mortality and long time inability to work. The by far highest costs according to the friction cost method incurred for work off days (4.2 billion EUR) as the costs for mortality, long-time inability to work and work off days were in a comparable range according to the human capital approach. Conclusions: Although there is no doubt about the relevance of the health damage and the productivity loss associated with smoking the estimated direct and even more the indirect costs of smoking differ due to the chosen methodological approach.

0265 Is there a chance for the QALY? Ethical aspects concerning the QALY concept Hessel F. (Institute for Health Care Managementof Duisburg-Essen, Germany) The question of a perfect allocation of health care is a crucial task for ethics as well as health economics. The most well-known concept to compare medical technologies for different indications and from differEur J Health Econom Suppl 1 · 2006

| S43

ent sectors of the health care system is the concept of quality-adjusted life years (QALYs). Based mainly on utilitaristic assumptions it is not only criticized for its methodological limits but also from an ethical and philosophical point of view. In this paper the following three basic ethical questions are identified and the possibilities of adapting these ethical aspects in the methodological framework are discussed (1) Do the numbers count? Is a maximization of the aggregated utility more important than the individually equal right for health care? (2) Fairness and goodness? Is it possible to combine the maximization of the aggregated utility with an equal distribution of health care? (3) Priority to the worse off? How to include the rule of rescue in a concept for maximization of the aggregated utility?might be one of the future challenges in health economics to increase the dialogue between ethics and economics and to consider these ethical aspects in recommendations for resource allocation in health care.

0563 Universal Coverage in a Mostly Private System: The Case of the USA Richard H. (University of Michigan, USA) Uniquely for the U.S., insurance for patients with kidney failure is provided under a universal, publicly funded program, in an environment of privately dominated, fee-for-service medicine. Surprisingly, the U.S, is ranked roughly in the median level of per-capita ESRD expenditures for ISHCOF countries, but achieves relatively good outcomes. ESRD patients, regardless of age, are entitled to coverage under Medicare, the federal health insurance program that serves primarily the elderly. Because the Medicare program is relatively more influential (a near single buyer) in the market for ESRD-related ser vices than it is for most other medical ser vices, price controls have been more stringent than elsewhere in the US health care system. Nonetheless, costs have grown substantially due to increases in both the prevalence of ESRD and in the volume of ancillary services provided (e.g., injectable medications) and the number of privately owned dialysis facilities has continued to grow along with the patient base. Overall treatment costs are also controlled to some extent by the relatively high rate of transplantation compared to other ISHCOF countries as ESRD patients with transplants have substantially lower costs than dialysis patients after the year in which the transplant procedure occurs. Proposed payment reforms in the US include bundling a broader range of dialysis-related services into a prospective payment system, developing models to case-mix adjust payment rates, and financially rewarding providers whose patients have better than expected outcomes (“pay for performance”).

0226 Which factors do individuals find important when choosing between primary care models? A discrete choice experiment Hjelmgren J. (The Swedish Institute for Health Economics, Lund, Sweden) Anell A., Nordling S. Objective: To examine which factors individuals find important when choosing between primary care models. Method: A questionnaire survey was sent to 1 600 individuals in Sweden. In the questionnaire, different primary care models were constructed on the basis of five attributes: choice of general practitioner (GP) or primary care team (PCT), waiting time and patient fee per visit, freedom to chose a provider, and degree of influence over the care received. Individual choices were correlated with age, gender, education, and health status (EQ-5D). Results: The response rate was 58 %. Important factors in the choice between models was: waiting time, patient fee, freedom to chose a provider, and the degree of influence over the care received (p<0,001).

S44 |

Eur J Health Econom Suppl 1 · 2006

Subgroup analysis also showed that older individuals and individuals with poor psychical health prefer an “own” GP whereas individuals who lived far from a hospital preferred a PCT. The most important attribute was “the degree of influence over the care they receive” (WTP SEK 229) followed by freedom to choose a provider (WTP SEK 161 per visit) and shorter waiting time (WTP SEK 112 per day).

0042 A health warning about modelling health production across countries. Hollingsworth B. (Centre for Health Economics, Monash University, Melbourne, Victoria, Australia) Spinks J. The production of health in a population is a multifactorial and complex issue. Consideration of the socioeconomic determinants of health in cross-country economic analyses has previously been somewhat limited by the lack of available data and suitable analytical techniques. The collation of more comprehensive cross-country databases in recent years provides an opportunity to apply ‘cutting-edge’ techniques to analyse the technical efficiency of health production at a country level. It has been proposed that cross-country comparisons of the technical efficiency of health production, estimated using such a technique, data envelopment analysis (DEA), have useful applications for policy makers. In theory such an analysis utilises measures of the socioeconomic determinants of health relevant to all social policy, not just health policy. Using OECD and WHO data, this paper proposes a conceptual framework, and critically analyses a number of outstanding theoretical questions regarding the use of DEA in this setting. It concludes that until such questions are addressed, the resultant implications for policy will be based on misleading information.

0461 Technical efficiency of health production and the socioeconomic determinants of health in OECD countries Hollingsworth B. (Centre for Health Economics, Monash University, Melbourne,Victoria, Australia) The production of health in a population is a multifactorial and complex issue. Consideration of the socioeconomic determinants of health in cross-country economic analyses has previously been somewhat limited by the lack of available data and suitable analytical techniques. The collation of more comprehensive cross-country databases in recent years provides an opportunity to apply ‘cutting-edge’ techniques to analyse the technical efficiency of health production at a country level. It has been proposed that cross-country comparisons of the technical efficiency of health production, estimated using such a technique, data envelopment analysis (DEA), have useful applications for policy makers. In theory such an analysis utilises measures of the socioeconomic determinants of health relevant to all social policy, not just health policy. Using OECD and WHO data, this paper proposes a conceptual framework, and critically analyses a number of outstanding theoretical questions regarding the use of DEA in this setting. It concludes that until such questions are addressed, the resultant implications for policy will be based on misleading information.

0215 Negotiations with the Pharmaceutical Industry – Lessons Learned Hollo I. (European Investment Bank, Luxembourg) In 2001, 2002, 2004 and 2005 the subsequent Hungarian administrations concluded agreements with the representatives of the pharmaceutical industry. The aim of the deals was to limit the outlays of the Health Insurance Fund (HIF) and to reduce the burden of patients (co-payments) by influencing the ex-manufacturer prices of pharmaceuticals. The outcome of these efforts is known: continued increase of pharmaceutical spending of the HIF and further increase of co-payments. The presentation will analyze the agreements and conclude lessons learned from this experience.

0185 Incentives for Exerting Effort and Patients Satisfaction Holmas T.H. (University of Bergen, Dep of economics, Bergen, Norway) Kaarboe O. We test if a demand response exists in the Norwegian primary care sector. The Nor wegian physicians are remunerated either by salary or by incentive contracts, and we have access to a large data survey that allow us to study the relationship between consumer satisfaction with primary physician services and the way physicians are paid. When a demand response exists we expect that patients are more satisfied when visiting a contract physician. As expected we find very small effects of the salary physician density on reported patient satisfaction in municipalities with excess demand. In municipalities without excess demand we find a negative association between salary physician density and patients satisfaction with their physician.

ment literature, which measures the productivity of similar production units to cover interval production data. Interval data represent a direct way to handle data uncertainty and problems with data comparability which seem particularly prevalent concerning health care applications. We illustrate the “mechanics” of the suggested method using a simple 1-input-2-output model on health care data from the OECD health statistics comparing the performance of the OECD health care systems.

0578 A framework for describing and classifying fourth hurdle decision-making systems Hutton J. (Department of Economics, Curtin University, London, UK) Objectives: to achieve better understanding of the potential for use of HTA in decision-making on the reimbursement of drugs and other technologies, by developing an analytical framework for system description and classification. Methods: key aspects of 4th hurdle systems were identified at 2 levels – policy implementation and individual technology decision. Characteristics of systems were grouped under 4 main headings – objectives, use of evidence, decision processes and accountability. A pilot application was carried out for 8 countries using publications, websites and expert opinion as data sources. Results: the framework was broad enough to encompass the main issues, but the proportion of information available to complete the framework from published sources was less than 50%.Conclusions: the framework will be helpful to researchers and policy-makers in understanding the nature and purpose of different 4th hurdle systems. Given the need for subjective input for its completion, more experience will be needed to judge whether the framework will provide the basis for more formal comparison of systems.

0388 0019 Investing in dental health in Australia

Anticipatory behaviour prior to the introduction of a prospective payment system for hospitals in England

Hopkins S. (Dept of Economics Curtin University, Perth, Australia) Kidd M.

Ikenwilo D. (Health Economics Research Unit, University of Aberdeen, UK) Farrar S., Amaya-Amaya M., Sutton M., Chalkley M. Yuen P.

Dental ser vices in Australia are available both privately and publicly. Access to public dental ser vices is free, subject to a considerable waiting period and is restricted to social welfare beneficiaries. Access to private services for private health insurance policy holders is subject to a co-payment. Those with neither private health insurance nor access to free public ser vices pay the full cost for privately provided dental care. We consider the relationship between dental utilisation and health insurance using an ordered probit model and an endogenous insurance variable. We find that Australians with private insurance and/or high incomes have more frequent dental visits. There have been some changes in access of the social welfare beneficiaries since 2000 when a higher proportion of Australians purchased private health insurance. The restricted access creates a problem of dental health for those on low incomes and as a result there are differences in the dental health of high and low income groups in Australia.

In April 2004 a new prospective payment system for hospitals – Payment by Results (PbR): was introduced in England. This national tariff, based on average costs, is intended to increase activity, efficiency and quality and reduce waiting times. In 2003/04, tariffs were used for all hospitals for changes in activity in 15 Healthcare Resource Groups (HRGs) from a possible 550. In 2004/05, for most hospitals (non-Foundation Hospital Trusts – NFTs) the system was extended to 48 HRGS, whilst a self-selected group of Foundation Hospital Trusts (FTs) moved to predominantly tariff-based funding. hypothesise that the response to the incentives will have been different for the soon-to-be FTs and NFTs and between tariffed and non-tariffed activity. For instance, soon-to-be FTs may contain their activity prior to the introduction of PbR allowing more scope for future increases in activity funded by tariff. We identify a matched control group using propensity score matching of FT status on a range of published performance indicators in 2002/3. We compare trends between FTs and NFTs using a matched differences-in-differences design and trends in tariffed and non-tariffed HRGs. We consider changes in means and measures of dispersion of activity and costs and the influence of market concentration.

0101 Health Care Productivity with Uncertain Data Hougaard J.L. (University of Copenhagen, Denmark) Overgaard C. Data uncertainty is a huge problem comparing the performance of health care systems and/or various health care producers. We suggest to extend the FDH-method, known from the efficiency measure-

Eur J Health Econom Suppl 1 · 2006

| S45

0267 Pricing policies for the pharmaceutical market – an international perspective Ilgin Y. (Johann Wolfgang Goethe University, Frankfurt am Main, Germany) Eisen R. Governmental price regulation of drugs has become a popular measure to contain health care spending. The aim of our research is to evaluate the effectiveness of selected demand and supply side measures, keeping in mind crucial factors in this market, for instance the free rider problem and moral hazard. To analyze reference price limits on reimbursement, a static two class product model is introduced. Beside this measure additionally patents, parallel imports as supply side price regulation and drug budget for physicians and generic drugs as demand side price regulation will be reviewed. By using a sequential price-setting process within the model it can be shown that applying marginal cost pricing for drugs clustered within Phase 1, welfare can be increased. If government sets the reference price equal to the marginal costs welfare can be increased without free riding on the sunk R&D costs of researching pharmaceutical firms because the patent protection has expired. To give a comprehensive evaluation of the other pricing policies the interactions between regulative measures are taken into account. In the past too many regulative measures to contain health care expenditures have been targeted primarily to supply side measures or to demand side measures. But, regulation that only applies on one side does little to control the rising expenditures. Without simultaneous use of demand side incentives and volume controls, pharmaceutical expenditures probably can not be reduced effectively.

0523 Investment in health Ivády V. (Semmelweis University, Budapest, Hungary) Sinkó E. Aims: How can we express the impact of health care on the economy, and how is it reflected in the involvement of resources (the economic basis of the economic stimulating role of health care)? Spending on health: burden or investment: challenges for middle-income countries to consider health sector as a major determinant of their development? What impacts do the economic effects of the operation of the health care sector have on the formulation and implementation of the aims of health care policy? Methods: Review of Literature; Analysis of official data; Analysis of documentary papers Results: In the health sector the outcome of the allocation of the resources is split between health gains and other gains. The category of “other gains” should include such important factors as the improvement of the feeling of security of the people, improved equity in access to care, implementation of patients’ rights, the employment and satisfaction of workers in the health care sector. According to our assumptions at the lower level of economic development a higher ratio of the health expenditures goes to increase the health gains of the active age groups (expenditures with direct economic return) than at a higher level of economic development, where an increasing ratio of the rising health expenditures goes to increase the health gains of the inactive agree groups and to other gains (so-called equity or burden expenditures). From among the typical aims of health policy would like to highlight two objectives that all countries should adopt: advance in the field of “equity” and “efficiency.” If, in the field of the improvement of equity, Hungary intends to promote investments, i.e. if the country would primarily like to expand the qualification of the population for treatment, furthermore, if it intends to ensure the widest possible access to services, then we must expect an increase in burden expenses. Putting those measures in focus that are aimed to improve efficiency will justify investments resulting in the increase of health gains.

S46 |

Eur J Health Econom Suppl 1 · 2006

Conclusion: New theoretical framework for government involvement in investing health.

0219 Prospective payment system for Trauma inpatients, in Israeli hospitals Ivancocsky M. (The Gertner Institute for epidemiology and health policy research, Ramat-Gan, Israel) Shmueli A., Oberman B., Freedman L.S., Peleg K. Abstract in Israel are currently compensated for all patients by some function of length of in-patient stay (LOS) using a per-diem rate, but this may be incorrect for trauma patients. The aim of this work is to suggest a fairer and more equitable prospective payment system. 409 patients hospitalized between 998-2000 in eight trauma centers included in the Israel National Trauma Registry, formed the training set and patients hospitalized between 2001-2003 formed the learning set. In addition, actual patient costs were obtained for two of these hospitals, a non-parametric tree building technique, was employed to create a model with “equal resource utilization” groups (“ERUG”s). Costdays, a variable based on LOS, was used as the dependent variable and patient admission data comprised the independent variables. Learning models were used to validate the predictive ability. order to evaluate whether “costdays” is an appropriate surrogate for actual costs, two training models, one based on “costdays” and the other on actual costs, were compared. These models were also compared with the eight hospital model. “Costdays” was found to be a good surrogate variable for actual costs, this was shown consistently by performing a large number of simulations. ERUGs were created, each defining a unique profile of patients.

0175 A study of income-motivated behavior among general practitioners Iversen T. (University of Oslo, Norway) Background: We study the effect of a mixed capitation and fee-for-service system on the amount of services provided by physicians. In particular, we study to what extent the variation in service intensity among general practitioners (GPs) may be explained by an observation that some physicians have fewer regular patients than they would like to have. Theory: By means of non-linear programming a GP’s optimal practice style is derived. The theory predicts that a GP, who experiences a shortage of patients, is likely to increase the number of services he provides to his patients if the marginal utility of leisure is less than the marginal utility of income from the extra services. Empirical strategy: The sample consists of all GPs (3650) during the period 2000-2004 in the nationwide list patient system in Norway. Models are estimated by generalized least squares regression that takes unobserved heterogeneity and heteroskedasticity into account. Self selection is adjusted for by means of a Difference-in-differences estimator. Results: We find that patient shortage increases a GP’s intensity of service provision (in particular the length of each visit) with 10 – 15 per cent. Results are also valid when possible selection bias is accounted for.

0065 Composite performance measures for hospitals Jacobs R. (University of York, UK) Goddard M., Smith P.C. Composite indices of health care performance are an aggregation of underlying individual performance measures. They are increasingly being used worldwide to rank healthcare organisations. While they pres-

ent the ‘big picture’ in a way that is easy to interpret, misleading conclusions may be drawn if attention is not paid to key methodological issues in their construction. We illustrate how variability in the underlying data and the resulting composite may undermine the robustness of performance measures in health care. We use panel data for English acute hospitals from 2000/01 to 2004/05 based on the “star rating” data – a composite performance indicator used in the English NHS. We illustrate generic methodological issues that will arise in any context by creating a composite index from the underlying performance indicator data, and using Monte Carlo simulations to examine the robustness of the composite. The variance in performance measures is partitioned into ‘controllable’ and ‘uncontrollable’ elements. Random variation gives rise to considerable uncertainty in performance scores. Our results suggest some hospitals can jump enormously in the league table of performance as a result of subtle changes in the methods of creating the composite. Great care is warranted in interpreting the results of composite performance measures.

0305 Is informal payment as equitable as physicians claim? Evidence from the Kyrgyz Republic Jakab M. (World Health Organization, Bishkek, Kyrgyz Republic) Kutzin J. There is increasing evidence about various aspects of informal payment in transition economies. However, the distributional impact of informal payment has not yet been studied. Physicians often make the case that informal payment is equitable as patients pay based on their ability to pay. However, qualitative research speaks to the contrary. In this paper, we formally test whether physicians are effective agents of redistribution. We use a unique survey of patients discharged from hospital conducted in the Kyrgyz Republic in 2004 (n=4,533). The survey asks detailed questions about payment for admission, personnel, drugs during hospitalization as well as socio-economic status. The survey data was merged with the database of the Mandatory Health Insurance Fund to obtain further information about the hospitalization episode (e.g. DRG weight to control for severeity, type of acility, etc). We use a two-part model to analyze the determinants of informal payments. We find that controlling for a variety of factors, socio-economic status does not explain whether patients give informal payment and the level of payment. We hypothesize several reasons for this finding.

0201 Information Technology and its Impact on the Integration of Health Care Delivery: Experiences from California Janus K. (Hannover Medical School for Epidemiology, Social Medicine and Health System Research, Hannover, Germany) Amelung V.E. Information technology (IT) supports different modes of governance for the organization of health care transactions. Based on Williamson’s reasoning (1996) that IT facilitates the exchange of information in networklike organizations care delivery in hierarchies, such as completely integrated delivery systems, becomes less worthwhile. This raises the question whether the emphasis on health care provision in integrated delivery systems is based on a different reasoning or does not take the emergence of technologies into consideration.paper deals with the impact of IT on the organization of health care services. Based on a long-term case study the interaction between IT and the integration of players in the Californian health care market have been evaluated. An emphasis is put on the understanding of integration which implied mere aggregation in the past instead of actual integration supported by virtual integration.with senior management executives in the San Francisco Bay Area

revealed that this is not an “either-or” decision, but rather a stage-like process in which IT first acts as a technological facilitator and change agent and later has a psychological impact on integration. Together with the management of human resources IT represents a constant challenge to health care organizations in industrialized countries.

0118 Are safety promotion interventions for elderly also cost-effective? Johansson P. (Stockholm Centre for Public Health, Stockholm, Sweden) Per T., Siv S., Antonio P.L. Aims. Several pharmaceutical interventions against injuries among the elderly are considered cost-effective. As alternatives to pharmaceutical interventions seldom have been subjected to economic evaluations, we report a cost-effectiveness analysis of an elderly safety promotion intervention. A cost-effectiveness analysis is performed on Safe Seniors in Sundbyberg, a 5-year safety promotion programme against elderly injuries in the metropolitan area of Stockholm, Sweden. The effects are avoided hip fractures. The societal perspective is adopted, with hip fracture costs and quality-of-life consequences based on Swedish data. Total societal intervention costs amounted to 7.3 million SEK. The effect evaluation shows differing effects among population groups, with a net number of avoided hip fractures of 14 (0,44 per 1000). Consequences during the first year after a hip fracture are societal costs of 130.000 SEK and a loss of 0.17 QALYs. A first estimate, only including first year consequences, gives a cost-effectiveness ratio of 2.3 million per QALY. The full economic evaluation will also consider consequences during subsequent years, including long-term hip fracture consequences on costs, mortality and quality-of-life. The first crude cost-effectiveness estimates of the safety promotion intervention shows that it cannot be considered cost-effective, but further work will also include long-term consequences.

0274 Patient self-management of anticoagulation Jowett S. (Health Economics Facility, Service Management Centre, University of Birmingham, Edgbaston, Birmingham, UK) Bryan S., McCahon D., Murray E., Fitzmaurice D. Patient self-management (PSM) is a means of enhancing patient choice and autonomy. In the management of patients receiving warfarin for atrial fibrillation, regular patient monitoring is undertaken in a primary or secondary care clinic. The development of near patient testing means PSM is now feasible. A randomised controlled trial involving 617 patients (the SMART trial) compared PSM with routine care. The economic evaluation alongside the trial revealed PSM had a higher mean health care cost per patient (£417 vs £ 122), with little difference in clinical outcomes or quality of life followed over 12 months. However, resource use under trial conditions may not reflect actual use by patients carrying out PSM routinely. A further study was therefore undertaken involving 78 patients followed over 12 months. Whilst PSM again had a higher mean cost per patient (£299 vs £118), the difference was much smaller. The reduction might not change the decision in this case, but such a significant reduction in costs under non-trial conditions should alert both researchers and decision makers to the importance of assessing the impact of trial conditions on the results of evaluations. This study explores the extent to which this problem is likely to arise in other contexts.

Eur J Health Econom Suppl 1 · 2006

| S47

0120 The costs of informal care in neurological disorders in Spain Oliva J. (Univ Carlos III. SESAM/Department of Economics, Getafe (Madrid), Spain) Osuna R., Lobo F., Cabanas M., Jorgensen N. Informal care is frequently an invisible part of total care in terms of costs and effects over patients and caregivers. However, if the societal perspective is adopted, informal care needs to be identified, measured and valuated. The purpose of the present work is to identify, measure and evaluate the costs of informal care for neurological diseases in Spain. Data collected in the Survey on Disabilities, Impairments, and State of Health (EDDES, for its initials in Spanish) of the National Institute of Statistics (INE) was used to estimate the hours of informal care. We used several approach (opportunity costs and proxy good methods) for the monetary estimation of the hours of informal care. We distinguish between hours of caregivers employed and not employed. Informal costs estimated range between 2,402 and 2,926 millions of euros (at year 2002 prices), depending on the method used. The estimated costs of informal care of each neurological disease were stroke (823 to 1,007); dementia (1,021 to 1,246); Parkinson disease (329 to 401); and multiple sclerosis (229 to 272). The estimated informal costs represent a 6.3 to 7.7 per cent of the total health care costs of the Spanish National Health System.

0272 Determinants of Health Care and Informal Care Costs of Alzheimer disease in Spain Juan O. (Universidad Carlos III. SESAM. Department of Economics, Getafe Madrid, Spain) Osuna R., Bastida J.L., Serrano P. Objectivesaim of the study is to explain the total cost of the illness, by distinguishing public and private health care costs and formal and informal care. We also analysed the decision making process inside the household related to the time dedicated to formal and informal care.In this study we analysed the determinants of the costs of Alzheimer disease from a survey to 237 affected people living in Spain. The survey included an extensive set of questions about the use of health care ser vices, time of caregiving (formal and informal), socioeconomic characteristics of the patients and caregivers, health-related quality of life (HRQOL), disease severity (Clinical Dementia Rating scale) and Barthel questionnaire. stage of the disease and the HRQOL are the most significant variables in explaining the public health care cost. On the contrary, income household level is the main determinant of the private health care cost. The total amount of time dedicated to the assistance of the person affected by Alzheimer does not seem to depend on income, and this variable only determines the amount of formal but private care.

0328 The cost effectiveness of alternative referral strategies for suspected bowel cancer Juarez-Garcia A. (Health Services Mangement Centre, University of Birmingham, United Kingdom) Baker R., Stokes T., Thompson M., Shaw E., Camosso-Stefinovic J. Introductionguidelines on the symptoms and signs indicating the need for referral have been published in the UK, but little is known about the cost implications of the referral recommendations.cost effectiveness study analyzing three referral strategies were examined)age plus bleeding and change in bowel habit;)all patients in (a) plus other patients with an abdominal and/or rectal mass and iron deficiency anaemia;)this strategy was based on the assumption that all patients presenting any

S48 |

Eur J Health Econom Suppl 1 · 2006

symptom or sign of bowel cancer are urgently referred.epidemiological parameters were based on a UK study of 8000 patients referred to a surgical outpatient clinic in which there were 601 cancers. The cost-effectiveness measure was cost per early detected cancer case.(b) showed an incremental cost-effective ratio of £2446 per early detected cancer case compared with strategy (a). The incremental cost-effectiveness ratio for each extra early diagnosed cancer increased to £20,927 in strategy (c) compared with strategy (b). of patients for referral should be based on combinations of symptoms, the two physical signs and the presence or absence of iron deficiency anaemia.

0468 Pharmaceutical innovation, reference pricing and therapeutic classes Jullien B. (Idei and University of Rosario Bogota, Toulouse, France) Bommier A., Bardey D. This paper is a first attempt to model the effects of reference pricing on the innovation effort of pharmaceutical firms. The model is based on a dynamic game involving three types of agents: pharmaceutical firms, consumers and a regulatory entity. The game includes research stages where the innovation efforts by the firms are determined and introductory stages where a price for a new medicament is fixed. We model the negotiation between the drug owner and the regulator to fix the price, first without legal constraint, second under the regime of reference pricing in therapeutic classes. We then solve the innovation game where the firms anticipate the results of the negotiation round on prices. We thus consider the effect of the therapeutic class regulation on both prices and the innovation pace. The final stage consists in calibrating the model with a small data on statin in France and simulates the effect of the change in regulatory regime.

0537 Capacity building for mental health economics in Romania: A case study estimating the cost of relapse in people with schizophrenia – a one year audit Micluţia I. (Faculty of Political and Administrative Science, Babes-Bolyai University, Romania) Junjan V., Popescu C.A. Economic evaluations remain scarce in Romania and among these studies looking at mental health are even rarer. There are few attempts to use economic evaluation methods in academic circles, but there are increasing signals from health policy makers for more such studies. Relapse in people with schizophrenia it is hypothesised strongly increases hospital costs, thus demanding a greater share of scarce hospital budgets. This study examines the resource utilization and costs for patients with schizophrenia who experience of relapse and attend a psychiatric clinic in Cluj-Napoca in 2005. The evolution of insurance coverage for specific medications for schizophrenia is also examined. This is the first study of its kind in the country and some of the practical challenges in conducting such studies in a system where capacity for health economics remains limited are discussed as well as the way in which such studies can inform the decision making process. As such studies remain rare, especially in the field of mental health in Romania, they can play an important part in the development of methods and study protocols that might in future be conducted at the national level. They can be of use to local level health decision-makers in developing medium-and-long term strategies and informing budgetary negotiations.

0368 The impact of social determinants of health on health inequalities in France Jusot F. (IRDES Research and information institute for health economics, Paris, France) Dourgnon P., Grignon M. This paper looks at the role of social determinants of health in the explanation of social health inequalities in France in addition to well established socioeconomic dimensions, such as income, education and occupation.literature suggests that social factors, particularly social capital, social support and relative social position within a reference group, are important in the construction process of health inequalities. Nevertheless, the impacts of these factors on health inequalities are only studied in isolation so far, the respective impact of each determinant on health is not established.the 2004 Sur vey on Health and Health Insurance – which is the unique French dataset providing micro-data on health status of the general population, as well as a large number of social and economic determinants of health – we propose to assess the direct and interaction effects of social capital, social support and relative social position on health status, measured as self-assessed health. First, we analyse the correlation between each social determinant with the socioeconomic status as well as the correlation between the various social determinants of health. Second, we conduct a multivariate analysis of the link between each of these variables with self-assessed health, controlling for socioeconomic status.

0262 Relationship between the EuroQoL-5D and Barthel Index – Examining the use of proxy outcome measures for older people Kaambwa B. (Health Economics Facility, Services Management Centre, University of Birmingham, UK) Bryan S., Barton PM. Intermediate care (IC) of older people is a key component of UK government health policy. Evidence on outcomes associated with IC is scarce. In many instances, health related quality of life (HRQoL) outcome measures are not available mainly because older people are physically or mentally not able to (self-) report their HRQoL. The resulting missing outcome values may lead to biased statistical results. Proxy outcome measures may help but their suitability as proxies needs to be tested and verified. We considered outcome measures from a total of 1589 IC patients participating in the national evaluation of IC for older people in the UK. We examined the direct relationship between a conventional clinical scale of functional status that is suited for proxy-assessment (Barthel Index (BI)) and a measure of HRQoL (EuroQoL-5D). Using jackknife regression, multiple comparison tests and multiple imputation methods, we established a plausible and significant relationship between the measures. This relationship was valid across all diagnostic groups in the sample. There is considerable documentation supporting the construct validity of using the EuroQoL-5D in geriatric care and these results add to the evidence that BI scores may be used as proxy for HRQoL in case of missing values on EuroQoL-5D.

0100 The influence of hospitals eimbursement schemes on internal wage contracts Kaarboe O. (Health Economics Bergen, University of Bergen, Norway) Hehenkamp B. The starting point of the paper is the empirical observation that more and more payers include rewards for quality deals. These deals reward

hospitals financially if a certain quality standard is met. Often these financial bonuses take the form of increasing the reimbursement of all ser vices by a certain percent of the price the hospital gets if the standard is not met. question we ask in this paper is how such a reimbursement scheme affects the optimal contracts within the hospital department? Furthermore, given that the payer wants to implement the quality standard, what is the minimal transfer she has to provide to ensure that the hospital management provides wage contracts for the employees that include incentives for exerting effort on quality?answer these question is a two-period game with three players, the payer, the hospital and the physician. We characterize how the reimbursement system that the sponsor offers the hospital affects the wage contracts within the hospital. Furthermore we analyze how the minimal price that the payer provides to the hospital to ensure that the hospital choose to implement the quality standard vary over time.

0524 Economic influences on end of life decision making Kahana B. (Cleveland State University and Case Western Reserve University, Ohio, USA) Kahana E. Advanced directives and other aspects of end of life decision making are generally arrived at by members of a health care partnership. These partnerships include patients, physicians and family caregivers who share in the process of decision making. This poster will focus on the influence of economic status of elderly patients on end of life decision making. Data were obtained from 284 community dwelling older adults paricipating in a panel study of adaptation to frailty. (Kahana et al., 2003) Interviews were also conducted with primary care physicians and family caregivers of elderly respondents. Findings indicate that family caregivers were significantly more involved than were physicians in end of life planning. Economic status of respondents served as a major predictor of making plans for end of life care. Economic factors may differentially impact each member of the health care partnership triad in regard to end of life decision making. Understanding of divergent and convergent interests of constituent parties allows for better understandig of barriers to successful planning for end of life care.

0483 Does therapeutic reference pricing always result in cost-containment? The Hungarian evidence Kalo Z. (Novartis Hungary Ltd., Budapest, Hungary) Muszbek N., Bodrogi J. Therapeutic reference pricing (RP) is one of the potential cost containment methods for pharmaceuticals. The most critical question of RP is how to select reference product(s) if their efficacy is different, especially if different strengths of the same substance are available. Authors describe the Hungarian experience related to the introduction of therapeutic RP for statin therapies as of 1 September 2003. Therapeutic RP was expected to reduce the expenditure on statins by switching therapy to cheaper alternatives and therefore decreasing the average price per prescribed unit. The National Health Insurance Fund expected price erosion not only for branded products directly affected by generics but even for patented ones. Despite switching many patients from the original simvastatin to generics, the average unit price of statins was reduced by only 3% at 6 month after the introduction of the RP system. During the same period the average DDD per prescription was increased from 1.14 to 1.65. The price of patented statins was not changed over this period. Introduction of therapeutic RP in Hungary for statin therapies neglected evidence-based medicine results and ultimately increased drug expenditure. Selection of the cheapest DDD per unit as the reference product resulted in growth of DDD per prescription, and conEur J Health Econom Suppl 1 · 2006

| S49

sequently increased price per prescribed unit of statins. The failure of the system could have been even more dramatic if increased utilisation of generic statins had not reduced the negative effect of therapeutic RP. Extension of RP to therapeutic categories from generics is not justifiable based upon the first experiences of the Hungarian implementation.

0405 Drug prices and product differentiation Kanavos P. (London School of Economics, London, UK) Costa-Font J., Eagan L. Little evidence is available on the determinants of generic penetration in countries subject to significant market regulation. This paper empirically analyses the determinants of generic market entry in European countries, the USA and Canada, by using IMS data over the 2000–2005 period on a quarterly basis. The paper also analyses the extent to which generic penetration explains cross-country price differences for generic products in different countries. We find that country size and country specific fixed effects are the most significant factors in explaining market penetration rather than cross-country price differences.

0577 Cost-effectiveness assessment in the Hungarian drug reimbursement, 2004-2005 Kárpáti K. (Corvinus University of Budapest, Hungary) Brandtmüller Á., Gulácsi L., Dávid T., Lengyel G. According to the Hungarian laws health technologies subsidised by the Hungarian National Health Insurance Fund Administration (NHIFA) are required to be efficient and cost-effective. However, Hungarian health economics studies or adaptation of international results are rarely available. In many cases, the size of the target population isn’t known. In the absence of sufficient information and of a guideline assisting adaptations, HTAs carry considerable uncertainty. The effectiveness of the covered technologies should be verified on a regular basis, since the effectiveness of resource allocation should be improved continuously and transparency of public financing is a requirement. To achieve this, significant assessment capacity and reliable databases are needed. A more systematic approach and setting explicit criteria (e.g. significant budget impact, high prevalence) for prioritization of HTAs would help to use scarce research capacities. Development of patient registers and availability of patient level data from clinical trials are highly needed, too. In Hungary, the subsidization of pharmaceuticals amounted to 289 billion HUF in year 2004. After expert consultations, according to our study the NHIFA spent more than 536 million HUF on clinically obsolete medicines for diabetes. Another 1.720 million HUF were spent on non-effective locomotor disorder and varicose vein therapies and 2.228 million HUF on nootropics with limited evidence. The 6 agents and formulations we examined are subsidized on a smaller scale (0-40%) or they are considered to be nutricial supplements in 8 EU member states. However, they total up to 2% of the Hungarian Pharmaceutical Budget (with 50-70% reimbursement), which is more than the whole subsidisation of H2-receptor inhibitors.

0613 Measuring health care quality at the international level: Challenges and findings from the OECD Health Care Quality Indicators Project Kelley E. (OECD, Paris, France) Measuring health care quality at the international level: Challenges and findings from the OECD Health Care Quality Indicators Project

S50 |

Eur J Health Econom Suppl 1 · 2006

Objective/Aims: The objective of this presentation is to present the background and findings from the first phase of the Health Care Quality Indicators Project. The long-term objective of the HCQI Project is to develop a set of indicators that can be used to raise questions for further investigation concerning quality of health care across countries. It was envisioned that the indicators that were finally recommended for inclusion in the HCQI measure set would be scientifically sound, important at a clinical and policy level and feasible to collect in that data would be available and could be made comparable across countries. This presentation will summarize the findings of the HCQI Initial Indicator Report, released in March 2006 as OECD Health Working Paper 22. Method: The HCQI Project includes 23 countries in its initial report. The project has been divided into two phases. The initial phase, for which this report serves as the summary report, concentrated on 17 important and readily available indicators of effectiveness of care. All of the participating countries with one exception submitted data on at least five of these indicators and twelve of the seventeen indicators have data from 15 countries or more. The OECD carried out extensive data validation on each indicator and undertook detailed analysis on a set of five questions posed by HCQI country experts. These data-based questions focus on data comparability issues across countries on particular indicators: – What is the appropriate reference population for age adjustment? – What is the impact of different policies for handling missing data? – What is the impact of notification policies on cases of vaccine-preventable disease? – What is the impact of variation in coding practices (for asthma)? – What is the effect of unique identifiers when dealing with mortality rates? Result: The HCQI Initial Indicator Report presents detailed data on 13 indicators retained as internationally useable for raising questions on quality of care and on 4 indicators that are not currently useable in an international quality indicator set. Across the indicators, no one country does best or worst on any of the indicators and most countries exhibit areas of possible best practice. All of the indicators raise questions about causes of differences in quality of care across countries. Two examples of this will be highlighted for exploration, AMI 30-day mortality and breast cancer 5-year survival rates. Conclusion/discussion: It is clear that the assessment of health care quality is topic of concern across the OECD. Deriving a set of indicators that can be used reliably to assess differences across countries is a major step for ward. This presentation, and the paper on which it is based, offer data on country-based quality differences and suggest how such data should be used to understand international differences, identify best practices and improve quality of care.

0387 Effects of Tobacco Control Policies on Smoking Behavior Kenkel D. (Cornell University of Policy Analysis & Management, Ithaca, NY USA) Lillard D. We use data from the European Community Household Panel to estimate whether and how tobacco control policies affect individual decisions about smoking initiation and cessation over the life course. We use variation in country specific policies such as taxes, marketing restrictions, advertising restrictions, and public health information campaigns to investigate how those policies affect the probability that a person starts or stops smoking cigarettes. These policies vary across countries and over time within a country. We match policies to individuals in each country in each year under the assumption that the decision to start smoking occurred between the years a person was 14 to 18 years old. The analysis will inform policy makers of the sensitivity of smokers to tobacco control policies and will make it possible to estimate causal effects of those policies because we will be able to control for country fixed effects using temporal variation in policies.

0536 Willingness-to pay for mental health: a contingent valuation study Kennelly B. (Department of Economics, National University of Ireland, Galway) O’Shea A. This paper uses a contingent valuation cost-benefit approach to examine public preferences for the allocation of resources to mental health care in Ireland. The study uses a willingness to pay methodology to elicit citizens’ preferences for the allocation of additional resources to mental health care relative to other health care programmes. Contingent valuation offers an opportunity to obtain direct valuations of mental health outcomes by the general public through a survey-based methodology which asks people direct questions on their willingness to pay to expand services to achieve a given health state or health gain in mental health. People are first of all asked to rank specific health care programmes to improve health outcomes in three fields: mental health, cancer and hearts. The survey asks respondents about their willingness to pay additional taxation and/or voluntary contributions to fund health care interventions in the three areas. People’s willingness to pay for an expanded provision in mental health care is then compared to the other two programmes

0534 The economic evaluation of primary and secondary mental disorder prevention: A systematic review Kilian R. (University of Ulm, Department of Psychiatry II, BKH Guenzburg, Germany) There is an increasing economic burden linked to mental disorders in most European countries. Therefore the prevention of mental illness should be considered as a key task of European health policy. A systematic literature search has been conducted to identify relevant studies on the health economic evaluation of primary and secondary mental disorder prevention programmes. Only a few studies were found in scientific journals but a great body of research was published in the grey literature. The majority of programmes identified were targeted at particular groups with an increased risk of developing mental disorders such as economically deprived families, young single mothers, or families with a parent with a mental illness. The most rigorous empirical evidence was found for the cost-effectiveness of early childhood development programmes implemented since the 1960s in the USA. Several of these programmes have been evaluated by randomised controlled trials over more than 25 years and largely were found to produce economic benefits, particularly in the long run. Results of the review suggest that costeffective programmes for the selective prevention of risk factors of mental illness exist but that further research is necessary on the cost-effectiveness of programmes for the direct prevention of mental disorders.

vation decisions are closely related through firms’ innovation activities, their relationship is still left unexplored surprisingly. In this paper, by measuring their relationship, we unveil how firms’ strategic decision shapes the overall demand structure and the consumer welfare. Consequently, we empirically emphasize the active role of public policy in inducing and directing innovation and shaping social welfare, we focus on the pharmaceutical market where technological innovation, such as new drugs, has huge impact on consumer welfare. In this market, as a consequence, earlier successful innovation strategies have altered not only consumer welfare, but also market structure and even science base to a large extent. Accordingly, innovation strategies have been substantially changed over the decades in response. However, we remarkably know little about how important firms’ strategic innovation behavior is in shaping social welfare and how it is determined. In addition, firms’ strategic behavior and its effect on social welfare have been out of radar in public policy research. With social welfare improvement in mind, public policy can be categorized into active and passive approach depending on the targeting processes in innovation. The bulk of existing public policy has been following mostly the passive approach. Passive approach in public policy in the form of demand-side regulations and subsidies is executed to directly improve consumer welfare, especially in pharmaceutical market, by changing consumer characteristics. For instance, including a new drug in Medicaid coverage can be accepted as a passive one. Therefore, researchers evaluate patient welfare benefits arising from pharmaceutical innovation and, subsequently, suggest demand-side modifications to improve consumer welfare, active approach in public policy, which covers firms’ strategic behavior and its effect, has not been fully explored. Active policy, in the form of incentives and regulations for supply-side, is implemented to guide the behavior of decision makers and/or the overall direction of innovation and, ultimately, welfare. For instance, funding firm research for cures of neglected diseases can be viewed as an active policy. However, we have neglected long enough the fact that the strategic innovation behavior of firms is one of the major determinants of innovation directions and the outcomes of it. For example, we know little about what the determinants are for pharmaceutical firms’ strategic decision whether to invest enormous amount of resources into a new drug innovation project for long term profit or into the development of several generics for short term profit and, consequently, its effect on consumer welfare. importantly, active policy for firms’ strategic innovation decision is especially important in the areas of neglected diseases. Because the innovation for neglected diseases is left behind in firms’ strategic innovation priorities and, in the perspective of public goods, it is urgently required to be actively pursued to meet the social need. For instance, Less than 10 percent of the global spending on health care R&D is dedicated to the major health problems of 90 percent of the world population (Pull Mechanism working Group 2004) The one of the main reasons why large numbers of neglected diseases have been outside of firms’ innovation scope is that the demand for each is relatively small.

0170 Distinguishing the wood from the trees

0001

Kind P. (Department of Pharmacy, Uppsala, Sweden)

The Role of Firms’ Strategic Innovation Decision in Shaping Patient Welfare in Pharmaceutical Industry

The measurement of health outcomes for economic evaluation places specific requirements on the researcher. For example, an index measure calibrated on a 0-1 scale is widely regarded as an essential property for CUA. Generic measures of health-related quality of life (HrQoL) have been widely reported in the health economics literature and these include AQLQ, EQ-5D, HUI and 15-D, together with an SF-36 derivative (SF-6D). These measures are designed to represent the (dis)utility of (ill)health states as descriptively reported by patients. The methods by which these measures have been constructed and the extent to which developers have been successful in calibrating instruments in terms of social preferences are rarely scrutinized. Such a critical review is an essential prior in the selection of any HrQoL measure for use in tech-

Kim W. (Yale School of Management, New Haven, USA) Cleanthous P., Kim W., Witthink D.R. This paper measures the firms’ strategic innovation decision effects on static and dynamic welfare change. The considerable difference between static and dynamic welfare change proves that public policy which actively shaping firms’ strategic innovation decision has more extensive impact on dynamic welfare change than that is not. Such recognition provides new perspective on optimal public policy and broadens its scope of implementation. Social welfare and firms’ strategic inno-

Eur J Health Econom Suppl 1 · 2006

| S51

nology assessment. This paper reviews the principal instruments reported in the economics literature in terms of the source of its descriptive system and the methods used to establish health state values, the representativeness of reference populations, the to discriminate between individuals and over time, and the validity / reliability in (a) descriptive system and (b) value system. This review utilises secondary data to demonstrate the comparative performance of these indirect assessment methods.

0423 The impact of antisocial behaviour in childhood and adolescence on health, social care and other service use and costs in adulthood King D. (London School of Economics Health and Social Care, London, UK) Knapp M. The aim of this study is to determine the effect of antisocial behaviour and conduct disorder in childhood and adolescence on service use and costs in adulthood. While primarily looking at health and social care, other societal services and costs, such as criminal justice costs and the cost of lost employment, are considered. Additionally, we aim to estimate the reduction in ser vice use and costs which would result from improvements in the average scores on measures of antisocial behaviour across the cohort.analysis uses recently collected data from the Cambridge Study of Delinquent Development – a prospective longitudinal study which initially studied a group of boys in 1961/2 when they were 8 years of age, and subsequently interviewed them at ages 18, 32 and most recently at age 48. The data are linked using standard methods for analysis of longitudinal data. Antisocial behaviour, assessed in childhood by peers and teachers and in adolescence using a measure based on criminal convictions, along with covariates, are linked to data on use of health, social care and other services in adulthood. Additionally, assessments of psychiatric health are included to determine if the pathway from antisocial behaviour to service use is influenced by mental illness.

0471 Quality, efficiency and scope in Norwegian hospital production Kittelsen S.A.C. (Frisch Centre, Oslo, Norway) Hernaelig;s K., Magnussen J. If there are economies or diseconomies of scope, the production of hospital services in a region could become more efficient by exploiting any cost savings that may stem from an optimal division of service production between units. Previous studies using hospital level cost function estimates have found no evidence of diseconomies of scope. Studies of individual patient groups or treatments have, however, found improvements in health outcomes with the size of ser vice production. Thus, medical economies of scale may well imply improved health outcomes or better ser vice quality, without cost savings, if hospitals were more specialised. Testing for economies of scope must then control for the quality level. This paper uses quality indicators at the individual patient level to estimate the overall quality level of hospitals. These indicators are then used with conventional input-output variables to estimate a multiple output cost function from a five year panel of data on Norwegian hospitals using the non-parametric Data Envelopment Analysis (DEA) method. To overcome the methodological assumption of convexity inherent in DEA, the sample is split into relative specialised and differentiated hospitals, before comparing costs. Bootstrapping methods are used to test the significance of results.

0505 Supporting healthy ageing in the European Union Kjaeserud G. (European Commission Health Strategy Unit, Public Health and Risk Assessment Directorate General for Health and Consumer Protection, Brussels, Belgium) Siddall C. Part of the organised session by the European Commission: Policy choices for Europe- ageing and sustainability. “Healthy and active ageing plays an important role in stimulating economic growth and in making public finances sustainable. The Community Lisbon programme adopted in July 2005, foresees EU measures to help Member States develop active ageing strategies and encourages governments to increase healthy life years as a means to boost employment. In March 2005 the European Council asked the Commission to reflect on “how to ensure the sustainable funding of our social model.” As a result the Communication “European values in the globalised world” was discussed by the Heads of State or Government at Hampton Court in October 2005. It proposes ways to balance required reforms with European values through increased employment, productivity, facilitating individual changes and mobility. Policies to ensure health ageing must be integrated into this work. The Commission proposal for a Health and Consumer Protection Programme (2007 – 2013) foresees ageing as a cross cutting theme and will enable dissemination of good practice, and support of projects that add to our knowledge effective policies to ensure healthy ageing. However, in line with the Treaty article 152 the Commission is working to improve the understanding of the impact of different policies and actions on health and health systems, and to how those impacts can be determined and taken into account in the policy cycle.”

0594 Introduction on the theory and policy practice of public health economics Klazinga N. (Department of Social Medicine, Academic Medical Center/ University of Amsterdam, The Netherlands) The role of economists in health care is increasing. Both the need for cost-effectiveness information to underpin decesion-making and a growing interest in the functioning of health care markets have give voice to economic reasoning and economic studies in healh care. The whole field of health economics ias such is a testimony of this growing interest. However, health economists are not necessarely taking a population health perspective as the start of their argumentation. As such, studies of health economists may represent the perspective of various stake-holders in the health care system and as in the broader economic debate may represent various in essence political visions on society and its development goals. This session will explore to what extent health economists are good for your health by looking at the health economists approaches in various countries from a public health perspective. The public health perspective takes the health of the population as the ultimate goals of health care efforts and reasons that health care costs are at the same time an investment in a production process that tries to maximize the health potential of a population.

0566 Global budget and Integrating Technology-Intensive Modalities: Germany Kleophas W. (Dialysis Center, Düsseldorf, Germany) Reichel H. Germany represents the case of a global budget with negotiated fees and competing medical insurances. Dialysis therapy is provided most-

S52 |

Eur J Health Econom Suppl 1 · 2006

ly by physicians in private practice and by non-profit dialysis provider associations. ESRD modalities are well integrated into the overall health care system. Dialysis therapy, independent from the mode of treatment, is reimbursed with a weekly flat rate. Health expenses including those related to ESRD are covered for more than 90% of the population by a mandatory health insurance (payments are divided equally between employees and employers). The remainder of the population is covered by private medical insurance. Access to treatment, including dialysis therapy, is uniformly available.

0392 Societal costs of allogeneic blood transfusion in the Netherlands Klok R. (Groningen University Institute for Drug Exploration (GUIDE), Departement of Social pharmacy, Pharmacoepidemiology and Pharmacotherapy, Groningen, Netherlands) Hulst M., Eschbach MA., Slappendel R., Postma MJ. Introductionestimating the cost-effectiveness of interventions to avert blood-transfusions it is important to estimate the costs of one unit of allogeneic blood. An example of such an intervention is administration of erythropoietin.of this study was estimating the costs per unit of erythrocytes in the Netherlands from a societal perspective.the estimation of the costs we followed the path from donor to patient. In this path 6 steps were elaborated donation, production, transport, storage and preparation, administration and the consequences of administration. In these different steps the cost-consequences were estimated.not all costfactors could yet be identified, the cost-estimate is €240,– for a unit of erythrocytes. The indirect costs are responsible for approximately 3% of the unit costs. The production and transport part by the Dutch blood banks is responsible for the majority of the costs. & Conclusioncost-estimates vary widely. Cost-estimates for the United States, United Kingdom, Sweden and Canada vary from approximately €130,– to €930. For the Netherlands we estimated the unit costs for a unit of erythrocytes at €240,– from a societal perspective.

0122 Economic evaluation of cancer screening promotion Koinuma N. (Department of Health Administration and Policy, Tohoku University Graduate School of Medicine, Sendai, Japan) Ito M., Takeyoshi H. Objective: New anti-cancer strategy aiming at a sharp decrease of both morbidity and mortality has been raised as the top-priority health policy in Japan. As long as the pressing goal to improve a five-year survival rate for 20% is put up, cancer prevention becomes the important pillar to be concrete and the effective screening programs are strongly demanded. The objective of our study is to find out the balance sheet of the cancer screening promotion. Methods: We performed simulation to assess the change of medical expenditure when the rate of cancer screening improved to 1.5-3 times in the projected population from the existing rate in five main cancers. We developed system model of cancer treatment along with Markov model, and performed a cost-benefit analysis to find out the incremental benefit for cancer patients detected in screening. Three elements; reduction of screening cost, improvement of detection rate, and detection in earlier stage are anticipated in the improvement of screening rate. We defined 8 ways of combinations of these elements and estimated them according to cancer, sex and age. Results: A reduction effect of medical expenditure of 396 million EUR is expected in total when the screening rate improves 1.5 times from the current rate accompanying 3 elements. The improvement of screening rate for 3 times saves 2.6 billion EUR. On the other hand, medical expenditures increase 185 million EUR and 742 million EUR respectively, when there is no improvement in 3 elements. For the promotion of cancer screening, reduction of screening cost is essential in female stomach, colon and lung cancers. The improvement of detection rate contributes largely to reduction of medical expenditures in male stomach, colon and lung cancers. Conclusions: From the view point of cancer economics, technical progress, i.e., reduction of screening cost, improvement of detection rate and the earlier detection are indispensable for the improvement of the rate of cancer screening.

0139 A difference in cost estimates using different cost survey method in asthma burden analysis

0195 Effects of a hospital merger

Ko S. (Health Insurance Review Agency, Seoul, South Korea) Kim C., Chae S.

Kok L. (SEO Economic Research of Amsterdam, Netherlands) Hop J.P.

Objectives: The purpose of this study was to compare cost estimates as different cost sur vey method was applied to get estimates in cost analysis. Methods: Two types of sur vey were implemented to sampled asthma patients. Firstly, main survey was conducted using face-to-face questionnaire and included 660 patients from 31 institutions. The survey participants were required to remember expenditures associated with asthma during the last 3 months (retrospective survey method). Secondly, these patients were asked to record their daily expenditures associated with asthma on cost diaries designed for this purpose for 3 months (prospective survey method). Results: Outpatient and pharmacy cost was US$411 and $370 in the retrospective method and in the prospective method, respectively. Inpatient cost was $162 in the retrospective method while $35 in the prospective method. Costs related to emergency visit were $19 and $5, respectively. In case of the alternative medicine cost, the figures were $104 and $90. The number of absence due to asthma was also lower in the prospective method. Conclusions: The accurate assessment of the burden of asthma is difficult and obtained values may differ according to cost sur vey method. Costs in the prospective method were notably reduced; recording diaries could affect disease control.

The paper describes the results of an ex-ante investigation in the effects of a merger case of two hospitals in the Netherlands. As patients do not pay a price for treatment in a hospital a straightfor ward SSNIPtest based on price-elasticities is not possible. The idea behind the SSNIP-test is that the merged firms might deteriorate their offer and still make more profits than before the merger. We applied three alternative approaches which are in line with the idea behind the SSNIPtest.first is the time elasticity approach which is developed by Capps et al (2001). This approach looks at the consequences of a hypothetical increase of travel time of the two hospitals. The idea is that patients do not pay a money price for their treatment but they pay in terms of travel time. Their utility of a hospital depends on quality and travel time. second approach looks at the consequences of a decline in quality of treatment in the merged hospitals. We used conjoint analyses to measure patients willingness to travel for quality. third approach measures the willingness to pay for an insurance including treatment in the two merging hospitals.

Eur J Health Econom Suppl 1 · 2006

| S53

0110 Validity of QALY weights derived from the SF-36. Empirical evidence from patients on renal replacement therapy Kontodimopoulos N. (Hellenic Open University, Faculty of Social Sciences, Patras, Greece) Niakas D. This study aimed to provide empirical evidence on the validity of QALY weights (utilities) derived from the SF-36 Health Sur vey. The Brazier algorithm was used to restructure the SF-36 into a single health-state measure known as the SF-6D. Data was obtained from Greek end stage renal disease patients with the KDQOL-SFTM psychometric instrument, incorporating the SF-36 and a kidney disease questionnaire. The sample comprised 642 hospital dialysis (HD) patients, 65 peritoneal dialysis (PD) patients and 167 transplanted (Tx) patients, representing approximately 9% of the national population for each modality. Resulting utility scores were 63.85 (HD), 59.87 (PD) and 71.73 (Tx), with the latter statistically significantly different (P<0.001). Validity was supported by expected correlations (P<0.001) between utilities and an overall health-rating scale included in the instrument. Linear stepwise regression revealed that kidney disease scales explained large portions of the variance of SF-6D utilities, specifically R2=63.5% (HD), R2=59.2% (PD) and R2=43.8% (Tx). Controlling for sex, age and education revealed that men scored higher, whereas age was negatively and education positively correlated with utilities respectively. Furthermore, comorbidities and hospitalizations were both negatively correlated with utilities. Differences were statistically significant (P<0.001) for HD patients and in the expected directions, but not always statistically significant for PD and Tx patients. Results provided initial evidence of valid patient preference measures from the SF-36 in this particular clinical setting. Further research is required to compare these utility scores with those from direct preference methods.

0466 How to value informal care in economic evaluations? Koopmanschap M. (Erasmus Medical Center, Rotterdam, Netherlands) In this presentation several ways to value informal care in economic evaluations of health care are to be discussed: the proxy good method, the opportunity cost method, contingent valuation and conjoint analysis. All valuation methods will be discussed from a theoretical point of view, but in addition empirical results of applying these methods will be shown. The proxy good and opportunity cost method are relatively simple to apply, but are problematic with respect to measuring the relevant amount of time invested in providing informal care. Contingent valuation and conjoint analysis are promising alternatives, as these methods should be capable of incorporating opportunity costs of time and the burden of providing care in one outcome. However, these methods need further validation in the field of informal care.

0572 Cost-effectiveness of fluvastatin therapy following successful percutaneous coronary intervention in the Hungarian health care settings Kósa J. (Novartis Hungary Ltd., Budapest, Hungary) Scuffham P.A. Aims: To estimate the costs, benefits, and cost-effectiveness of fluvastatin used following first percutaneous coronary intervention (PCI). Methods: A cost-effectiveness analysis was undertaken using efficacy data from the Lescol Intervention Prevention Study (LIPS). That study

S54 |

Eur J Health Econom Suppl 1 · 2006

was a randomized double-blind placebo-controlled trial undertaken in 77 centres, predominantly in Europe, of patients with moderate hypercholesterolemia who had undergone their first PCI. Fluvastatin (Lescol) 40mg twice-daily plus dietary counselling was given to the intervention group; the control group received dietary counselling only. A deterministic Markov model was used to estimate the incremental costs per quality-adjusted life year gained, with cost data drawn from the Hungarian National Health Insurance Found. Effectiveness data on fluvastatin was derived directly from LIPS and utility weights from previous studies on heart disease. One-way sensitivity analysis was conducted around key parameters and analyses were conducted for subgroups identified in LIPS. Results: Treatment with fluvastatin cost an additional HUF 417,179 and resulted in an additional 0.107 QALYs per patient discounted over 10years compared with controls. The incremental cost per quality-adjusted life year gained with fluvastatin versus control was HUF 3,895,745. The key determinants of cost-effectiveness were the effectiveness of fluvastatin in reducing Major Adverse Cardiac Events (especially cardiac deaths), the utility weights used, the cost of fluvastatin, and the time horizon evaluated. Fluvastatin was substantially more cost-effective in patients with diabetes, renal disease, multi-vessel disease or LDL-cholesterol >3.4 mmol/L. Conclusions: Fluvastatin is a viable and economically efficient pharmaceutical to reduce heart disease in Hungary when given routinely to all patients following PCI.

0327 Antipsychotics and antidepressants Koskinen H. (The Social Insurance Institution, Research Department, Helsinki, Finland) Maljanen T., Jaana M. Background: Prescription drug expenditures in Finland (population 5.2 million) have been snowballing since the 1980’s, creating considerable economic pressure to the society. Particularly the growth can be seen with antipsychotic and antidepressant drug spending, which has more than doubled since 1998 amounting to 150 million euros in 2004. Objective: To describe and analyze antipsychotic and antidepressant drug expenditure and utilization trends in Finland from 1998 to 2005. Methods: The data was collected from the Social Insurance Institution’s registers and it comprises of nearly all the prescribed outpatient antipsychotic and antidepressant purchases. The outcome measures included utilization of the drug groups measured as number of patients per population, number of prescriptions filled and volume of Defined Daily Doses as well as the costs per patient, cost per prescription and the cost per one Defined Daily Dose. Results: The rise in antidepressant spending can be mostly explained by the growth in number of prescriptions filled, whereas with antipsychotic drugs the cost increase resulted mostly from the increased prescribing of newer and more expensive drug alternatives. Conclusions: Factors underlying rising prescription drug costs vary considerably according to therapeutic classes. Therefore, in planning cost containment measures there should be solid understanding of the reasons causing drug expenditures to rise in each specific drug group.

0426 Which generic HRQoL instrument reflects best the direct TTO valuations of own health by general population? Kotomäki T. (University of Helsinki, Dept. of Public Health, Finland) Honkalampi T., Sintonen H. To be valid for QALY calculations, the HRQoL scores produced by generic instruments for health states should reflect a true trade-off between quality and length of life. At least theoretically, TTO valuations of own

health should do that. To find out how the average scores produced by 15D, HUI3, EQ-5D (with UK TTO valuations), AQoL and SF-6D in different age/gender groups of general population reflect the direct TTO valuations of own health in these groups.. Population survey with a representative sample from 16 age/gender groups (n=250 per group, 17+ years). The trade-offs between length and quality of life implicit in the average scores produced by the instruments in these groups were made explicit. The respondents in each group chose between the number of years implied by different instruments that the respondents on average should be willing to give up for full health or indicated their own preferred number or unwillingness to trade-off. The average 15D scores were almost identical with the average TTO valuations up to the age of 75, but thereafter lower. The HUI scores were second closest, whereas the scores of other instruments were much lower, especially in older age groups.

0424 Early retired pensioners and Longevity:Factors affecting Pension Schemes in Hungary Kovacs E. (Corvinus University of Budapest, Hungary) Szule B. Hungarian pay-as-you-go pension system was transformed into a multipillar pension scheme in two steps. Public and private pensions were expected to finance a future retirement benefits. The present paper extends the model to the case where individuals live longer but the number and ratio of early retired, disabled persons are still high. The problem is twofold, increasing demand for pension and decreasing contribution from the employed persons. In the other hand the attempt to raise revenue in order to finance current pensions was not successful. The labour force participation rates are lower for both sex after the transition period. Social transfers are not enough to support disabled persons in Hungary. The major purpose of the paper is to estimate this pension gap. Methodsmodel was set up and tested using demographic and macroeconomic Hungarian data to measure the effect on pension schemes. The base year is 2004, projection is prepared for 50 years. Analysis will be undertaken to observe the effect of changing the benchmark assumptions on participation rate, early retirement, replacement ratios and mortality reduction.

0177 Managed care and a process of integration in health care sector. A case study from Poland Kowalska K. (Department of Economics, Warsaw University, Poland) The aim of the paper is to identify a causal nexus between contract type (financing rules) and tendency towards integration between Polish health care providers. After reforms in 1999, in the context of underdeveloped information systems and poor enforcement mechanisms, unit financing rules proved to be one of the weaknesses of the Polish health care system. By the way of enhancing structure-disintegrating process, cost-per-case contracts destroyed those of organizational and professional networks (both formal and informal ones) that used to improve the quality and effectiveness of health care process. From the other hand, in 2002, two of the seventeen Sickness Funds implemented contractual arrangements typical for the managed care system. The primary care providers who realized so called Pilot Programme, took on responsibility for coordinating treatment of the enrolled patients and for management of financial resources assigned for the health care packages broader than usual in Poland. Capitation payment was the major technique of financing those providers. By the use of a case-study descriptive analysis it is argued here that capitation prospective payment for wide packages of health care encourages spontaneous (bottom-up) integration between primary and secondary care providers.

0202 The direct and indirect costs attributable to alcohol consumption in Germany Konnopka A. (Health Economics Research Unit, Department of Psychiatry, University of Leipzig, Germany) König HH. This study aims to estimate direct and indirect costs caused by morbidity and mortality attributable to alcohol consumption in Germany in 2002 from a societal perspective. Methods: Using the concept of attributable risks and the prevalence-based approach, age- and genderspecific alcohol-attributable morbidity and mortality was calculated from neoplasms, endocrinologic, alcoholic, nervous, cardiologic, digestive, skin and perinatal disorders, injuries and poisonings. The literature provided data on drinking prevalence by age, gender and dose in Germany and relative risks. Direct costs were calculated based on routine utilization and expenditure statistics. Indirect costs were calculated based on the human capital approach using a discount rate of 5%. Results: 5.5% of all deaths and 970,000 years of potential life lost were attributed to alcohol consumption. Total costs were 24,398 million EUR, equalling 1.16% of Germany’s gross domestic product. Direct medical and non-medical costs were 8,441 million EUR. Indirect costs were 15,957 million EUR (69% mortality and 31% morbidity costs). In contrast, protective health effects of alcohol consumption saved 4,839 million EUR. Conclusion: The magnitude of alcohol attributable health damage and costs demands more preventative efforts.

0206 Validity of the time trade-off method of utility assessment in patients with mental disorders König HH. (Health Economics Research Unit, University of Leipzig, Germany) Roick C., Günther O., Angermeyer MC. To analyze feasibility, discriminative ability and validity of the time trade-off (TTO) method of utility assessment in patients with mental disorders. Methods: In 172 patients with affective, 166 with schizophrenic and 160 with alcohol-related mental disorders TTO utilities were administered through a standardized interview. Measures of quality of life (EQ-5D, WHOQOL-BREF), subjective (SCL-90R) and objective (CGI-S) psychopathology, and functioning (GAF, GARF, SOFAS, HoNOS) provided comparison. Feasibility was assessed by completion rates. Discriminative ability was analyzed by assessing frequency distributions of TTO utilities. Validity was analyzed by assessing the correlation of TTO utilities with all other scores. Results: Of patients with affective/ schizophrenic/alcohol-related mental disorders, 89/86/91% completed the TTO elicitation. 29/43/28% of the respondents were not willing to trade a reduction in their length of life for an improvement of health status. The mean TTO utility was 0.66/0.75/0.61, median 0.85/0.95/0.75. In patients with affective and alcohol-related mental disorders, TTO utilities were significantly correlated (mostly moderate: 0.3
Eur J Health Econom Suppl 1 · 2006

| S55

0259 Advertising and Generic Market Entry Königbauer I. (University of Munich, Department of Economics, Munich, Germany) The effect of purely persuasive advertising on generic market entry and social welfare is analysed. Advertising in the ethical pharmaceutical market is generally allowed, if targeted towards physicians, on the grounds that it provides necessary information which might dominate the downsides of advertising. This paper, however, emphasizes the persuasive aspect of advertising and thus the negative aspect of advertising and shows that, as in the literature on informational advertising, advertising is no barrier to market entry. Advertising creates product differentiation and can induce generic market entry which is deterred without differentiation due to strong Bertrand competition. This advantage can dominate the negative brand-loyality. Hence, the persuasive aspect of advertising can contribute further to the welfare which is gained from the informational content. It is analysed under which circumstances this is the case and how these conditions can be positively influenced by the health authorities.

0234 The German Way of Managed Care Kuchinke B.A. (Ilmenau University of Technology, Ilmenau, Germany) Under the given regulation one can find separated, vertically not integrated providers for physician, hospital and rehabilitation ser vices in the German health care system. These three levels of medical treatment have been largely separated vertically for historical reasons. Nowadays in Germany the need of managed care and vertically integrated providers of medical ser vices are discussed and politically pushed. So far first steps have been made in Germany. Some sorts of co-operation between providers such as physicians and hospitals or hospitals and rehabilitation facilities are allowed by law. In this paper the economic effects of vertically integrated health care providers under the given regulation in Germany are worked out theoretically. In the existing literature especially the potential positive effects of such systems are pointed out. For example the transaction and production costs may fall. The quality of care could be higher. Additionally in this paper it is shown that these positive effects are just one side of the same medal and have to be seen in comparison to the potential negative effects. For example the freedom of patients will fall potentially because they are not free in choosing a doctor or a hospital and the problem of market power may occur. As a conclusion the given regulation is discussed and it is shown that a re-regulation has to be made to implement an efficient health care provider system.

0136 Economic transition and the health financing reform imperative in the former Soviet Union Kutzin J. (World Health Organization/EURO, Copenhagen, Denmark) Objectives: The aims of this study are to demonstrate how the health financing system inherited by the successor states of the USSR was not able to cope effectively with the economic and fiscal impacts of transition; the association between the magnitude of fiscal contraction and the share of health spending coming from out-of-pocket payments; the association between economic transition, fiscal contraction, and the mix of health system inputs funded from public sources; and the experience of selected countries with comprehensive health financing reforms that have mitigated some of these negative trends and enabled improvements in health system efficiency, equity, and transparency. Methodology: The approach combines conceptual and empirical analysis. The conceptual approach is based on applying the framework of

S56 |

Eur J Health Econom Suppl 1 · 2006

health financing functions (collection, pooling and purchasing) to the “Semashko” health system inherited from the USSR, as well as to the implementation of piecemeal reform in Kazakhstan and comprehensive reforms in Kyrgyzstan and Moldova. The empirical work includes analysis of cross-country data relating the extent of fiscal contraction (both levels and changes in total government expenditure as a percent of GDP) and the share of health spending coming in the form of out-ofpocket payments. Administrative data from Kazakhstan and Moldova, as well as extensive household and hospital patient survey data from Kyrgyzstan, is used to analyze the effects of financing reforms in these three countries. The administrative data are combined with a description of reform implementation to examine the impact of reform on the level of public funding for health in the 3 countries. Because the Kyrgyz reforms were phased in over several years and the data sets enable oblast-level comparisons, a combined cross-sectional longitudinal approach is used to assess the effects of the reforms on the efficiency of input mix (changes in the mix of fixed and variable costs over time), the magnitude and frequency of informal payments, and the targeting of benefits to vulnerable groups. The association of reforms with changes in equity in the use of services is also explored. Results: The cross-country analysis illustrates the radically different context faced by the countries of the Caucasus and some countries in Central Asia (Tajikistan and to some extent Kyrgyzstan) that experienced extreme contraction of the state, as compared to some other countries that experienced more modest levels of contraction (Kazakhstan, Uzbekistan, Moldova, Ukraine, Russia), and one that has had essentially no transition (Belarus). There is a clear positive relation between the extent of fiscal contraction and the share of health spending coming in the form of out-of-pocket payment. Hence, much of the difference in the funding mix that we observe in these countries can be explained by this aspect of the macro context. However, there remains considerable unexplained variation for countries with similarly sized public sectors, suggesting that differences in policy as well as the extent to which governments prioritize health in resource allocation make an important difference. The analysis of selected reform experiences suggests that the introduction of a new funding source (payroll-taxes for compulsory health insurance) by itself may have a negative impact on the overall financing system unless it is coordinated with existing budget funds and combined with a fundamental change in the way that public funds for health are pooled and the way ser vices are purchased from these funds. In Kyrgyzstan, the change in pooling and purchasing methods led to a substantial reduction in fixed costs and a reduction in informal payments, while also being associated with improved equity in the utilization of care across income groups. Conclusions: The economic transition created specific challenges to the inherited Semashko health financing system. Fiscal contraction led to a growing share of out-of-pocket payments in the health financing mix, while the wider shift to a more market-based economy also led to increased prices for key inputs such as energy and medicines. This change necessitated countries to reduce the fixed inputs costs of the system or be faced with an ever-decreasing amount of public funds available to support direct patient treatment inputs. The experience of the CIS countries that are on the leading edge of reform (Kyrgyzstan and Moldova) suggests that a priority for financing reform is to replace the fragmentation and duplication inherent in the Semashko system with a more centralized pooling of funds (at least at oblast level, and potentially at national level, depending on population size), the financial separation of purchaser from providers and use of output-based purchasing methods (i.e. an end to the hierarchical line item input-based budgeting process), and the adjustment of the wider public finance system to accommodate these changes in the health financing system.

0520

0232

Country-level cost-effectiveness analysis using who-choice: An example of alcohol interventions from Estonia

Choice, Price Competition and Complexity in Markets for Health Insurance

Lai T. (Department of Public Health, University of Tartu, Estonia) Reinap M., Habicht J., Chisholm D. Aim: To adapt WHO-CHOICE approach for country-level use and assess the population-level costs, effects and cost-effectiveness of different alcohol control strategies in Estonia. Design: Regional cost-effectiveness estimates from the WHO’s CHOICE project were used as a starting point for the contextualisation and regional data was substituted to better reflect the situation in the country. Population-specific data on Estonian demography, alcohol epidemiology, inter vention costs and effectiveness was included. A simplified modelling framework of WHO-CHOICE was used to estimate the total costs and effects of interventions compared to a situation in which no interventions are implemented. Costs were assessed in Estonian Kroon (EEK) for the year 2000 (Euros were also computed), while effects were expressed in disability-adjusted life years (DALYs) averted. Results: Most cost effective interventions to reduce burden of disease from hazardous alcohol consumption is increased excise tax. Increasing taxation costs 759 EEK (€49) per DALY averted to reduce hazardous alcohol consumption. Adding another inter vention costs 1331 EEK (€85) for every additional DALY averted from hazardous alcohol consumption. The first choice to be added would be advertising ban. Compared to WHO baseline estimates of the analysed inter ventions there was a consistent decrease in intervention costs resulting in more favourable CER in all cases. Conclusions: All inter ventions analysed were found to be cost-effective with population-wide measures like taxation having the best CER in Estonia. Personal interventions such as brief advice to heavy drinkers are effective but much more costly and therefore less cost-effective. The differences between WHO priors and adjusted results underline the importance of contextualising WHO-CHOICE results for small countries like Estonia as more accurate data on cost-effectiveness of interventions allows better health policy decisions on national level.

0056 Adjustment of existing EQ-5D TTO value sets for use of an EQ-5D five level descriptive system Lamers L. (iMTA, ErasmusMC, Rotterdam, Netherlands) The EuroQol Group started to develop a five level EQ-5D (EQ-5D-5L), which can be designed by adding a level 1½ and 2½ between the existing levels. Since EQ-5D is often used to calculate QALYs, the objective of this study is to estimate decrements for level 1½ of a 5L version using the Dutch existing EQ-5D-3L TTO value set as the starting point. In the Dutch 3L valuation study 17 health states were directly valued. In this study two moderate state, two severe and the worst EQ-5D state from this subset were selected for direct valuation as well as 12 mild states, which were created by changing levels 1 and/or levels 2 into 1½ of (very) mild states of the Dutch 3L subset. 100 students valued these health states using TTO. The combined TTO dataset of this 5L study and the Dutch 3L valuation study was used to estimate a random effects regression model. The estimated coefficients for level 1½ were only significant different from zero for the dimensions self care and anxiety/depression. The results showed a context dependency effect for TTO valuation. The practical implications of our results for further adjustments of existing TTO value sets will be discussed.

Lamiraud K. (Institute for Health Economics and Management (IEMS), Lausanne, Switzerland) Frank R.G. Lamiraud K. This paper examines choice exercise and price competition in health insurance markets. Switzerland offers a unique opportunity to conduct such an analysis as numerous sickness funds provide community-rated premiums, homogenous benefits, open enrolment and get risk-adjusted payments. Quite unexpectedly one observes low rates of switching and little price convergence. We examine whether some explanations grounded in behavioral economics could help us understand such a puzzle. We make use of publicly reported data in Switzerland in combination with a survey of insured households. Our results are consistent with the presence of both decision making overload and status quo bias in health insurance markets. Consumers facing large numbers of health plan choices were less likely to switch health plans. So were people with longer periods of attachment to a particular plan. Moreover, people making new health plan choices chose to enroll in different funds than those who had not switched in some time. At a moment in history where more and more nations are thinking of favoring consumer choices in health insurance markets our results may offer some cautions regarding the needs for decision supports and mechanisms that simplify such choices if the desired price and quality outcomes are to be encouraged.

0116 An Economic Analysis of the Effects of Obesity on Wages Lange R. (U.S. Food and Drug Administration, College Park, USA) Several economists have connected obesity to wage penalties, particularly for women. This analysis extends previous literature examining the effects of body mass index (BMI) on income. This analysis improves on previous work by using a dataset that can allow health effects to be better examined. the first series of regressions, log wage income is regressed on a series of variables including BMI. The coefficient on BMI is significant, negative, and very small. Separating the sample into men and women and running separate regressions on the two groups yields interesting results. Body mass index has no significant effects on the wages of either gender.second series of regressions replaces the BMI variable with BMI categorical variables under weight, normal, over weight and obese. In the regression run only on women, overweight women experience only a very slight wage penalty, but it is not significant by conventional measures. By contrast, underweight and obese women do experience significant wage penalties. Obese men also experience significant wage penalties. Given these results, it is possible that any sigma associated in the past with being overweight, at least with regard to income, has subsided. However, being obese does appear to hurt both men and women.

0278 A Model of the Effects of Patient Comfort on the Decision to Seek Treatment Lari N. (North Carolina State University, Raleigh, NC, USA) Diabetes is a common chronic disease that leads to serious complications when inappropriately treated. Basic care includes a complete physical exam once each year and HbA1c testing every three months. Diabetics have a greater risk for developing damage to the eyes, kidneys, nervous system, cardiovascular and circulatory systems. Patients must also monitor symptoms of such damage. Thus, the physician plays an important roll in the patient’s future health. As the level of comfort with the Eur J Health Econom Suppl 1 · 2006

| S57

physician increases and the opportunity cost of visits declines, patients are more likely to visit their physician. This paper attempts to model the effects of ‘patient comfort’ among pediatric diabetes patients and investigate the effects of comfort on their future health. Increased patient comfort, which includes factors such as comfort with the physician and his/ her staff, ease of making and attending appointments, language and cultural barriers, etc., is expected to improve the patient’s long-term health as the patient seeks routine care and/or care when symptoms first present themselves.

0082 Do Health Shocks Lead to Earlier Retirement When Workers are Universally Insured? Larsen M. (The Danish National Instititute of Social Research, Copenhagen, Denmark) Gupta N.D. Health shocks significantly alter workers’ retirement plans, more than existing health conditions. The labour supply response following a health shock depends on access to pensions and/or public/ private insurance. Within an employment-contingent health insurance system as in the U.S., individuals may have to continue working to be able to pay the costs of their treatment. When workers are universally insured, no such mitigating effect on labour supply can be expected. Therefore, health shocks would trigger greater retirement in a general tax-financed health-care system with universal access and may be one reason behind the earlier retirement of workers in such economies. We explore this question by estimating panel retirement age models of the effect of new health disturbances on data drawn from the Danish Longitudinal Registers. Our aim is to compare the retirement effects of negative health shocks in a universally-insured health care system to those found in U.S. studies for similarly defined health conditions. We apply truly objective health measures based on medical diagnoses made at the time of hospital discharge from the Danish National Patient Registry records. As objective measures need not necessarily correlate with work incapacity, we focus only on diagnoses for acute discharges that impose serious work limitations.

0266 The effect of price and quality on health insurer choice Laske-Aldershof T. (Erasmus University, Rotterdam, Netherlands) Aim: In 2005, the government published comparative information about health insurers on the Internet. In order to monitor the sensitivity of consumers to premium and quality differences, a baseline study is performed in 2004. Method: Two datasets from 2004, concerning health plan choice of sickness fund enrollees (n =1,100) and individual member satisfaction ratings on several quality aspects (n = 4,000), were combined. For each sickness fund we constructed an overall quality rating. We estimated a probit model of sickness fund choice with premium, quality aspects, age, gender, education level, family size, and health status as the most important explanatory variables. Results: There appears to be little variation in satisfaction ratings between sickness funds. The probit estimation results show that there is no significant effect of quality on health insurer choice, but people who pay higher contribution rates are more likely to switch to a cheaper fund. Furthermore, the decision to switch is significantly negatively related to age and family size and is significantly positively related to knowledge of the health insurance. Conclusion: In 2004 people were sensitive to premium differences, but not to quality differences.

0335 Policy making for rare childhood diseases Leal J. (Health Economics Research Centre, Department of Public Health, University of Oxford, England) Wordsworth S., Gray A., Oerton J., Philips P., Dezateux D. Medium chain acyl-CoA dehydrogenase deficiency (MCADD) is a rare inherited error of fatty acid metabolism, with high mortality and morbidity which may be easily preventable through early diagnosis and treatment. Several European countries are already providing universal MCADD newborn screening. The UK National Screening Committee (UKNSC) is currently reviewing its policy for MCADD. Our study examines the cost-effectiveness of universal MCADD screening compared to no-screening and is part of a larger evaluation funded by the UK Department of Health. A Markov model examines the lifetime costs (screening, diagnosis, morbidity and treatment) and effects (sur vival) of the two strategies. Unit costs were derived from reference costs, published literature and primary data collection. Effectiveness data were obtained from a current UK MCADD screening pilot and surveillance study, international prospective studies, and the literature. Probabilistic sensitivity analysis was used to explore model uncertainty. We present the model results and discuss the challenges in using outcomes measures, in particular QALYs, in evaluating newborn screening. We consider the implications of uncertainty for a potential decision of implementing MCADD screening for UKNSC. Finally, we consider whether screening for rare childhood diseases needs to be evaluated differently from screening for more common diseases.

0400 Multilevel analysis vs other methods – expenditures on health promotion in Poland Krajewski-Siuda K. (Faculty of Public Health, Medical University of Silesia, Bytom, Poland) Romaniuk P., Hubicki L., Kociel P. Background: Self-governments has become since 1999 responsible for realisation of health promotion and prevention programmes, although this task is non-obligatory for the units. The survey is an attempt to resume counties’ experience concerning health promotion and prevention. Aim: The aim of presented research is to analyse data concerning counties’ expenditures on health promotion and prevention programmes in 1999 – 2004. Methods: The questionnaire was send to all units via e-mail. Expenditures were corrected by (health) inflation rates. The expenditures per capita and percentage of total budget were a dependent variable in the multivariate analysis. The t-test, GLM (general linear model) and MLA (multilevel analysis) were adopted to validate whether the choice of statistical technique can modify final conclusions. Results: The response rate (after one year) was 197/380 (52%). The results of multivariate analysis (both GLM and MLA) are not coherent with the results from t-test in this particular study. The significance level of the region, time and type of unit (rural/urban) varied from not significant up to high significant (p<0.00000001). The results from GLM and MLA were simillar. Conclusions: A multi-level analysis should be considered in health accounts research.

0419 Measurement of Public Service Productivity Lee P. (The Office for National Statistics, London, United Kingdom) A great deal of work has been carried out on improving measures of inputs, output and productivity of the health service in the UK in recent

S58 |

Eur J Health Econom Suppl 1 · 2006

years by the Office for National Statistics (ONS), the Department of Health for England, and a project group involving the National Institute for Economic and Social Research and the Centre for Health Economics at the University of York. This work has been supported and in many instances driven by Sir Tony Atkinson’s Review of the Measurement of Government Activity for the National Accounts.ONS published a first article entitled Public Service Productivity Health in October 2004, and a second article on health is scheduled for publication in 2006. The article is designed to explore public service productivity within the context of the UK National Accounts, drawing on other information to paint a more detailed picture of public service output and productivity than the National Accounts themselves.conference paper will present the work carried out by the ONS to improve existing measures, focusing in particular on health service inputs, output and productivity as essential material for public accountability and better management of the UK health service.

0164 The Migration of Highly-Skilled Workers Léger P.T. (HEC Montréal at Université de Montréal, CIRANO and CIRPÉE, Montreal, Canada) Dostie B. The theoretical and empirical economic literature on who and where individuals migrate is quite vast. At its most basic form, the economic model of migration posits that individuals will migrate if the expected benefits of moving to an alternate location are greater than the expected benefits of remaining in their current location (net of transaction costs). This model suggests that individuals will migrate if they can expect to earn more in an alternate location than in their current location, where the choice of location will be the one that maximizes expected earnings. Viewed in this way, migration is simply an investment in human capital. In a heterogenous population, however, the returns to migration (i.e., the returns to investment in human capital) are also likely to be heterogenous. As a result, the basic model of migration can be augmented to allow for sorting based on skills. According to Borjas and al (1992) different locations (in their case, U.S. states) are characterized by their own wage generating process. Consequently, highly-skilled individuals will wish to migrate to regions with a high-skills premium (i.e., with a relatively large variance in wages conditional on mean wages) whereas lowskilled individuals will wish to migrate to regions with a relatively lowskills premium (i.e., with a relatively small variance in wages conditional on mean wages). Although the link between skills and migration is an important one, our understanding of this relationship is limited for several reasons. First, there are very few studies which examine this issue empirically, partly because of a lack of suitable data. Moreover, there is little consensus as to what constitutes a highly-skilled worker. Theoretically, a high-skilled individual is defined simply as being more productive. Because of a lack of suitable data on individual-specific productivity, highly-skilled workers are often defined by their level of education and/or their profession and/or their wage. However, none of these measures are entirely satisfactory as they only take into account observable differences which may not accurately or fully reflect productivity differences between workers. Furthermore, the issue of unobserved skill differences between workers who leave and workers who stay is complex, and ultimately difficult to resolve (Card, 2003). In response to these limitation we wish to contribute to the recent empirical literature on sorting of workers (via migration) based on skills and location-specific returns to skills (Borjas, 1987). However, unlike many other papers (Hunt and Mueller, 2004 and Chiquiar and Hanson, 2005), we consider not only observable skills but also unobservable skills (and their respective location-specific returns). In doing so, we will also contribute to the literature on the fundamental drivers of mobility of highly-skilled workers. In order to do so, we focus on the inter-provincial migration patterns of Canadian physicians. We choose this particular group for sev-

eral reasons. First, physicians (especially specialists) are often singled out, in Canada, as a highly-educated group who frequently experience both international and inter-provincial migration. Furthermore, focusing on physicians allows us to study the migration decisions of a relatively homogenous set of workers in terms of formal education. Finally, Canadian physicians are generally paid on a fee-for-ser vice (i.e., a fee-per-consultation) basis which are set at the provincial level. Thus, wage rates (or fees) can be considered exogenous to both observable and unobservable individual characteristics as they reflect exclusively the physician_specialty and province of practice. Consequently, we can use the physician_total billings as a measure of his or her output. Because of this environment, we are able to offer a precise definition of a highlyskilled physician (or worker) a physician who has greater billings (i.e., has a greater output of health-care ser vices) than would other wise be predicted by his or her observable characteristics and their corresponding exogenous wage-rate. In other words, we consider a physician to be more skilled (or equivalently, more productive) if they have greater total billings than observably identical physicians. Finally, because (i) interprovincial migration is much more frequent than international migration, (ii) health care is a provincial responsibility in Canada, and (iii) we observe we observe physicians prior-to and post inter-provincial migration, we develop a two-stage model of wage determination and migration decisions. In the econometric section provided below, we first estimate province-specific earnings equation and use the estimated parameters to predict the physician_potential earnings for each potential destination (i.e., for each of the other provinces). We then estimate a conditional-logit model with unobserved heterogeneity to examine the effects of individual and destination-specific characteristics (particularly earning differentials) on the choice of physician location. Our setup also allows us to examine whether differences in the returns to observed human capital and unobserved skills across different provinces are an important factor in physician mobility.

0557 The assessment of drug benefit by the German Institute for Quality and Efficiency in Health Care (IQEH) in an international perspective: what role for health economics? Leidl R. (Institute for Health Economics and Health Care Management, GSF-National Research Center for Environment and Health, Neuherberg, Germany) In Germany, the Federal Joint Committee (FJC) decides whether or not new drugs are classified into reference pricing groups. On demand by FJC, the newly created IQEH carries out drug benefit assessments. This renders the evidence basis for the classification decision. FJC rules for the assessment, IQEH study scopes derived at the beginning of the process, and first assessment examples provide the basis for an analysis of this regulatory procedure. By law, economics is not part of the assessment, but indicators used belong to both the cost and the effect side of an economic evaluation. Findings contrast strongly the highly transparent procedure used in the British National Health Service, but also decision making in the better comparable Dutch health insurance system. Methods and results of the assessment of selected drugs are compared across the three systems. For Germany, it is suggested to extend IQEH’s responsibility to economic assessment, and to standardize the methodology used therefore. Featuring a very comprehensive task, this will provide better transparency in the overall result when assessing cost and effect impacts of a drug. In an international perspective, reducing regulatory variety would likely diminish costs and regulatory uncertainty to innovators.

Eur J Health Econom Suppl 1 · 2006

| S59

0407 Working conditions and health status of older workers Lengagne P. (IRDES, Paris, France) Debrand T. To preserve health of older workers is a major issue to health and social policy in European countries, because success of policies consisting in increasing labor force participation rate among ageing people depends on it. Even if the relationship between health status and participation in the labor market appears obvious, the causal relationship is not clear. Two contradictory effects seem to operate: on one hand arduous work may result in worse health, particularly towards the end of working life, and on the other poor health may result in early departure from the labor market. Several studies show that health status modifies labor participation of older workers. But few articles introduce working conditions as a determinant of health, and thus, labor supply. In this paper, using the data from SHARE (Sur vey on Health Ageing and Retirement in Europe – which represents the population of individuals aged 50 and over for ten countries on the European continent), we measure the impact of working conditions on several health indicators: self-assessed health, longterm illness, limited activities…

0531 EFPIA HTA Key Principles: A Payer Perspective Lengyel G. (National Health Insurance Fund Administration, Budapest, Hungary) Health technology assessment is a mandatory requirement in the reimbursement application process of a new medicine since May 2004 in Hungary and the requirements of the analysis are officially stated (Directive of the MoH: Cost-effectiveness analysis guideline, issued in 2001). As HTA became a formal element in reimbursement decision making it helps to maximize the health-gain of reimbursed medications. HTA outcomes can be implemented in the decision making process only if the input data is local and the analysis is adopted to the local settings. The potential positioning of the new drug in the clinical practice and in the disease management may define its reimbursement opportunities. Besides clinical benefits the cost of the new therapy is critical as the financial resources are limited. Innovative drugs bring therapeutic advantages but affordability is always the key question and potential savings may be realized only in the long term or they fall outside the health care sector. In order to reward the added value the payer has to re-allocate the resources or the budget will be overspent. The most difficult and most interesting question is how to finance innovation and manage financial concerns at the same time, in parallel with serving the societal health-care demands.

0037 Complexity and Economic Evaluation in Health Care

transdisciplinarity. For the health economic field, the use of complexity theory will involve the acknowledgement of complex, interdependent relationships with broader contextual, economic, social, cultural, political and other non-technical factors. This would lead the field towards an accountability and epistemology based on pluralism and uncertainty, requiring new forms of lay-expert engagement and roles of lay knowledge into decision-making processes. A turn towards complexity theory would provide the basis for the engagement of health economics and politics, and economic evaluation knowledge and broader society.

0217 Contract setting in health care when quality depends on an irreversible investment decision Levaggi R. (Dipartimento di Scienze Economiche, University Degli Studi di Brescia, Italy) Moretto M. The cost and quality trade-off in contracts for hospital care is studied by assuming that quality is a running cost. We argue that quality is the result of an investment decision. This assumption shifts the focus of the incentive-compatible contract from cost revelation to intertemporal investment decisions. Levaggi and Moretto (2004) shows that to incentivate quality the purchaser should set an intertemporal purchasing rule allowing hospitals to increase their activity only if they invest when the technology is new.study the game between the purchaser and the provider of health care that determines the best price and quality contract given that the purchaser faces a trade off between price, quality verifiabily and hold up problems.use a two-period model à la Abel et al. (1996) where the hospital can expand its capacity by making an investment now or in the future through a real option approach. The technology is innovative only at the beginning of its life when learning costs are higher and future operating costs are not known. In subsequent periods the technology is consolidated and the hospital that has invested in the first period produces a positive externality to the system.

0527 Governance and Corruption in Health Care Systems in Transition and Developing Countries Lewis M. (Center for Global Development, Washington D.C., USA) This paper defines governance and corruption in health care, and then reviews the ample if fragmented evidence. Cross country regressions, and data from institutional and governance surveys on the perceptions of corruption in the sector provide a context for the summary of evidence on the performance of health care systems. Drawing on comparative quantitative measures such as absenteeism, petty theft, informal payments and leakage of public funds the paper reviews evidence at the country level. Effective country policies and initiatives that have bolstered accountability and improved performance are reviewed to guide policy.

Lessard C. (University of Montreal, Quebec, Ca nada) With the growth of health care costs, managers and policy makers turn towards health technology assessment, with a focus in the area of health economics. Despite the aesthetics of models and the precision of computation, decision analyses may oversimplify complex health care decisions. These analyses often ignore important health consequences, contextual elements, relationships or other relevant modifying factors, which might not be appropriate in a multiobjective, multistakeholder issue. One solution would be to develop a new paradigm based on the issues of perspective and context. Complexity theory may provide a useful conceptual framework for economic evaluation in health care. Complexity thinking contributes to develop awareness of issues including uncertainty, contextual issues, multiple perspectives, broader societal involvement, and

S60 |

Eur J Health Econom Suppl 1 · 2006

0251 Variations in the Prescribing Patterns of Statins Link C. (University of Delaware, Newark, United States) Condliffe S. Rationale: The identification of high cholesterol as a risk factor for coronary heart disease (CHD) has prompted implementation of prevention programs, with statins playing a major role. Studies have shown for many drugs that disparities in prescribing patterns exist that are unrelated to clinical guidelines. (e.g., across insurance types, regions, and various demographic characteristics).

Objective: The objective of the paper is to estimate multinomial logit models of the determinants of prescribing patterns for statins for persons having CHD and dyslipidemia. The key question, do the above mentioned disparities exist for statins? Three large databases provide so cioeconomic, demographic, and diagnostic information. These include office-based physician settings (National Ambulatory Medical Care Surveys) and hospital settings (National Hospital Ambulatory Medical Care Surveys). The Medical expenditure Panel Survey (MEPS), a panel, has richer socioeconomic data about respondents. Results based on NAMCS and NHAMCS: Preliminary results found variations across minorities, regions of the US, and insurance categories. It is evident that variations do exist. Between 1992 and 2002 statin usage increased by between one and five percent per year. We are in the process of running the analyses for 1996 through 2002 using the Medical Expenditure Panel Survey (MEPS).

0288 Effect of the generic substitution policy on entry, prices and expenditures of pharmaceuticals Linnosmaa I. (Department of Health Policy and Management, University of Kuopio, Finland) Kangasharju A., Valtonen H. Objective of this study is to examine the impact of the generic substitution policy on the entry of generic pharmaceuticals, prices and expenditures of pharmaceuticals in the Finnish pharmaceutical markets. We use quarterly sales data for individual pharmaceutical products in seven ATC groups for the years 1997-2005. Generic substitution is allowed for drugs in four ATC groups (drugs used to treat axid related disorders (ATC group A02B), high cholesterol (ATC group C10A), high blood pressure (ATC groups C07-C09), and depression (ATC group N06A)). Generic substitution is not allowed in three ATC groups (drugs used to treat psychosis (ATC group N05A), depression (ATC group N06A), and epilepsy (ATC group N03)). The three latter groups are used as a control group in the study. use panel data models to study the influence of the generic substitution policy on the entry of generic pharmaceuticals, prices and expenditures of pharmaceuticals. Endogenous variables are regressed on the dummy variable, indicating the time of introducing the policy, variables characterizing markets and other relevant covariates. Appropriate estimation techniques, like the instrumental variable estimation technique, will be applied to solve any problems due to endogenous explanatory variables. Results and conclusions will be presented at the conference.

0192 A multilevel analysis on diabetes care practices in primary care Lippi Bruni M. (Department of Economics, University of Bologna, Italy) Nobilio L., Ugolini C. In recognition of their generic potential to influence doctor behaviour, bonuses paid over and above the physician_base income have been increasingly introduced in primary care to try and improve allocative efficiency and quality of care. Although many studies have confirmed that physicians respond to financial incentives, there is scarce empirical evidence on how such schemes influence quality of treatment. This problem reflects the lack of empirical data but also the complexity of general practice in which many confounding factors are likely to influence physician behaviour. Given this background, we investigate the impact on care quality of having introduced economic incentives in the primary care contracts in the Italian region Emilia Romagna. We concentrate on the treatment of patients affected by diabetes mellitus type 2 and we test the hypothesis that, other things equal, local health authorities that introduced financial incentives aimed to increase primary care assumption of these patients were more likely to register a decrease in hospital

admissions for hyperglycemic emergencies and in belated hospitalizations. To this end, we examined the combined influence of physician, organisational and patient factors through the use of multilevel modelling. Estimations are based on the year 2003.

0315 The intergenerational transmission of health Llena-Nozal A. (Faculty of Economics, Universiteit Amsterdam, The Netherlands) Lindeboom M., Klaauw B. This paper investigates how much of health is transmitted across generations through genetics and how much through behavior and socio-economic status (SES). Low SES children start adulthood with lower education and health and this may constrain their economic position in adulthood. The intergenerational correlation of income may be linked to health but it is not clear what mechanisms are driving this. Low SES parents may face financial constraints that lower parental investment, they may have lower parenting skills, or they may have worse health outcomes that constraint economic opportunities and these may be transmitted across generations. Both unobserved heterogeneity and parental health need to be accounted for in order to disentangle the correlation. We use a cohort study, the National Child Development Study, which follows individuals since their birth in 1958 until their 40s. We exploit the richness of the data and incorporate information on twins, adoptees and the cohort members’ own children to disentangle the nature and nurture components. Our findings suggest that income is important for the child’s health but this effect disappears when we control for parental health and time investment. On the other hand, when financial difficulties are used to measure poverty, the effects persist.

0584 Effectiveness, Efficiency and Equity to accomplish the Millennium Goals in Cuba Pardo C.M.L. (Habana University, Faculty of Economics, Cuba) In the World Health Report 2003, the World Health Organization presents the MDGs that refer to health, with its goals and indicators. From all indicators (6 in total) are examined, regarding their accomplishment in the case of Cuba, the following: reduce child mortality, improve maternal health and combat HIV/AIDS, malaria and other diseases. It is demonstrated that Cuba has showed high effectiveness (understood by the achievement of the defined goals), efficiency (considering the relationship between results and resources) and equity (according to the territorial distribution) in the accomplishment of the objectives considered in the context of Latin America and the Caribbean, Canada and the United States. This presentation also discusses that, in terms of the analysis, indicators to quantify effectiveness, efficiency and equity has been built.

0494 EFPIA HTA Key Principles: An Industry Perspective Loth K. (Novartis International AG, Brussels, Belgium) Health Technology Assessment (HTA) is an impor tant means of enabling governments and other health payers to develop mechanisms for evaluating the clinical and cost-effectiveness of medicines and other health technologies. This presentation, illustrated by examples, focuses on three key principles. First, that assessments should be based on a clear, sophisticated and differentiated view of what constitutes value. A narrow definition of value that does not incorporate the perspective of patients, carers, healthcare professionals, and ignores cost implications elsewhere in the system, whether personal or public, distorts the true Eur J Health Econom Suppl 1 · 2006

| S61

value of a medical intervention. Secondly, any assessment or appraisal must be transparent and balanced and should conform to the principals of the European Transparency Directive i.e. the process must be non discriminatory, be based on objective and verifiable criteria and there must be a right to appeal a decision. HTA should be implemented as a means to achieve better health outcomes, not to delay or exclude new medicines from reaching patients. Thirdly, the appraisal process should remain separate from the regulatory review process. The decision on whether to fund a medical intervention remains a subjective one, based on societal values and the local funding situation. As long as member states retain the responsibility for health care, national authorities or politicians must remain accountable for the decisions on whether a medical intervention is deemed to be cost effective.

0179 Identifying constructs in measuring patient empowerment Loukanova S. (University of Heidelberg, Germany) Bridges J.F.P. Patient empowerment is a key issue in health care quality improvement. There is a need for a validated instrument for patient empowerment assessment. The principle aim of this research is to define patient empowerment and to identify constructs in measuring empowerment. An expert panel of key informants was formed to identify their views of empowerment from a wide range of perspectives. In-depth interviews were conducted with selected informants to explore their opinion about what is important in terms of patient empowerment. Data were analyzed using the Framework analysis. Initial discussion was concentrated on two sets of factors, influencing empowerment – human capital (health and health care status, ownership and responsibility, literacy, health outcomes), and health care system resources (access, advocacy, decision making). Further analysis suggested more conceptual constructs. The result indicated that empowerment is the state of control over own’s health through a dynamic, multidimensional process of taking responsibility to be involved in decision making and treatment. To be empowered, patients need their human capital and a patient oriented health care system that accepts (legalizes) their involvement. This definition should be used as a base for development of empowerment measure instrument among patient groups.

0454 Do different models provide different evidence? Implications for policy making. Lourenço O. (Faculdade de Economia da Universidade de Coimbra and CEIS-UC, Coimbra, Portugal) Quintal C., Ferreira P.L. Barros P.P. The robustness and soundness of empirical evidence provided by health economics may affect the way how policy makers use it to decision making. However, the output of research might depend on the methodology and hypotheses adopted. It is therefore important to assess the stability of results across models. Our objectives are: to compare the results of different regression models that assess the impact of socio-economic variables in health care utilization in Portugal, these models being the NEGBIN, Hurdle and Latent Class models (LCM). Our data comes from the latest National Health Survey. The results suggest that LCM models perform better than the others; regarding aggregate results for the whole population, different models generate results that are quantitatively similar; the two classes, derived from the LCM model, diverge in terms of the economic effects of regressors in the utilization of care: contrarily to low users, where most of nonhealth variables are economic and statistically significant, in the high users class, most of non-health variables are not statistically significant.

S62 |

Eur J Health Econom Suppl 1 · 2006

Results for the population are similar for the three models, suggesting some consistency, still, LCM models, by allowing us to identify latent classes of users, generate more detailed information relevant for policy making.

0338 The impact of smoking restrictions on smoking prevalence Ludbrook A. (Health Economics Research Unit, University of Aberdeen, Scotland) Bird S. A number of countries across Europe have either planned or implemented restrictions on smoking in public places. Apart from the health gains from reduced exposure to environmental tobacco smoke, such restrictions may have the potential to contribute to reducing smoking prevalence. However, determining whether and by how much prevalence has been or will be reduced as a result of this policy intervention is difficult and provides an example of the problems which will be encountered more generally in evaluating policy inter ventions.available evidence relating to smoking prevalence is drawn from studies of workplace smoking bans and clean air laws. A variety of study designs have been employed, some more robust than others. The paper discusses the potential sources of bias in the different study designs and the problems of attempting to pool data from different types of study in quantitative meta-analysis. A number of estimates for prevalence effects are produced and translated into health gains. The paper argues that in a context where ‘gold standard’ RCT data are not available, careful analysis can ensure that policy impacts are estimated with sufficient precision for policy making and policy evaluation purposes.

0306 Injury Severity and Compensation Luke C. (The University of Queensland, Brisbane, Australia) In this paper we analyse the claims activity and claims outcomes for compensable individuals who have been injured in automotive accidents in the State of Queensland for the period 1999–2004. We investigate the impact of a range of covariates, including injury severity (as measured by the New Injury Severity Score), demographic characteristics, income, rehabilitation treatments and expenditures, as well as data on claims management processes and demographic factors on (i) claims propensity, (ii) claims success, and (iii) payout magnitudes. Our study is somewhat unique in the sense that it includes objective (i.e., non-self-reported) data on injury severity as well as the usual fields that are available under compensation schemes of this kind. Our results shed some light on an important, but somewhat neglected field within health economics, i.e. the economics of health and health care under compensation and common law schemes.

0350 Informal and formal care among single-living elderly in Europe Lundborg P. (Lund University Centre for Health Economics (LUCHE), Lund, Sweden) Lindgren B., Bolin K. The aims of this study were (1) to analyse whether informal care, provided by children or grandchildren to their elderly parents, and formal health care are substitutes or complements, and (2) whether this relationship differs between different parts of Europe. The analyses were based on the newly developed SHARE (Survey of Health, Age, and Retirement in Europe) database. We found (1) that informal- and formal home care are substitutes, while informal care is a complement to doctor- and hos-

pital visits, and (2) that these relationships in some cases differ among the European countries that were studied.

0565 Health Care Organization and Financing of End Stage Renal Disease: The Spanish Model Luño J. (Hospital General Universitario, Madrid, Spain) Marañón G. Renal replacement therapy in Spain is completely covered by the universal and free National Health Service. While the prevalence of ESRD in Spain is high (925 people per million), the incidence (130 people per million) is only a third of the US ESRD incidence and is similar to Europe’s average incidence. Spain’s remarkable rate of cadaveric organ donation, 35 per million population in 2004, is the highest in the world. Whereas most European countries have a renal transplant incidence rate of only 30 transplants per million, Spain’s is 50 per million. Spain spends relatively little on general health care (1319 Euros per capita), compared to other countries, but ESRD care represents a larger portion of the overall budget (7.7% of the Gross Domestic Product). Quality of care in the NHS is not determined by free competition or performance rewards; instead, it is monitored by several health agencies. Nevertheless, nephrologists with low salaries have relatively few professional and economic incentives to improve quality. One may wonder whether the Spanish system can be sustained at salaries that represent only a third of the income received by the nephrologists of the other countries and without stronger professional incentives.

0347 International comparisons of health system performance with aggregate data Luoma K. (Government Insitute for Economic Research, Helsinki, Finland) Räty T. We assess the performance of health care systems with data envelopment analysis (DEA) using different sets of output and input variables available in OECD Health Data and data on mortality amenable to health care. We investigate how sensitive efficiency scores and rankings are with regard to which input and output measures are included in the analyses. In alternative model specifications output is measured either by life expectancy at birth, infant mortality, mortality amenable to health care, disability adjusted life expectancy or some combinations of them. The alternative input variables include total health expenditure, health employment, the number of doctors and nurses. In most of the analyses we use pooled data for 24 OECD countries from years 1998-2002. The analyses which use mortality amenable to health care as an output variable use only 1998 data. DEA results and country rankings based on them proved to be quite sensitive to which input and output variables were included in the model. In applying DEA for assessing the performance of health care systems one should be careful and parsimonious in selecting input and output variables. The excess number of variables and variables loosely connected to health status tend to inflate efficiency scores and many countries can be deemed to be efficient by default.

0188 The impact of service provision on patient movements in a list patient system Luraas H. (Health Service Research Unit,University Hospital, Lorenskog, Norway) Iversen T. In the Nor wegian list patient system general practitioners (GPs) are paid by a mix of capitation and fee-for-ser vice. Some GPs experience

that their preferred number of patients is greater than the number they actually have on their lists. In previous work we have shown that fee-forservice encourages these GPs to increase the service provision to their patients. In particular, we find that they provide the better paid consultation of a long duration more frequently than GPs with enough patients. In the present paper we focus on a related economic motive If patients prefer many services rather than few, the provision of extra services may discourage patients from switching to other GPs. By means of Generalized Least Squares Regression that both takes unobserved heterogeneity and heteroskedasticity into account, this hypothesis is tested on panel data covering all Norwegian GPs during the period 2001-2004. We are not able to show any effect of service provision on patient switching. Hence, extra service provision does not compensate for negative characteristics (patients’ impression of competence, empathy etc) of less popular GPs. Hence, patients’ revealed preferences show that benefits are less than costs at the margin, and we conclude that extra service provision is socially inefficient.

0433 Expenditure, Organization and Outcomes in Health. America and Europe Maceira D. (Center for the Study of State and Society (CEDES), Buenos Aires, Argentina) Barbieri E., Lignelli B. Health care systems widly differ in their coverage, financing and outcomes, since they are associated with the average richness of each nation and its distribution. As a result, the objective is to compare health care systems in four different regions: Latin America, United States and Canada, Western and Eastern Europe, identifying the main characteristics related with the socioeconomic context ruling in each region. The paper also analyses the association between income and participation of the public financing and providing of health care services. The results support the hypotheses of the positive association between the GDP per capita and the health care systems performance. In this way, the effectiveness of the health care systems in poor nations is questioned, taking them to face two challenges in light of rich countries’ experience: (I) to build a health care structure that compensates the financial limitations giving the best services and (II) to enforce the institucional capability of the state to guarantee the good working of the health care sector, equity in the access and quality to all the users.

0002 Development of a Capitation Scale for IDF Career Soldiers in Israel Magnezi R. (Department of Health Systems Management, School of Health Sciences, The Ariel College of Judea and Samaria, Medical Services and Supply Center, Medical Corps, Israel Defense Forces, Israel) Weiss Y., Cohen Y., Shmueli A. The Israeli National Health Insurance Law allocates a national healthcare budget to the sickness funds, which provide the medical care the civilian population. Medical care for members of the IDF is financed through the budget of the Ministry of Defense and is not included in the national healthcare budget. Benefits provided to soldiers serving in the permanent forces are far more extensive than those provided to civilians. Because of no co-payments, poor management, and the costbased budget, military healthcare costs in Israel are expected to exceed civilian healthcare costs, adjusting for age and sex. The present paper derives age- and sex-based capitation rates for military personnel, and compares military and civilian age-based expenditure and capitation rates. The study population comprised career soldiers and civilians aged 21–54 years. Expenses of career soldiers were calculated to provide inforEur J Health Econom Suppl 1 · 2006

| S63

mation on the financial costs of medical services for each age group in 2003. Overall expenses for women were higher than for men in all age groups. As expected, the older the group, the higher the total expenditure for both men and women. In-patient care represented a higher percentage of the total costs for men (28.3%) than for women (22.1%). Emergency room care was higher for women in the 22–24 age group but comparable to that of men in higher age groups. Specialist visits represented a significantly higher percentage of the total costs for women than for men in the 22–24 and 25–34 age groups (by 6% and 15%, respectively). The difference decreased to 13% in the 35–44 age groups and, in the 45-54 age group, the difference for men was 14% higher than for women. Military costs were similar to civilian costs in the 22–24 age groups, higher in the following two groups, and lower in the 45–54 age group. Like in other organizations, military healthcare ser vices might benefit from outsourcing. The inequality in medical services to soldiers and civilians, the over-use of the military healthcare system, and the decrease of standards and budgetary resources will compel the establishment of more creative means of providing these services through contracts and agreements, perhaps through the civilian sickness funds. Keywords: Capitation; Health Services; Israel Defense Forces (IDF)

ease severity that equate to levels that a psychiatrist would regard as a case likely to require treatment and exclude many with mental health problems. People with symptoms below this threshold often suffer distress and impairment of functioning and many of these may not have access to mental health care. Their needs are the ‚hidden needs’ which are rarely taken into consideration in the planning of mental health services. Results show that while there is inequality in the distribution of mental health problems among social classes, there is no clear evidence of inequity in the use of services. The degree of association between psychiatric morbidity and social classes varies between diagnostic groups and the gap between self-perceived and clinically assessed needs and utilisation of services is remarkable.

0607

The aim of this study was to evaluate a four year catch-up hepatitis B virus (HBV) vaccination campaign for high-risk adults at risk-group level, starting in the Netherlands in 2002. Groups considered were men having sex with men, heterosexuals with a high rate of partner change (HRP), commercial sex workers and hard drug users (HDU). Based on observed data collected in the first two years, and extrapolating these data for the two following years, total number of participants and program costs were estimated. By applying linear regression to the data collected on anti-HBc prevalence, the probability of susceptible individuals in risk group i to become infected with HBV in their remaining life was estimated. Correcting for observed vaccination compliance, the number of future HBV infections avoided was estimated per risk-group. By combining these numbers with estimates of life years lost, quality of life losses and health care costs of HBV infections, the benefits of the campaign were estimated for each risk-group. The program was cost-effective in all four risk groups. Vaccinating HDUs was most cost-effective with € 6,700 per QALY gained, whereas vaccinating HRPs was least cost-effective (€ 15,200 per QALY gained).

The cost-effectiveness of infliximab fro rheumatoid arthritis in Hungary Májer I. (Corvinus University of Budapest, Hungary) Brodszky V., Péntek M., Gulácsi L. Objec tive: To estimate the cost-ef fec tive ness of infliximab plus methotrexat (MTX) compared with MTX plus placebo in the treatment of rheumatoid arthritis in Hungary. Methods: Markov simulation model was adapted to assess costs and quality-adjusted life-year (QALY) values in the long run. Baseline distribution of candidates for biologic therapy according to the Hungarian guidelines, utility and costs of each disease severity state were obtained from a multicentric cross sectional survey in Hungary. Published transition probability data of the ARAMIS (Arthritis Rheumatism and Aging Medical Information System) cohort and the ATTRACT (Anti-Tumor Necrosis Factor and Concomitant Therapy) infliximab clinical trial were used. Sensitivity analysis was performed to ascertain the variability of the results. Results: Assuming 54 weeks of treatment the incremental QALY was 0,263 in the first year and 0,493 on lifetime horizon with an incremental cost of 12,390 € and 12,080 € respectively, resulting in 45,355 €/QALY or rather 24,485 €/QALY. Applying other modelling assumptions according to cost scenarios, discount and mortality rates, the incremental costeffectiveness ratios (ICERs) remained in a reasonable range. Conclusion: Compared with other published results, ICERs based on Hungarian data were lower that suggests lower Hungarian costs and resource use.

0485 Social class and mental health of the UK population 1993–2000 Mangalore R. (London School of Economics, London, United Kingdom) Knapp M. This paper analyses the distribution of mental health problems and use of services in the UK, by social classes, using data from Psychiatric Morbidity Sur veys 1993 and 2000. The aim is to present the changes between the two years, estimate the extent of unmet needs and evidence of inequity in the system. Concentration index approach is used for the analysis of equity issues. A distinction is made between self-perceived and clinically assessed mental health problems as the rules of international classification systems for psychiatric diseases set thresholds of dis-

S64 |

Eur J Health Econom Suppl 1 · 2006

0276 For which high risk adults is a catch-up HBV vaccination campaign cost-effective? Mangen M.J. (National Institute for Public Health and the Environment, Bilthoven, The Netherlands) De Wit G.A., Kretzschmar M.

0354 Proposal for an unconventional risk-adjusted cpitation formula for allocating funds to regions in Italy Mapelli V. (University of Milan, Italy) Current formulae for allocating resources from a central fund to subnational governments or to health schemes are based on multivariate regression models that adjust for risk or need, but fail to consider the variability of the average costs of services provided, thus ending up by “redistributing” inefficiencies and wastes across health authorities. The proposed formula for the Italian NHS explicitly considers the determinants of health expenditure in the following way: C = π · f · q · c, where C is capitation, π is the probability of being (chronically) ill, f is the frequency or propensity to use specific services when ill, q is the quantity and c is the unit cost of services used. Epidemiological studies show that almost 80% of expenses is determined by chronically ill and by dead. The capitation can be calculated either at an aggregate level, by multiplying the regional prevalence rates of 15 major chronic diseases times the national average treatment cost (f · q · c) for each disease, thus recognizing the difference in regional risks/needs; or at an analytical level by setting standard or target values for (f), (q), and (c), thus correcting also for excessive frequencies, inappropriate quantities or unnecessary costs. The basic information for the formula can be derived from administrative databases containing individual records on chronic diseases and expenditure.

0414 Voluntary Private Health Insurance and Equity in access to health care Margherita G. (Department of Economics, Finance and Statistics, University of Perugia, Italy) Masseria C. As in most EU countries, the last ten years have been characterised by a trend in shifting health care costs from the state to individuals. The Italian health care systems (SSN) provides comprehensive health care coverage. Despite this, health care expenditures have been characterised by the increasing importance of the private component, including Voluntary Health Insurance (VHI). There is evidence at EU level that access to complementary VHI is problematic for people with low incomes. aim of the paper is to measure, income-related horizontal inequity in the access to voluntary health insurance in Italy. In order to standardise for individual needs we use a measure of inequity which is based on the indirect standardization approach proposed by Wagstaff and Van Doorslaer (2000). The data are taken from the Multiscopo survey, 19992000. Income information, was imputed by using the Eurostat sur vey (1999) for Italy. Horizontal inequity indices for VHI are generated comparing them across regions. Results show that access to VHI is subject to income-related inequity in most Italian regions. We measured the relative contribution of VHI to inequity in access to inpatient, GP and specialist services. It appears that VHI access is a source of inequity. This should raise concerns on increasing the role of VHI in Italy.

the next 20 years. The paper looks at options for policy and macroeconomic interventions to improve performance relative to status quo.

0558 The use of Patient-Reported Outcomes to support a regulatory claim in the USA and Europe Marquis P. (Mapi Values, Boston, USA)

0066

The objective of this session will be to present FDA and EMEA patientreported outcome (PRO) submissions and reviews. Areas of convergence and divergence will be identified and discussed. The creation of the Study Endpoints and Label Development group (FDA/CDER/OND/ SEALD) three years ago and the development of a draft PRO guidance document illustrate the US regulators’ commitment towards developing and maintaining consistent criteria for review of PRO and other study endpoint evaluations. In the meantime, the EMEA has developed a reflection paper on health-related quality of life. So the FDA and EMEA have both gained experience in reviewing PRO and health-related quality of life evaluations. During the presentation, an emphasis will be put on practical experiences when assessing and submitting PRO/healthrelated quality of life data at different stages of the clinical program, with a focus on early phases (e.g. development of hypotheses, development/ validation of questionnaires, strategy for analysis and interpretation). Although most of the recommendations when implementing PROs in global clinical development programs are agreed upon by different regulators, issues related to definition of concepts, PRO paradigms, PROs as primary end-points, role in evaluation of treatment risk/benefit, crosscultural issues, and clinical significance, merit further discussion.

Giving greater independence to providers: does it make a difference?

0575

Marini G. (Centre for Health Economics, University of York, UK) Jacobs R., Marini G. Foundation hospitals are a new type of organisation introduced into the English National Health Ser vice in 2003. Foundation status offers hospitals financial and managerial independence. The rationale is that giving hospitals more self-determination will provide incentives to improve performance. Foundation hospitals are being introduced in a phased manner – only the highest performing hospitals are allowed to become foundation hospitals. At present there are 25 foundation hospitals (from a total of 173 acute hospitals), but within 5 years the government expects all acute hospitals to be performing at levels high enough to qualify for foundation status.panel data for English acute hospitals from 2000/01 to 2004/05, we investigate the characteristics of the “early-adopters” : those hospitals that are in the first and second waves of reform : with those of other hospitals. Difference in difference methods are used to examine the difference between hospitals before and after the adoption of foundation status on a range of characteristics. The foundation hospital scheme has been criticised for generating inequality in the distribution of healthcare supply (a “two-tier” system). Our analysis sheds light on this and is generalisable to other devolved systems in which hospitals achieve more independence and freedom from central control.

0590 Economic Costs and Impact of Non-Communicable Diseases: Experience from Russia

An economic evaluation of physical activity interventions Marsh K. (Matrix Research and Consultancy Ltd., London, UK) Objective: NICE commissioned Matrix RCL to determine the cost effectiveness of four types of intervention aimed at increasing physical activity levels: brief interventions in primary care; pedometers; exercise referral; and walking and cycling programmes in the community. Method: A review of effectiveness studies was used to identify the resource use and change in physical activity resulting from the interventions. An economic model was constructed to measure the health benefits of the changes in physical activity levels, and the consequent impact on participants’ quality of life and NHS cost savings. Results: Defining costs as just those incurred implementing the interventions, the model found that the cost per QALY gained estimates for the interventions varied from c£20 to c£470 (when compared with ‘usual care’). When costs are defined to include the future health care costs saved as a result of health states avoided, the interventions dominate usual care, both improving health outcomes and being less costly. The most cost-effective interventions are exercise prescription or interviews. Discussion: A number of important lessons are drawn for undertaking economic evaluations of public health interventions. A key challenge is the epidemiological modelling required to determine long-term health impacts of public health interventions. A number of pieces of information are crucial to this part of the model: an accurate measure of baseline risk of suffering health states; change in risks as a result of the intervention; and the impact of this change on risk of suffering health states.

Marquez P. (WHO) Shurke M. This paper would present results on costs and economic growth impact of NCDs in Russia. Costs are estimated at 3 percent of GDP and, at its current pace, NCDs could have a profound impact on GDP growth in Eur J Health Econom Suppl 1 · 2006

| S65

0084

0282

On insurance and the cost-sharing of pharmaceutical R&D

Sources of small area variations and the effects of self-dispensing practices in the use of antibiotics in the community

Martinez-Giralt X. (Universitat Autonoma de Barcelona, Bellaterra, Spain) Barros P.P. An old debate that is being developed since the late 1950s, is the link between the level of profits of a company (and thus, its monopoly power in the market) and the source of funding of R&D activities. The recent years have witnessed the uprising of the globalization of economic activities in the developed world. Together with globalization, increasing costs associated with the pace of technological change force companies to review their R&D organization and spending. DiMasi et al. (2003) estimate the total R&D cost per new drug in 2001 at $802 million. The Tufts Center for the Study of Drug Development (2003), increases the estimate further to $897 million in 2003. The sharp increase in R&D expenses combines in the case of the pharmaceutical industry, with the traditional differences of drug prices across countries due to a number of factors. Among them, different regulatory regimes and insurance systems, together with the fact that governments are usually monopsony buyers of drugs. In reply to this new environment, companies have reacted developing partnerships, and proposing a single price across countries (at least within the EU). On their part, governments in high price countries have also reacted by introducing these prices differences as an element in their negotiations with pharmaceutical corporations We propose to focus on the issue of how to impute the overall R&D costs in the pharmaceutical industry across countries. Ramsey pricing has been proposed in the pharmaceutical industry as a mechanism to price discriminate among markets while allowing to recover the (fixed) R&D cost (see e.g. Danzon and Towse, 2003). However, such analyses neglect the role of insurance and the associated moral hazard distortion of (the elasticity of) demand. By incorporating this new element, we aim at providing some building blocks towards an economic theory merging Ramsey pricing, equity concerns by governments and the strategic incentives, as governments also determine the reimbursement level in countries with a NHS-like system.

0111 Evaluation of comorbidity indices in risk adjustment models for inpatient mortality Martins M. (National School of Public Health/Oswaldo Cruz Foundation (Brazil) and Medicine School/Montreal University, Canada) Blais R. Objective: To evaluate the validity of severity measures based on comorbidities, comparing the Charlson comorbidity index (CCI) to a new empirically developed index. Methodology: The research design was a retrospective cross-sectional study. The study was limited to the Ribeirão Preto region in the State of São Paulo, Brazil, from January 1996 to December 1998. We used information from the database on hospitalizations of Ribeirão Preto University, covering 100% of hospital admissions in the region. We included only admissions in which the principal diagnoses were respiratory and circulatory diseases. Hospital admissions of patients less than 18 years of age were excluded from the analyses. Based on these criteria, the final sample consisted of 69,302 hospitalizations. Results: Evaluation of the CCI indicates that revision of the clinical conditions studied by Charlson, as well as their weights, increased mortality model predictive capacity. In addition, the development of a new index through the utilization of other comorbidities excluded from the original CCI allowed the improvement of the models’ discrimination. Conclusions: The results of the validity analysis for comorbidity indices favor the utilization of empirically developed indices. However, age and principal diagnosis are the most important predictors of mortality risk.

S66 |

Eur J Health Econom Suppl 1 · 2006

Masiero G. (Faculty of Economics of Lugano, Switzerland) Filippini M., Masiero G., Moschetti K. Health authorities worldwide are concerned with the increasing effects of bacterial resistance and the inappropriate use of antibiotics. Wide differences in the utilization of antibiotics in ambulatory care are observed across European regions. Filippini, Masiero and Moschetti (Health Policy, forthcoming) recently investigated the factors affecting regional variations in outpatient antibiotic use in Switzerland. However, the literature lacks empirical evidence on the determinants of small area variations. We investigate the sources of small area variations in outpatient antibiotic use and address the issue of inducement in antibiotic prescriptions by self-dispensing practices. We sketch a theoretical model of the demand for antibiotics capturing the interaction between patients and general practitioners in different market areas. The main hypotheses are tested using an econometric model. The dependent variable is the per capita outpatient antibiotic sales (240 Swiss areas for the year 2002, quarterly data) whereas the explanatory variables include socioeconomic and epidemiological factors, access to practices and to antibiotic treatment, and aspects of practice regulation and practice style. We find that not only do self-dispensing practices improve access to health care services, but they may induce higher levels of antibiotic consumption in the community. Following Phelps and Parente (1990) we then estimated that the welfare loss from small area variations in the use of antibiotic is indeed remarkable.

0200 Inequality in health in 10 European countries Masseria C. (London School of Economics and Political Science, LSE Health and Social Care, London, UK) Allin S. Using a unique, comprehensive dataset across 10 European countries, Sur vey on Health Ageing and Retirement in Europe (SHARE), this paper measures socio-economic inequalities in health among the over 60 across 10 European countries, highlighting the role of different determinants such as gender, education, income, wealth and participation in social activity. This study aims also at addressing a main methodological issue the choice of an appropriate indicator in the analyses of inequalities in later life among income, education and wealth. results show that inequalities in health do not disappear in later life. Health status deteriorates when age increases and women tend to report worse health status than men. Smoking and BMI are statistically significant determinants of limitations in daily activity in many countries, while participation in social activity (voluntary work, caring activity, community life, courses, sports, etc.) is a protective factor. Individuals with higher education level enjoy better health in all countries, and the inequality is particularly large in Denmark, Greece and Austria. Income and wealth are important determinants of health disparities among older people as well. The index of income-related inequality is positive everywhere, and particularly so, in Denmark and Sweden. Similar results are found for wealthrelated inequality.

0214 Socioeconomic and geographical differences in antenatal screening practice in Italy Masseria C. (London School of Economics and Political Science, LSE Health and Social Care, London, UK) Socioeconomic differences in health care use have been the focus of several studies, but antenatal care has not received sufficient attention. However, inequalities in access to health care in the antenatal and perinatal period are likely to contribute to health inequalities in later life. The aim of this study is to investigate the role of mother’s education level and region of residence in the take-up of antenatal screening in Italy using data from the Multiscopo sur vey. Antenatal screening has become more common in recent years. National guidelines are vague, and each region decides its antenatal screening program; large variations are found across geographical areas. Antenatal screening is more common in the North 90% of pregnant women undertook morphologic-fetal ultrasound in the North compared to only 76% in the South. Further discrepancies are found across regions in each geographical area. Moreover, it appears that pregnant women with higher level of education have almost four times greater probability of receiving information about available screening tests, and are more likely to undertake tests such as amniocentesis, tri-test and fetal ultrasound than women with lower level of education across all ages and regions.

0157 Perverse incentives from waiting list management programmes : evidence based on DRG data Mateus C. (Escola Nacional de Saude Publica, Lisboa, Portugal) Barros P.P. Waiting lists are present in many western countries, and in general constitute a major political issue. Governments are pressed to find some solution. One such solution was implemented in Portugal batches of the waiting list were auctioned. Both private and public facilities were allowed to enter the auction. In particular, public hospitals holding waiting lists were not excluded from treating their own patients, outside the normal schedule of course. Naturally, treating these patients was paid on top of the typical fixed salary (which is independent of the level of activity). It is clear to see that perverse incentives may set in, with possible reductions in normal activity in order to treat the more compensating “waiting list” cases. We term this effect “crowding-out”. We empirically evaluate whether this is the case, or not. We make use of a detailed data set, based on DRG activity, for both normal and “waiting list programme” treatments in public hospitals. Our econometric analysis shows that indeed crowding out was present. This has, of course, strong policy implications for waiting list management programmes.

0541 Two decades of case mix in Portugal: where are we now? Céu M. (National School of Public Health, Nova University of Lisbon, Portugal) In 1984 DRGs were formally introduced in Portugal and the aim was the development and implementation of a fully integrated information system for the management and financing of hospitals. Portugal, as other European countries, faced raising costs and increasing pressures from the demand side in the hospital sector when it was decided to adopt DRG as a management and payment tool. Adoption was decided at the MoH and a top-down approach profiting from a very centralized political system was chosen. It is sought to present the rationales behind the option for a prospective payment system using DRGs, as well as some results and controversies related to the implementation of such system.

Major achievements of implementing DRGs are more related to benefits in the management of hospitals and building of a national database with demographic and clinical data. The Portuguese experience shows that to implement a PPS more is required besides DRGs. A case mix culture has been created but further developments will depend on the support by the MoH to research in this area.

0093 Russian healthcare reform during the 1990s – what lead to the institutional trap? Mathivet B. (University Paris I Pantheon Sorbonne, Paris, France) Our main goal is to understand why, after more than ten years of introduction of the reformed mechanisms, the Russian healthcare system is unable to cope with the sanitary challenges of the economic transition, and in some cases even contributed to worsen the situation.reform was dual and consisted first in the decentralization of the former centrally and budgetary financed system to the regional and local authorities. The second aspect of the reform was to introduce a different source of financing (payroll) and new means of coordination in the system including the introduction of market regulation (with private insurance companies as the main actors of the Mandatory Health Insurance System). This system was supposed to quickly supply the majority of the necessary resources. Market mechanisms were also supposed to be able to ensure both quality and cost-control of medical practices.analyze the confrontation with initial conditions (main characteristics of the former system, and socio-economic conditions at the date of introduction of the new system), in order to show that it inevitably lead to an institutional trap, with lack of incentives, few sanitary improvement, and a worsening of inequalities of access to healthcare. paths for future reforms are proposed.

0365 Do new technologies cost more? Outcome and economic results of the GERSHWIN-Study (German Stent Health Outcome and Economics Within Normal Practice), a controlled evaluation of drug-eluting versus bare-metal stents for the reduction of coronary restenosis McBride D. (Charité University Medical Center of Social Medicine, Epidemiology and Health Economics, Berlin, Germany) Brüggenjürgen B., Willich S.N. Objective: To evaluate long-term implications of drug-eluting stents (DES) vs. bare-metal stents (BMS) in the treatment of coronary artery disease (CAD). Method: In this prospective intervention study in 35 hospitals in Germany, CAD patients undergoing PCI were treated with BMS or DES. Standardized patient and physician questionnaires 3, 6, and 12 months following PCI documented major adverse cardiac events (MACE), disease-related costs and patient quality of life. Results: 602 patients were treated with DES (mean age 63±9, 87% male) and 295 with BMS (mean age 65±10, 79% male). There were no significant baseline differences in cardiovascular risk factors and severity of CAD. After 12 months, 18% DES vs. 30% BMS patients had suffered MACE (p<0.05). Initial hospital costs were higher with DES than with BMS (5,939±85 vs. 3,730±78 Euro), but respective 12-month follow-up direct and indirect costs were lower (5,052±642 vs. 6,052±590 Euro and 753±459 vs. 2,013±422 Euro). Overall, disease-related costs were similar in both groups (DES 11,765±827, BMS 11,826±760, p = ns). Conclusions: Compared to patients with BMS, patients with DES experienced considerably fewer MACE during 12-month follow-up. Despite the higher initial costs associated with DES compared to BMS, the overall costs 12 months following PCI were similar.

Eur J Health Econom Suppl 1 · 2006

| S67

0456 Healthcare Output and Productivity Measurement McCormick B. (Department of Health, London, UK) The Department of Health in England published a paper in December 2005 reporting work to improve measures of health output, used in the National Accounts, to include measurement of quality change. This was a response to the Atkinson Report on Measurement of Government Output and Productivity for the National Accounts (January 2005). The DH paper, Accounting for Quality Change, combined a summary of research commissioned from York/NIESR (separate ECHE submission) with proposals for other quality adjustments, including annual rise in value of health in line with GDP per head, a value (rather than cost) weight for prescribed statins, evidence of clinical improvements in primary care, sur vey evidence of improved patient experience and longer term survival from heart attacks. Measures were acknowledged as provisional and in some cases controversial. Further development plans were described. A presentation at Budapest would also report on subsequent work and discussions, with links to separate proposal by the UK Centre for the Measurement of Government Activity.

0535 Evaluating the cost effectiveness of suicide prevention strategies: practical and methodological challenges McDaid D. (LSE Health and Social Care, London School of Economics and Political Science, London, UK) Knapp M., Halliday M, Mackenzie M., Maxwell M., McCollam A., McLean J., Platt S., Woodhouse A. Reducing the suicide rate is a key public health target across Europe, where it is no coincidence that the highest suicide rates in the world can be found. A systematic review was undertaken to identify information on the development of national and area based suicide prevention strategies worldwide, ascertain the extent to which they had been subject to any type of economic evaluation, and consider in the absence of such information how best to aid policy makers in assessing the value for money of such programmes. Results indicate that few area based initiatives to tackle suicide had been implemented. Fewer still have been subjected to evaluation either of effectiveness or cost effectiveness although some evidence is beginning to emerge. Threshold analysis however indicates that if modest success can be achieved in reducing suicides then many programmes have the potential to be highly cost effective. A number of practical and methodological challenges for the use of economic evaluation in this area are discussed. As with many other areas of population level public health interventions cost benefit analysis may well be the most suitable approach to adopt. Much may also be learnt by comparing long term trends in suicide with changes in socio-economic risk factors such as economic recession and structural transition within countries.

0329 Visiting the Doctor, Noise and Self-Assessed Health McGregor P. (School of Economics and Politics of Ulster, Belfast, Ireland) McKee P., Neill C. Self-assessed health (SAH) is frequently used as a noisy, scalar measure of health status. This paper examines its role in a model of GP utilization in England employing data from the British Household Panel Survey (BHPS). The framework developed is one of utility maximization within which SAH is treated as a measure of utility, conditioned upon health status. This provides a different rationale for the inclusion of socio-economic variables than in conventional analysis.of SAH requires an alternative and more ‘objective’; measure of health status to

S68 |

Eur J Health Econom Suppl 1 · 2006

purge the errors in the former. This is provided by self-reported health conditions (of which there are 15 in the BHPS) and whether the individual attended hospital out-patients. In the UK the GP acts as a gatekeeper for such consultations and thus attendance implies that a threshold level of severity of the condition has been exceeded. empirical exercise consists of a simultaneous model of GP utilization where SAH and out-patient attendance comprise the additional equations. This allows health status to be measured by two variables in the GP utilization equation since the effects of the individual conditions is now contained in these variables. The results suggest that SAH is the dominant influence and that the role of socio-economic variables has important consequences for the consideration of inequity in utilization.

0385 Using a discrete choice experiment to design a treatment profile for peripheral neuropathic pain McIntosh E. (University of Oxford, Institute of Health Sciences, Economics Research Centre, Oxford, UK) Ossa D.F., Briggs A., Dworkin R., Collett B., Marciniak A. As part of the development of new treatments for a disease, the impact of the disease and its treatment on patients’; health-related quality of life should be explored. The aim of this study was to estimate the value of hypothetical treatments for peripheral neuropathic pain (PNP) in terms of attributes important to clinicians and patients. A discrete choice experiment (DCE) was designed and administered to a sample of clinicians and patients in the US and UK. Attributes related to PNP were identified from a literature review and through interviews with clinical experts and patients. Final attributes for the DCE, which were the most important to patients, include pain, quality of sleep, general activity, mood, concentration, daytime drowsiness, dizziness and nausea. In addition, an outof-pocket cost attribute was included in order to estimate willingness-topay values. The survey returned 354 questionnaires. Final analyses were performed on 10,620 observations that showed consistency, reliability and face-validity using a random-effects probit model. Most attribute levels were significant predictors of choice with non-linear effects observed across different attribute levels. These results could be used to aid decisions regarding optimal design of a treatment for PNP, or as a basis for a cost-benefit analysis within clinical trials.

0361 The generic cost-effectiveness of health care Meerding WJ. (Department of Public Health, Rotterdam, Netherlands) Mackenbach JP. Objective: To assess the generic cost-effectiveness of health care. Methods: We assessed the contribution of prevention and medical care towards population health, their costs and cost-effectiveness, for selected disease clusters: infectious diseases, cancers, and cardiovascular diseases. The current burden of disease (BoD) was calculated using the GBD methodology, combining incidence, prevalence, mortality, and disability weights into DALYs. Epidemiological estimates were made internally consistent with advanced incidence-prevalence-mortality modelling techniques (DISMOD2). We then calculated the BoD in a ‘null’ situation without prevention and medical care with use of historical data and literature on incidence, survival, mortality, and prevalence of risk factors, combined with trend analysis, and accounting for the role of autonomous factors (e.g. improvements in housing and nutrition). The difference between the current and ‘null’ BoD was combined with cost of illness data to estimate the generic cost-effectiveness of medical care for the selected disease clusters. Results: available in due course. Conclusions: The results mainly depend on the definition of medical care, and the estimated contribution of autonomous trends.

0582 The Answer of Health Unified System in Brazil and its financing challenges to achieve the Millenium Development Objectives Mendes Á. (Faculty of Economics of Faap, Sao Paolo, Brazil) Marques R.M. Brazil is achieving rapid progress in the direction of the most Millenium Development Objectives. The collected data of the national information systems shows that the country will achieve objectives related to poverty, hunger, education and its reduction of sex differences, if social and economic policies follow the same way of the present trends. Nevertheless, the country has important challenges for the near future, as in the case of health issue, which demands hard work, mainly in Health Unified System (HUS) financing. Alhtough the necessary efforts to achieve the main objectives have been drawn, it is important to say that it is not enough to have the programme defined. It depends on the financial responsability of each level of government – federal, state and municipal, all HUS managers –, specially from the federal level which accounts for the biggest participation in the total of health public expenditure in the Brazil. After 18 years of Health Unified System institutionalization on brazilian constitution, its annual definition of financing is not free of conflicts. These conflicts take place int the struggle by resources between the different areas of Social Security and / or in the appealing attempts of the economic federal government team in reducing its expenditure or in changing the definition of what is understood by public health services and actions. Even after the Constitutional Ammendment 29/2000 approval, the objectives of the present federal government economic policy restrain its development, placing in schock with the fundamentals principles of Social Security as well as HUS. This presentation tries to discuss that Brazil will face difficulties in achieving the goal which refers to health issue because it depends on the definition of HUS institutional financal rules, in order to garantee the responsability of federal government in the system financing.

0532 Economic Evaluation for Colorectal Cancer Screening Mennini F. (University of Rome Tor Vergata, Italy) Palazzo F., Federici A. The aim is a cost effectiveness study. The population subjected to the screening is for all three models from 50 to 64 years of age. A study modality has been adopted that refers to a Quality Management (QM) system unrelated to the study typologies previously indicated. The screening strategies have been compared utilising threshold analysis, to identify values of certain parameters and make the choice of screening strategy indifferent compared with any other. The model that we are implementing has characteristics in common with the most important studies analised even if adapted to the Italian system. The aim is to use cost for life gained. Later, should the nature of the information permit us to use QALY, it would also be feasible. It can be confirmed that the aim is a cost effectiveness study.

0602 DRGs in Italy: perspectives and problems Mennini F.S. (University of Rome Tor Vergata, Italy) Spandonaro F. In Italy a financing system based on DRGs was introduced in 1992. A Federal/Regional process (2001) shifted upon Regions the burden to manage the fee system. The use of that system during the last ten years permits us to evaluate: 1) the effects of DRG system on hospital activity;

2) the effects on the integration of supply; 3) the differences of impact as a consequence of the public structure of NHS in Italy vs the primary experience in USA; 4) the relationship between ageing and hospital admission with the DRG system; 5) the use of fee policies, autonomously, from Regions to guarantee financial stability of the Regional health systems. In conclusion it is possible to argue that: a) the Italian DRG system applied upon public health system has not created such competition as in USA; b) the supply integration goes slowly in respect of the expectations with negative relapses, particularly for the elderlies; c) Regions use the fee system to create dumping; d) the introduction of the DRG system increased the attention toward the economic aspects of health system; e) the DRG system represents an important instrument to increase hospital productivity.

0493 Measuring people’s preferences for distribution of health. On the methodology. Mestad O. (Chr Michelsen Institute, Bergen, Norway) Norheim O.F. Several empirical studies have demonstrated that people do not evaluate health programmes solely based on aggregate health gains they also care about the distribution of health. Distributional concerns can be incorporated into cost-effectiveness analysis by using distributional weights, but existing empirical studies provide little guidance on how this can be done in practice. In this paper, part of the reason for the weak link between the empirical literature and applied cost-effectiveness analysis is ascribed to the lack of a coherent analytical framework that can build this bridge. For instance, several empirical studies fail to distinguish appropriately between marginal and infra-marginal health gains. We show that this failure may lead to seriously misleading estimates of distributional weights, and we propose a way of dealing with the problem.

0391 Physician authority under managed care: A theory of utilization review Meyer E. (LMU München, Abteilung für Versicherungswissenschaft, München, Germany) Utilization review (UR) is an important instrument used by managed care plans and other health insurers to monitor physicians’ use of health care services. It is widespread in the US, but also established in several European health care systems. For example, in terms of prospective UR, physicians need insurer authorization for hospitalizations or other highcost treatments. Sometimes, patients who seek surgery are also required to obtain a second opinion. In the US, these practices have aroused physicians’ and patients’ anger about interventions into medical decision making, resulting in the socalled managed care backlash. Despite the attention paid to it, empirical studies suggest that the cost-reducing effect of UR is rather low and that many insurers prefer a non-intrusive review style with low denial rates. Building on the principal-agent theory of authority introduced by Aghion and Tirole (1997), this paper proposes an explanation for the empirical facts. The model is adapted to a context of managed care, considering the physician’s diagnosis and referral behavior under a standard cost-sharing contract. It is shown that the insurer’s (or a third-party reviewer’s) effort to identify unnecessary referrals will have a negative effect on the physician’s initiative, as measured by his diagnosis effort. Therefore, it is efficient Eur J Health Econom Suppl 1 · 2006

| S69

for the insurer to commit to a lower level of monitoring. This can be achieved, for example, by delegation to a third party who has weaker incentives to monitor.

0075 Does insurance make us blind? Milcent C. (PSE, Paris, France)

0444 Costing hospital services in the European Union: results from the HealthBasket project on the review of concepts, methods, and feasibility Mogyorosy Zs. (Centre for Health Economics, University of York, UK) Smith P.C.

How the variation of insurance coverage and the insurance regime can affect the consumer_search behavior of the drug price ? In France, 64 % of individuals declare they do not know the price of the use of healthcare. Here, the effort is defined by the consumer_search behavior. Without insurance, the effort is optimal and we show that the presence of insurance market deteriores this level of effort that corresponds to the moral hazard phenomena. The distorsion due to the moral hazard is the part of the price that the agent do not realise to spend, a sort of a blindness due to the insurance. Whatever the form of organization, this distorsion is always higher for the high-risk type in presence of insurance market. The overall drug expenditure is also affected by the moral hazard due to the presence of the health insurance market and its forms of organization. The extra overall drug expenditure is the part of the price that agents would not spend without insurance market. This extra overall expenditure depends on the participation of the high risk and low risk types. With a voluntary insurance market, the high-risk can be the only type to have access to the prescription drug. Therefore, this extra overall expenditure leads to an increase under a compulsory insurance system and may lead to a decrease under a voluntary one.

Aims and objectives: One of the aims of the EU-funded Healthbasket project is to provide a comprehensive assessment of best practice in costing health ser vices with a special focus on international comparative studies. Methods: English language publications were identified and critically appraised using pre-set quality criteria through searches of electronic databases, Internet websites, local library files and reference lists of published articles and books. Results: The review revealed that the choice of costing methods is guided by multiple criteria, and accounting cost is frequently used as an estimate of opportunity costs. A top-down costing approach can be reasonably accurate for ser vices where marketed health technologies are responsible for most of the resource use, whereas a bottom-up or mixed approach is to be preferred when service provision is based on complex organisational arrangements. Furthermore, accurate cross-country cost comparison can be compromised by differences in accounting systems, national costing guidelines, provider payment systems, and the lack of standardized unit cost databases. Conclusions: Researchers and decision-makers should carefully consider the trade-offs between information accuracy, comparability, and the costs of securing that information, in order to assess whether the value of detailed cost data justifies the additional costs of obtaining them.

0517

0521

Hospital debts in Poland – public assistance – attempt of structural solution Miskiewicz P. (WHO Country Office in Poland, Warsaw, Poland) Kuszewski K. In 1999 Poland introduced universal health insurance system. Before that, state liquidated all hospitals’ debts but not the main causes. Most unfortunately Bill on “203”(Growth of Average Wages in Enterprises) 2001 passed through Parliament, without indicating source of financing. Apart from increase of labor costs, costs of medical procedures, cost of pharmaceuticals increased as well and resulted in running hospitals into debts again, generating additional credit interest rate burden. Hospital debts reached in middle 2005 value of 6 billion PLN. Parliament approved the bill on “Public Aid and the Restructuring of Public Health Care Establishments”. The bill regulated the issue of dealing with public law liabilities, civil law liabilities and individual claims of employees. To apply for public support hospitals were obliged to develop and present the restructuring plan. Funds secured for public help were approximately 2.2 billion PLN. From 740 hospitals one third was in good condition, 240 were holders of 80% of debts. Analysis of debts structure, management, restructurisation and level of contracting with National Health Fund will enable to asses weather public support is helping hospital to function and how many hospitals will in reality benefit from it. Rationalization elements are hospital network, change in financing EMS and optimalisation of drug purchasing.

Patients’ preferences for Choice of Hospital Monstad K. (The Norwegian School of Economics and Business Administration, Bergen, Norway) What determines patients’ choice of hospital, in a setting where hospital stays are rationed by waiting lists and where travel distances within the country are substantial? Through a reform implemented in 2001, Nor wegian patients are given generous formal rights to choose hospital for elective treatment throughout the country. This presentation is an attempt to infer the willingness to pay for shorter waits by studying the observed allocation of operations. The trade-off between distance and quality is likely to differ according to patient characteristics and socio-economic background. Patients’ preferences are examined using a unique data set with individual patient data on one specific patient group, namely elective hip replacements in Norway during the years 2001-2003. After a discussion of the institutional setting, the paper focuses on the trade-off that the patients make between distance and waiting time, and explores whether quality competition may be traced in the Norwegian hospital sector. The main results are that distance and waiting time are both highly significant attributes, and that patients are willing to wait a considerable length of time to avoid travelling. Preferences are found to vary over time and according to age and gender.

0309 ACE-Obesity: Priority setting for obesity prevention Moodie M. (The University of Melbourne, Victoria, Australia) Haby M., Swinburn B., Markwick A., Magnus A., Tay-T.K., Sach T., Carter R. The ACE-Obesity study uses an evidence-based approach to evaluate interventions aimed at reducing the prevalence of obesity in Australian youth. It informs decision-makers about the benefits of individual inter-

S70 |

Eur J Health Econom Suppl 1 · 2006

ventions and the packaging of a coherent strategy for obesity prevention and management. To avoid methodological confounding, the approach employs standardised methods including a two stage concept of benefit; a common comparator, setting and decision context; Australian data; and extensive probabilistic uncertainty testing. The technical cost-effectiveness results (cost per DALY) for each of the selected inter ventions will be reported. Modelling is undertaken to convert changes in behaviour to BMI outcomes and then to DALYs, and issues of the attribution of costs across multiple objectives arise. Due process is achieved by involving stakeholders on a Working Group, and by consideration of second stage filters (such as equity, acceptability and feasibility). The results are brought together in a ‘league table’ in which all the interventions are ranked in order of economic merit without the usual methodological concerns about results drawn from studies lacking in comparability. In packaging interventions to meet particular budget allocations, the divisibility, mutual exclusivity and returns to scale of individual interventions are considered, as well as issues of program logic, target group coverage and a range of settings.

0384 Variations in Health Expenditure Growth Trends across OECD Countries, 1992 – 2003 Morgan D. (Organisation for Economic Co-operation and Development, Paris, France) Orosz E. This paper, based on OECD Health Data 2006, discusses expenditure trends from a perspective of the pressure on governments to adjust public expenditure on health to economic growth. Public expenditure on health over the period 1997 to 2003 increased on a yearly average by 4.7% across OECD countries – twice the rate of economic growth – resulting in a considerable increase in the ratio of health expenditure to GDP. This compares to the preceding period (1992–1997) when the annual growth rate of public health expenditure was 2.4% – the very same as that of the economy as a whole. The first part of the paper examines the cross-country variations and the different causes behind these trends deliberate strategies to relieve tensions arising out of tough costcontainment policies adopted in the early to mid nineties; the failure of cost-containment policies in some countries; and, a third pattern in Germany, slow public health expenditure growth (1.5%) accompanied by even slower economic growth. second part of the paper looks at the links – and possible explanations – between the growth in public expenditure on health and GDP by comparing the European Union and the United States. Whilst the last decade has seen similar overall growth rates in health expenditure, the short term fluctuations in the annual growth rate have followed the changes in economic growth more closely in the European Union than is the case in the United States.

0394 Estimating patients’ willingness to travel for higher quality hospital care Moschetti K. (University of Lugano, Lugano, Switzerland) Balsan D., Rochaix L. This paper aims at developing a methodology to estimate patients’ willingness to travel for higher-quality hospital care. The estimation method uses as dependent variable the additional distance that patients are willing to travel to reach such a hospital when a similar service in terms of DRG competence is offered nearer to their homes. The willingness to travel is modelled as resulting from a two stage decision process or from one stage decision process. Referring to these two alternative hypotheses, we apply a generalized Tobit model and simple Tobit model to a large French database containing micro-level data of hospital discharges

for both the private and public sectors. Because France does not rely on a gatekeeping system to access secondary care, and after excluding emergency cases and transfers, one may legitimately assume that the choice of hospital is made primarily by patients themselves. Unlike previous studies, we use all DRGs to give an across the board estimate of patients’ willingness to travel for quality. The study provides estimates of distance versus quality elasticities and shows that perceived quality significantly affects patients’ willingness to travel. In so doing, our work contributes to a better understanding of patients’ preferences concerning hospital care and the role they may play in quality competition among hospitals.

0530 EFPIA HTA Key Principles: A Patient Perspective Mossman J. (Kilmacolm, UK) As new therapeutic approaches to disease management are developed, they need to be tested in robust clinical trials to assess their effectiveness. Increasingly, in an environment where healthcare costs are capped, the emphasis is on cost effectiveness in addition to clinical effectiveness. Incorporating health economics into the data collection of clinical trials has largely ignored the priorities which patients and their carers would assign to the benefits and unwanted effects of the treatments being tested. The data may be statistically reliable but may be irrelevant to the decisions patients make when faced with treatment options. The diverse range of factors which influence patients might be difficult to assess in the robust way in which sur vival is able to be judged, but is nevertheless important if patients are to adhere to the treatment protocols which will maximise the clinical effectiveness. There is, therefore, an imperative for researchers to take more notice of what matters to patients. Ideally, patients should be involved in the health technology assessment process from the trial design stage through to the formal assessment of efficacy and effectiveness thus ensuring that decisions on which treatments offer value for money reflect the value patients put on quality of life and survival benefits. Patients understand the problems of a limited budget and are well placed to judge the relative merits of different treatments. The challenge is: how to achieve input from patients and carers that is useful in the HTA process.

0395 International comparison of skillmix between GPs and nurse in primary care Mousques J. (IRDES (Institut de Recherche et Documentation en Economie de la Santé), Paris, France) Bourgueil Y. The aims of this study is to examine ways in which some countries define the roles and skill of nurse in primary care and what kind of changes have taken place over recent decades in terms of cooperation, delegation and substitution between GPs and nurses. How these changes have been implemented, and what impact they have had on the health care system. This international comparison was based on literature review and structured questionnaire and/or interviews of researchers, stakeholders and health care professionals. This study show that close collaboration between GPs and nurses is not a standard. Where this model does exist, whether in an experimental form (Italy, Ontario, Quebec, the Netherlands, Germany), or more generally (UK, Sweden, Finland), this is as a result of concerted policy on the part of the state to develop the professions’ traditional modes of working. Organisational model move away from the -integrated health centre- option towards collaborative practice in a self-employed group practice setting. Nurses’ participation in primary care is also evolving through the development of new tasks in this setting of collective organisations, rather than through the direct transfer of tasks previously carried out by GP. Eur J Health Econom Suppl 1 · 2006

| S71

0124 All-Payor Hospital Rate Setting in Maryland Murray R. (Health Services Cost Review Commission, Baltimore, Maryland, USA) Rationale: US health reform has been characterized by vacillations from heavy-handed regulation in the 1970s, to an emphasis on marketbased strategies in the 1990s. Extreme approaches have failed to rationalize the market for health ser vices. Performance in achieving other policy goals (access, equity and public disclosure/transparency) has also been disappointing. Objectives: To present a description of the Maryland Hospital Rate System and illustrate how a regulatory structure can be crafted to combine both regulatory and competitive forces, to correct for market failure and better achieve key health policy objectives. Results: The Maryland system, it has outperformed the rest of the US in the areas of cost control, access, equity, accountability and financial stability: 1) the lowest rate of growth in hospital costs of any state over the period 1976 – 2003; 2) financing of $550 million of uncompensated care annually, obviating the need for public hospitals; 3) equitable payment with no cost shifting; 4) higher degrees of public accountability/ transparency than in other states or countries; and 5) standardization of billing for payers and financial stability for hospitals. Hospital Rate Setting also holds great promise in a number of important areas including the development of a comprehensive “Pay for Performance” initiative.

0529

Background: The Hungarian adaptation of DRGs have been used for reimbursement hospital care since 1st of July 1993. It was in 1986 when the viability of a case-mix classification method for healthcare provision began to be discussed in Hungary. There was a growing demand for healthcare activities and performance to be measurable and appraisable. Between 1987 and 1991 data was collected from an full-scale(?) range of sources. More than 18 hospitals participated in the micro) cost study. Hungary developed information system, groupers, define the reimbursement parameters, rules. Objectives: The launch of case-mix development was clearly motivated by its applications in the area of reimbursement. Health policy-makers sought a rational financing method replacing the traditionally used input financing system which would: 1.ensure the cost-effectiveness of treatment, 2.bring the allocation of resources into line with performance, 3.provide incentives to improve care, 4.ensure the sustainability of available resources. Present system Hungary apply the DRGs system for reimbursement of all acute hospital cases as a prospective payment system: a. The relative weights are nationally uniform. b. The trim-points of duration for each group are determined taking the guarantee criteria into consideration. c. The base-rates are nationally uniform, and predetermined. The predictability of the system has increased substantially. d. The basic fee does not currently cover the amortization cost of the equipments only the operating costs. However, it also – put a cover on the payment of physicians. Result: The DRGs system in Hungary lived up to expectations since it is capable of precisely expressing the competencies of hospital care, it is suitable for determining output and can be linked up incentive rules in the interests of optimising the use of resources or achieving professional goals.

Treatment costs of different phases of breast cancer in Hungary

0548

Muszbek N. (MEDTAP Institute at UBC, AstraZeneca, London, United Kingdom) Benedict Á., Horváth K.

Development of a risk adjustment and risk sharing scheme for Hungarian Managed Care Organizations

Objectives: The objective of the study is to determine the breast cancer treatment costs of different phases of breast cancer in Hungary. Methods: Resource use and costs were collected retrospectively on three cohorts: early breast cancer (EBC) patients, patients with local regional recurrence (LRR) and metastatic breast cancer. Data were obtained from the Oncology Report, the inpatient, outpatient and the pharmaceutical database of the National Health Insurance Fund. Cohorts were determined using ICD, WHO, ATC, TNM and other codes. Age above 55 served as a proxy for postmenopausal status. Patients receiving hormonal therapy were assumed to be hormonal receptor positive. The three cohorts consisted of the first 100 relevant patients in each category entering the Oncology Report database in 2003 and were followed up for minimum 1 year. Relevant resources were predetermined with the help of treatment guidelines, oncologists and coding experts. Results: The final cohort includes 86 patients with incident EBC, 80 LRR, and 100 metastatic breast cancer patients. The frequency and type of diagnostic and surgical procedures, consultant visits, prescriptions will be summarized in natural units. Total cost/patient will be presented based on Hungarian national fee schedules. Multivariate analysis will be carried out using demographic and hospital characteristics.

0489 DRGs in focus of reimbursement of acute hospital care – Hungarian experiences Nagy J. (National Health Insurance Fund Administration, Budapest, Hungary) Dózsa Cs., Boncz I., Borcsek B. DRGs in focus of reimbursement of acute hospital care – Hungarian experiences.

S72 |

Eur J Health Econom Suppl 1 · 2006

Nagy B. (Unit of Health Economics and Health Technology Assessment Department of Public Policy and Management Corvinus University of Budapest, Hungary) Boncz I., Falusi Zs., Gulacsi L., Reszegi Cs. Introduction: In the Hungarian Managed Care Program 16 health plans cover a wide range of health care benefits for 2,3 million people. The annual budget is allocated by the National Health Insurance Fund through capitation. The capitation formula is not accurate enough to fulfil the requirements of equitable and efficient resource allocation. Meanwhile the financial pressure on health plans not to exceed the capitation can be substantial. Therefore, the improvement of the capitation formula with risk adjustment and also the use of risk sharing arrangements are required. Method: The current system to handle risk is a mixture of rudimentary risk adjustment and a high level of risk sharing. The capitation formula is weighted by three variables age, sex (R2=4,57%), and chronic dialysis condition (R2=18,65%). Risk sharing incorporates outlier risk sharing where surpluses and deficits are handled in a solidarity fund. Also carve outs with high cost, high risk services are established to prevent extreme variations in the health plan’s budgets. These elements are under investigation. Results: Preliminary results show that the improvement of risk adjustment scheme is an optimal choice to improve both equitable and efficient resource allocation. For this purpose, individual level diagnoses, pharmaceutical data and aggregate level socioeconomic data are scrutinized. These provide a good hybrid system of health based and non-health based risk adjusters. We demonstrate that risk sharing is supplementary along with risk adjustment, since this method successfully eases the financial pressure on health plans but also reduces incentives to control expenditures. Conclusion: Priority on risk adjustment is recognized and that the improvement of risk adjusted capitation has to be balanced in line with the reduction of risk sharing arrangements.

0568 Vertical equity – measurement and impact on health related decision making of the Hungarian GPs. Nagy L. (MSD Hungary Ltd., Budapest, Hungary) Kovács F., Erdész D. One of the factors that may explain the limited impact of economic evaluations is that decision-makers believe that non-preference based outcome measures (i.e. QALYs) fails to address many potentially relevant dimensions of healthcare decisions. Externalities, like option value which are typically non-health effects of healthcare play an important role in determining people’s distributional judgments. Evidence shows a rejection of health maximisation as the sole criterion for allocating resources several other factors exist like: patient age, socioeconomic status, prevalence of the disease, severity of disease, type of illness (acute vs. chronic), time horizon of health gain, distribution of future health gain, number of people treated. Existing studies provide indications that people may make trade-offs, but do not provide useable equity weights. The present paper explores the potential to incorporate the equity dimension into healthcare decision-making. As a pilot study in a face to face interview a Choice Based Conjoint method was used in a randomly selected stratified GP population (200 physicians) to determine equity weights of the attributes described above. As a result we suggest to scale down the normative ambitions of economic evaluations, at least until these can be based on solid evidence about people’s preferences and values.

0581 Measuring the outputs and productivity of long term care Netten A. (PSSRU, University of Kent, England) In response to international requirements to improve the basis of National Accounts an approach is being developed in the UK to reflecting the welfare gain from care services. The aim is to use routine sources as far as possible in an index that is designed to reflect changes in characteristics of services, of service users and quality of care. The presentation will demonstrate the application of the theoretically grounded approach to the measurement of outputs of long term care services for older people. Services are described in terms of the capacity of service users to benefit from them and a quality measure applied to reflect the degree to which services do actually deliver this benefit. National surveys of user satisfaction and regulator reports of care home standards provide the basis of the quality measures. Data from research studies are used to adjust data from routine sources to reflect the changing contribution of welfare resulting from changes in patterns of provision and increased targeting of services. The approach has potential for a number of policy and practice applications including facilitating international comparisons of the productivity and efficiency of long term care of older people.

0106 Measurement of informal care giving time in a study with dementia patients. Neubauer S. (GSF – Forschungszentrum für Umwelt und Gesundheit, Institut für Gesundheitsökonomie und Management im Gesundheitswesen (IGM), Neuherberg, Germany) Gräßel E., Holle R., Großfeld-Schmitz M. Background and aim: Due to the rising life expectancy in developed countries the prevalence of dementia is increasing. One aim of dementia care is to postpone or avoid institutional care. Performing an economic evaluation of interventions with this aim from a societal perspective has to include informal care. We report on our concept and practical experiences in measuring these costs.

Methods: Within the Ger man IDA study (‘Initiative Demenzversorgung in der Allgemeinmedizin’) three intervention groups are to be compared with respect to time to institutional care over a period of two years. Informal care is measured by a German adaptation of the RUD Lite questionnaire, which distinguishes three categories of informal care activities: ADL, IADL, and supervising. The informal caregiver has to state the number of days and average hours per day of care provided to the patient during the last four weeks. In addition to the original version, we try to include the time of all informal caregivers who are involved in caring for the patient. The questionnaire is applied as a computer-assisted telephone interview with the primary caregiver. Results: Our first experiences show that most caregivers are willing and able to answer the questionnaire. We will report on our findings from the baseline assessment of caregiver time and describe factors influencing the completeness and plausibility of the data.

0237 Should one brush twice a day? Determinants of dental health among young adults Nguyen L. (Centre for Health Economics – CHESS, National Research and Development Centre for Welfare and Health (STAKES), Helsinki, Finland) Häkkinen U., Knuuttila M. The aim of this paper is to explore the determinants of dental ill-health with a focus on the occurrence of caries. Derived from health production and demand theory, we employed a recursive two-equation probit model that consists of a production function of dental health and a demand function of dental care. The data are from a longitudinal study of the Northern Finland 1966 Birth Cohort that had the latest follow-up when the cohort reached the age of 31 (n=5020). The factors controlled for related to birth, youth, the living condition of the family, and socioecononomic variables. The total effects of the explanatory variables on the occurrence of caries were computed for the recursive probit model. Factors among females that increased caries were a body mass index and the intake of alcohol, sugar and soft drinks. Among males, caries was positively related to the metropolitan residence and negatively related to good diet and current education. Smoking increased caries, whereas brushing teeth at least twice a day, visiting the dentist regularly and using dental care reduced caries. The findings suggest that to prevent caries, attention should be focused on appropriate policies aimed at providing dietary advice as well as raising dental education. Brushing twice a day is a useful and obvious cost-effective means of self-care that all adults should be encouraged to adopt.

0390 Eliciting preferences of menopausal women for a Hormone Replacement Therapy using Discrete Choice Experiments Nguyen F. (GRESAC, Lyon, France) Moumjid N., Brémond A., Carrare MO. All women experience menopause after their reproductive years, and eighty percent of them suffer from symptoms that may affect their quality of life hot flushes, osteoporosis. After 70 years, fractures of the femoral neck occur in one out of six women.order to alleviate these symptoms and risks, 30 to 50% percent of women between 48 and 64 years of age use Hormone Replacement Therapy (HRT). But recent studies have shown that HRT is also associated with increased risks of heart disease, pulmonary embolism and breast cancer.HRT presents (1) risks, (2) benefits and (3) monetary costs, the aim of our study is to analyse women’s preferences for HRT, using Discrete Choice Experiments (DCE). The method consists in building hypothetical scenarios with different treatment characteristics, and presenting them to women from a representative sample who then indicate their preferences.this study, we will be Eur J Health Econom Suppl 1 · 2006

| S73

able not only to specify the influence of benefits, risks and monetary costs of HRT on women_choice to use this treatment or not, but also to determine the potentialities of DCE as a support to collective decisionmaking for allocating collective health resources.

0564 Universal Coverage with Aspects of Rationing and Performance: The Case of England and Wales Nicholson T. (Public Health Sciences & Medical Statistics, University of Southampton, UK) Roderick P. England and Wales are comparatively high cost chronic kidney disease countries. Health care provision is predominantly publicly funded through a tax based National Health Service (NHS). Renal services have expanded significantly in the last decade, particularly hospital haemodialysis (HD). Here public funds have been used in some areas to commission private providers of satellite HD, though such facilities have NHS medical input and are linked to NHS renal units. Regional variations do exist, partly driven by historical patterns and in response to population need (e.g. ethnic minority populations). Whilst acceptance rates onto renal replacement therapy (RRT) are low by international standards, they are commensurate with many other northern European countries. There is no explicit rationing of care, and no robust data on mismatch between the incidence of established renal failure and acceptance onto therapy. However, the latter is likely to be most apparent in older ages and in those with significant co-morbidity. Currently, formal palliative care provision is low but growing, offering an alternative to RRT. Health outcomes appear to be improving in England and Wales. This should continue through the recent establishment of a Renal Services National Service Framework that sets out standards and markers of good practice. Monitoring is possible through clinical governance frameworks and the UK Renal Registry (that covers approximately 90% of patients on RRT).

0585 How can the popular health insurance programme support the achievement of MDGs in Mexico? Nigenda G. (National Institute of Public Health, Cuernavaca, Mexico) González L.M., González M.C. In Mexico, a heavy burden of diseases is still posed on the poorest populations despite the government´s efforts to achieve universal health coverage. The recently launched Popular Health Insurance Programme is the first programme in the country´s history to provide insurance to the jobless and poor populations using fresh public funds. Various strategies have been implemented alongside the programme to target the most vulnerable populations and deal with their most relevant diseases. The impact on the improvement of health conditions in these populations is estimated to reduce infant and maternal mortality in the following years, two indicators that are key to achieve MDGs in health in Mexico. The way this strategy is being implemented to improve these indicators is the objective of the presentation. The paper shows the amount of human and financial resources involved in the production of services, the groups of population that have been most benefited, the articulation between levels (federal, state and municipal) of government to achieve the strategy goals and the perspective of managers and providers in the progress made so far to achieve targets.

0086 Comparing the central limit theorem and bootstrap methods for estimating incremental net benefit Nixon R. (MRC Biostatistics Unit, Cambridge, UK) Wonderling D., Grieve R. An increasingly common approach in cost-effectiveness analysis on individual patients in a single randomised controlled trial is to estimate the incremental net benefits (INB), along with its 95% confidence intervals, and compute cost-effectiveness acceptability curves and confidence ellipses. Two common alternative non-parametric methods for estimating INB are to employ the central limit theorem or a bootstrap method. Both these methods are described and the validity of their underly ing assumptions dis cussed. Neither method is exact. However, both methods give unbiased estimates of the incremental net benefit, but give dif ferent estimates for the variance of this estimate. We used Monte Carlo simulation to compare the two techniques, assuming the population effect to be normally distributed and costs gamma distributed. We repeated this experiment a number of times varying the sample size, skewness, and willingness-to-pay threshold. In almost every experiment the variance of the estimated incremental net benefits was better estimated using central limit theorem than the bootstrap, leading to better coverage of the confidence inter vals and confidence ellipses, and more accurate CEACs.

0334 Evaluating the level of generalisability of economic evaluation studies Nixon J. (Centre for Reviews and Dissemination, York, United Kingdom) Ulmann P., Déméothis C., Serrier H., Rice S., Drummond M., Pouvourville de G. Objectives: In a previously published study we reported the preliminary validation of the EURONHEED transferability checklist. The aim of the present study is to report the results of a validation survey conducted with French and UK decision-makers. Methods: Postal survey of a panel of 40 prominent health economists and health care decision makers in France, which was matched for the UK. The survey instrument asked respondents to choose items that they considered to be import in the assessment of generalisability. They then assigned a weight (1, 2 or 3) to each selected item, which reflected its relative importance. Qualitative comments were also sought concerning the checklist, and suggestions for further items that were not included in the present version. Results: The response rate was 75% in France and 65% in the UK. The survey results provided a good level of validation of the checklist, which can now potentially be modified to produce a weighted solution. Discussion: Differences among the responders, in conjunction with their qualitative comments, suggest that more analysis and research is needed regarding decision-makers’ knowledge and understanding of economic evaluations. The EURONHEED checklist is close to becoming a viable instrument to assess the transferability of published economic evaluations.

0193 A Comparison of GP Visiting in Northern Ireland and the Republic of Ireland Nolan A. (Economic and Social Research Institute, Dublin, Ireland) McGregor P., Nolan B., Oill C. In Northern Ireland, all residents are entitled to free GP services while in the Republic of Ireland, only the 30 per cent of the population who

S74 |

Eur J Health Econom Suppl 1 · 2006

qualify on the basis of an income means test are entitled to free GP services. Given the similarity in the institutional features of GP services in Northern Ireland and the Republic, this allows us to analyse the impact of charges on GP use. use comparable data from the 2001 Northern Ireland Household Panel Survey and the 2001 Living in Ireland Survey to estimate pooled cross-section models of GP services use controlling for a variety of individual and household demographic and socio-economic characteristics. We estimate five separate ordered probit models for each income quintile, with a dummy indicator variable for those who live in the Republic. results indicate that for those in the lowest income quintile in Northern Ireland and the Republic, there is no significant difference in visiting between Northern Ireland and the Republic (because the majority of those in the lowest income quintile in the Republic are eligible for free GP care). However, for the second, third and fourth income quintiles, visiting is significantly lower in the Republic (where the majority must pay), and for the fifth (and highest) income quintile, visiting is also significantly lower in the Republic, but not to the same extent as for the middle income quintiles. This suggests that the effect of charges in the Republic is important, although less so for those in the highest income quintile.

0152 An Economic Evaluation of the Incredible Years Child-Parenting Programme Ó’Céilleachair A. (Canolfan Economeg a Pholisi Iechyd, Bangor, Wales) Edwards RT., Bywater T., Hutchings J. Background: Conduct disorder is estimated to affect 10% of children in the UK and has both short and long-term cost implications for families, public services and society. Objectives: To investigate the relative cost-effectiveness of the Incredible Years Child-Parenting Programme compared with a six-month waiting list control. Design: An incremental cost-effectiveness analysis of a randomised control trial. Setting: Sure-Start areas in north Wales, mid Wales and Borders. Inter vention: 116 children aged 36-59 months with conduct disorder as measured by the Eyberg Child Behaviour Index (ECBI) were randomised to either a group receiving the child-parenting programme or to a six-month waiting-list control. Primary Outcome Measure: Incremental cost per unit of improvement on the intensity score of the ECBI. Results: We calculated an ICER point estimate of £23 per one point improvement on the intensity score of the ECBI (95% confidence interval –£1–£51). With a Cost-Effectiveness Acceptability curve cost ceiling of £45 there is a 93.9% chance of the intervention being cost-effective. Conclusions: The Incred i ble Years Child-Parent ing pro gramme improves child behaviour as measured by the intensity and problem scores of the ECBI at a small additional cost and consequently could be cost-effective as compared with currently provided interventions.

0280 Patient Migration in Central Europe: The Case of Dental Care Oesterle A. (Vienna University of Economics, Vienna, Austria) Delgado J. With recent ECJ decisions, patient migration became a top priority issue in European health policy. Apart from patient migration in the context of regional co-operations, dental care is one of the few health sectors in Europe where cross-border migration of patients has already reached considerable levels. The paper studies the determinants of cross-border dental care in Central Europe by focusing on respective migration flows between Austria and Hungary. The underlying hypothesis is that,

in this particular sector, individual expenses to be borne by patients work as a decisive factor for migration decisions. Starting from previous work on patient migration using choice models, the study identifies four major determinants: quality of care, service quality, patients’ expenses for the service provided and transaction costs. Investigating these determinants the study uses administrative data (which only covers those reimbursed by social health insurance funds) as well as data from interviews with patients. First results confirm that dental care migration in Central Europe is strongly patient driven, with prices and service quality being key determinants for the respective decisions.

0297 How good is the quality of health care in France? Or Z. (IRDES,of Research on Health Economics, Paris, France) France has been classified by the WHO as having one of the best health systems in the world. However, there is surprisingly little systematic information on the quality and safety of health care provided in France. Despite many recent initiatives to improve quality and its measurement, the available data remain partial, inconsistent and not easily accessible. In order to obtain a global picture of health care quality problems and to develop quality-improvement strategies, it is important to collect national data on a consistent basis within a coherent framework. However, unlike in most English-speaking countries, no attempt has been made in France to develop a national reporting system on quality. This paper pools together both published and unpublished data from many different sources on the quality of care in France using the internationallyrecognised framework developed by the U.S. Institute of Medicine. The dimensions covered are safety, effectiveness, appropriateness, responsiveness and accessibility. Where possible, comparisons are made with the situation in other OECD countries. This paper thus provides an overview of what is known about the quality of health care delivered in France within an international perspective, and aims to identify the major deficiencies in terms of measuring and reporting quality.

0012 Person trade-off in QALY and DALY models Østerdal L.P. (University of Copenhagen, Denmark) A considerable number of papers have proposed, or have been in.uenced by, a speci.c method for integrating social preferences in the QALY framework (e.g. Nord 1992, 1994, 1995, 1996, Nord et al. 1993, Richardson and Nord 1997, Dolan and Green 1998, Ubel et al. 2000, and Schwarzinger et al 2004). In this method the quality-adjustment factor is estimated from person trade-o¤ scores, i.e. assessments of the number of people that should receive a speci.ed health improvement (for example one extra life year at a certain health state) so that it is socially equally preferred to axed number of other people receiving a given health improvement (for example one extra life year in perfect health). Likewise, the DALY model used by WHO for assessment of region- and disease-speci.c burden of disease is also based on disability weights estimated by person trade-o¤ protocols (WHO 2004) (for details, see Murray and Lopez 1996). Despite the popularity of quality-adjustment indices based on person trade- o¤s, theoretical support for its use in a QALY framework has not been investigated (Dolan 1998, Green 2001, Mansley and Elbasha 2003). In this paper, we show that if the quality-adjustment factor is constructed from person trade-o¤ scores, then any evaluation criterion based on QALYs (or DALYs) generally violates the Pareto principle.

Eur J Health Econom Suppl 1 · 2006

| S75

0324 Explaining the Differences in Kidney Organ Donations Across Countries Osterkamp R. (Ifo Institute for Economic Research, Munich, Germany) Roehn O. The scarcity of kidney organs with resulting waiting lists for kidney transplants is a key health sector problem in nearly all industrialised countries. The main dimensions of the problem are individual pain and high social costs, both due to dialysis. However, donation figures per million population and p.a. differ widely across countries. The article asks for the reasons of the differences and whether a best practice can be identified. Within a multi-variate panel analysis for 25 industrial countries several possible determinants are explored. Factors considered are the legal framework for organ donation, centrality of the clinical system, density of intensive care units (ICU), availability of “incentive” personnel in clinics with ICU, fatal traffic accidents, prevalence of obesity, practising dialysis specialists, donation-related spending for professional and public education, the interaction between cadaveric and living donation. The elements of a best practice system can be identified. Spain seems to be nearest to best practice.

0088 Supplementary health insurance as a tool for risk-selection in mandatory basic health insurance markets Paolucci F. (iBMG – Erasmus University, Rotterdam, Netherlands) Schut E., Beck K., Voorde C., Greß S., Zmora I. Abstractthis paper, we examine the role of supplementary health insurance (SHI) as a potential tool for risk-selection in five countries with a competitive basic health insurance (BHI) market Belgium, Germany, Israel, the Netherlands and Switzerland. We assess the potential role of SHI as a selection device in each of these countries by investigating the potential gains from risk-selection, the market share of SHI and the links between basic and supplementary health insurance markets. From this analysis we conclude that SHI is a powerful selection tool in Switzerland, while in the Netherlands incentives to use SHI for risk-selection are increasing. In Germany and Israel, incentives for risk-selection are strong but opportunities for using SHI to attract favourable risks are severely restricted by regulation and the prevailing market structure. In Belgium, there are ample opportunities to use SHI as a selection device but the incentives to do so are limited though increasing.

0016 Patient trajectories and Health areas Parent O. (Université de Saint-Etienne, France) Boureille B., Normand M., Trombert-Paviot B. This study attempts to analyse the trajectories of patients within the French health care system. The system of patient trajectories has variable features, specific determinants, and consequences on people_ health that should be investigated in a regional perspective. We present a trajectory-based analytical approach to assessing patterns, determinants, and outcomes of health care with application to breast cancer.. The approach uses economic and statistical methods for sample selection and analysis, as well as statistical methods concerning model fitting and validation strategy. It also uses spatial econometrics methods allowing patient trajectory identification.. Using a new dataset of the French National Case Mix Database over the period 2003 – 2004 for the Rhône-Alpes region, we find different patient trajectories for breast cancers pathology. Differences between the Loire area (with 2,129 patients and 70 hospitals and medical centers) and a metropolitan city Lyon

S76 |

Eur J Health Econom Suppl 1 · 2006

(with 1,210 patients and 44 medical centers) are set forth.. We illustrate the feasibility and profitability of the patient trajectory system to identify patterns, determinants, and consequences of regionalized health care system. This approach could be used to analyse other pathological health-care systems.

0571 Health economics from the consumers’ perspective; a study of consumers’ preference for coronary heart disease Parsa A.D. (Division of cardiovascular medicine, University of Nottingham, UK) Gray D. Every healthcare system in the world is struggling to meet demand for health care. Health systems throughout the world face escalating health care costs and prioritisation of health care issues is one approach to costcontainment. Willingness to pay (WTP) is increasingly being used as a measure of valuation in health technology assessment. Willingnessto-pay studies have assessed the value of a particular inter vention or to compare the relative value of different treatments. Specifically, willingness to pay (WTP) may be used as a guide to resource allocation. WTP implies that consumers of health services can express their preferences through their ability and willingness to pay for services, whether prevention or treatment of CHD. It can be used to explore the relationship between supply and demand for both existing ser vices and new services. Objectives: Different groups of ser vices consumers may react differently to price increases. The main objective of this study is to find out people’s willingness and ability to pay for treatment and prevention of CHD among British and Iranian population (people involved with CHD and without CHD involvement). The influences of age, gender, socioeconomic group, extent of knowledge of disease and its treatment, geographic location and educational level have been assessed. Methods: Design The study has been undertaken in both quantitative and qualitative setting. Setting and Sampling: Data have been collected from the study population in different geographical areas of Iran (Tehran, Abaddan and Ahavaz) and in the UK (Nottingham). This study is based on data from mainly a self administered questionnaire, and partly interviews will be conducted for target population. The study population comprises of about 500, patients with diagnosis of CHD or previous diagnosis of CHD who had attended cardiology clinic or admitted in the University hospital in Nottingham and Iran versus healthy volunteers (with no previous experiences of heart attack). Participants in volunteers group are selected randomly but will be stratified according to age, sex and so on. Additionally it is proposed to include non-English speaking (Farsi) participants (Iranian) due to undertaking part of study in Iran. Number of predictor variables, including age, gender, ethnicity, employment status, household income has been examined. The format consisted of two ways a self-administered questionnaire and a face-to-face interview. Perspective: The perspective of study is societal and the results will be relevant to decision making within the Iranian health system. Data analysis: Both quantitative and qualitative analysis using SPSS and ANOVA are used. Discussion: unfortunately we have not got all data yet but it is more likely to get the final result before the conference date. Conclusion: Till now, it is expected that there are significance differences in valuation of healthy heart and risk reduction programme in selected populations. This study stresses that coronary heart disease is a main priority not only among patients but also all other people in both developing and developed countries.

0615 Overview of priority setting, priority setting methods and the use of PBMA: results from international research Peacock S. (British Columbia Cancer Agency & University of British Columbia, Canada) Donaldson C., Bate A. The development of frameworks to allow decision-makers to manage scarce resources is one of the most important challenges facing health economics. Managing resource scarcity involves making choices about which types and amounts of health care to provide for different individuals and populations from a given budget. This paper describes an economic approach to priority setting – Programme Budgeting and Marginal Analysis (PBMA) – to assist decision-makers in making these choices. PBMA has been used in over 70 priority setting exercises in Australia, Canada, and the UK. It is based on three concepts: the need to consider opportunity costs; the need for marginal analysis of costs and benefits; and, the existence of a fixed budget which implies that some services must be contracted if others are to be expanded. Recent developments in PBMA will be described, which have focussed on the complex nature of both objectives and constraints in priority setting. Developments include modelling decision-makers’ objectives using multiattribute utility theory, incorporating ethics into economic approaches, and examining the constraints of political economy and organisational context that decision-makers face in setting priorities.

0601 Transferability of utility results of rheumatoid arthritis cost-effectiveness models into Hungary Péntek M. (Rheumatology, Flór Ferenc County Hospital, Kistarcsa, Hungary) Brodszky V., Májer I., Rojkovich B., Polgár A., Szekanecz Z., Poór Gy., Czirják L., Genti Gy., Kobelt G., Gulácsi L. Objectives: Several models have been developed to study the cost-effectiveness of new therapies in rheumatoid arthritis but the degree of utility changes and the methods used were different. Adaptation of international results is crucial in Hungary once there is a lack of epidemiological data and cost-effectiveness analysis is required for reimbursement. The aim of our study was to assess the transferability and adaptability of utility results of rheumatoid arthritis cost-effectiveness models into Hungary. Materials and methods: A cross-sectional sur vey was performed by hospital based rheumatologists from June to August in 2004 involving 255 rheumatoid arthritis patients. Disease activity (DAS), self-completed questionnaires (HAQ, EQ-5D, RAQoL) were used and visual analogue scale (VAS) to measure health status, pain, patient’s and physicians’ global assessment. Statistical analysis was performed to analyze correlations. Utility on different disease stages was studied. Results: Correlation between EQ-5D, HAQ, RAQoL, DAS and painVAS was significant. Utility equations were: EQ-5D=0,85-0,282×HAQ (R2 0,413); EQ-5D=0,895–0,027×RAQoL (R2 0,427); EQ-5D=0,8410,0081×painVAS (R2 0,312). Disease activity correlation was: EQ5D=1,014–0,25×HAQ–0,041×DAS (R2 0,44). Using 6 HAQ states, mean EQ-5D was 0,784; 0,576; 0,504; 0,367; 0,211; 0,022 (p<0,01), mean RAQoL scores were: 6,7; 12,23; 16,05; 20,96; 23,29; 22,37. Conclusion: Utility equations can be set up in the Hungarian sample but the difference can achieve the clinically minimal important level in comparison with models. Coincidently with Kobelt’s previous study, disease activity has impact on utility loss independently from HAQ, pain seems to have key role. Using HAQ based Markov model stages the EQ-5D change was significant but not consistent, the greatest loss was detected in the mildest groups. Disease specific quality of life was constant once a severe functional disability was achieved. Our cross-sectional survey covering several disease parameters offer comparable results

to models involving observational cohorts. Country-specific data might have impact on QALY gain specially in Markov models.

0430 Income-related inequality in health care in relation to inequalities in need Pereira J. (Escola Nacional de Saúde Pública, Nova de Lisboa, Portugal) Lopes S. Objectives: To assess income-related inequities in the delivery of health care in Portugal. Methods: Study draws on a method proposed by Wagstaff and van Doorslaer. Data are from the 1998/99 Portuguese National Health Interview Survey. Specialist and GP visits, as well as lab tests, are used as measures of health care utilization. Several indicators of ill-health are used as predictors of need. Results: Findings provide evidence of unequal treatment for equal need. Inequity indices reveal that the worse-off use more GP visits and the better-off use more specialist care and lab tests, even after taking need into account. The most significant inequities in health care are in specialist visits. The distribution of GP visits slightly favours the poorer income groups. Although these findings apply to all morbidity measures used in the prediction of need, when a measure of self-assessed health status is used, health care provision appears considerably more favourable to the rich. Conclusions: The nature of health care provision in Portugal very likely the main reason for differential utilization of services between worseoff and better-off. A major finding is that levels of measured inequity are particularly sensitive to the need indicator that is used, with selfassessed health showing very pro-rich inequity. This question has implications for cross-country comparisons.

0218 Misleading cost estimates when neglecting the centre effect in multicentre RCTs Petrincoa M. (Department of Public Health and Microbiology and Department of Statistics and Applied Maths “Diego de Castro”, University of Turin, Italy) Paganob E., Ferrandob A., Bigic R., Desiderid A., Gregorie D. In multicentre RCTs the cost evaluation is affected by the heterogeneity in the subjects’ clinical factors (between-patient sampling variability) and in the cost structure of each centre (between-location variability) (Manca et al.,2003). To avoid misleading conclusions the centre effect needs to be appropriately considered. We analysed cost data from the COSTAMI trial: an international RCT evaluating alternative discharge strategies in the post infarction. 10 centres (8 Italians and 2 Turkishes) participated (main differences in total and cardiologic hospitalisations and in intensive-care beds). Treatments costs were modelled as function of patients related covariates (age, sex, previous ischemic history, comorbidities) with a Cox proportional hazards model. The centre effect was modelled as shared frailty from Klein (1992) and Nielsen (1992). Frailty estimate was obtained by applying the Therneau penalty function (2000). The early discharge strategy showed a cost-ratio of 0.598 (95%CI: 0.48– 0.74) when modelled with the adjusted marginal Cox model. The estimated cost-ratio for the centre effect, based on the gamma frailty, turned out to be higher (0.783; 95%CI: 0.65–0.95) than the effect of the discharge strategy. In multicentre trials, to avoid misleading conclusions in comparing services costs and relating the results to clinical characteristics, the centre effect should to be evaluated.

Eur J Health Econom Suppl 1 · 2006

| S77

0290 Investments in medical technology and the role of doctors Pignataro G. (University of Catania, Italy) A theoretical model is developed to study the determinants of hospitals’ choices on how much to invest in medical technologies. A twostage process is modelled: in the first stage, the manager of a hospital has to choose how much to invest in medical equipment. In a second stage, doctors use the equipment, to maximise their objective function, defined according to a work-preference assumption. Medical technology can be used by doctors to serve two different purposes: 1) increasing the number of treatments; and 2) allowing more complex cases to be treated, or more “sophisticated” treatments to be provided. They will choose how much effort to put in using this equipment for treating more patients and how much effort to devote in using the technology for treating more complex cases (or providing more “sophisticated” treatments). The manager maximises revenues (positively related to the number of patients treated in the hospital, as weighted by the complexity of treatments), net of investments opportunity costs. First, the amount of investment chosen by managers is characterized in terms of doctors’ preferences for the different ways of using technology, the hospital’s remuneration system, doctors’ effort productivity. Second, the results are extended in an asymmetric information context, where hospitals’ managers are uncertain about doctors’ preferences.

0580 The influence of the ratio bias on the value of a statistical life Prades J.L.P. (Universitat Pompeu Fabra, Barcelona, Spain) Pérez J.E.M., Perpiñan J.M.A. Objective: to test the consistency of values of a statistical life (VOSL) using the Standard Gamble-Contingent Valuation (SG-CV) method. We analyze if these values can be influenced by the Ratio Bias. Background: the SG-CV method has been proposed as a method more consistent than the traditional one in order to estimate VOSL. We refer to the “traditional” method as that one that estimates the VOSL asking willingness to pay (WTP) questions for very small (usually X in 100.000) risk reductions. The SG-CV method is a two stage procedure. First, subjects are asked WTP and willingness to accept (WTA) questions for a health state under certainty. Using the response to the WTP and WTA questions, the Marginal Relation of Substitution (MRS) between risk and money is estimated. Second, the SG method is used to estimate the relative value (RV) of a health state in relation to death. Using these two questions, the VOSL is estimated as MRS x RV. The ratio bias (RB) consists in the tendency of subjects to concentrate on the absolute number of events (the negative part of the numerator) when they do judgments about risk, not giving enough weight to the denominator (the base of the risk). Methods and results: 180 subjects where inter viewed using the CVSG method in order to estimate the VOSL. In this group, we asked the SG question using 1000 as the denominator. We then asked the risk of death that subjects would accept in order to get rid of health states X and Y as the number of cases out of 1000. We also asked for WTP and WTA for health states X and Y. Using this method, we estimated a VOSL of about 9 million euros. In a second group (n=102), we replicated the same methodology using 500 as the denominator. We then asked the risk of death that subjects would accept in order to avoid health states X and Y, as the number of cases out of 500. We found that subjects were not very sensitive to the change in the base of risk (from 1000 to 500). This reduced the relative value (RV) of the health states in relation to death and then the VOSL. We found a VOSL of about 4.5 million euros if the SG questions were asked as number of events out of 500 (instead of 1000 as before).

S78 |

Eur J Health Econom Suppl 1 · 2006

Conclusion: the SG-CV avoids some of the problems observed in the literature related to the difficulty of conveying small probabilities to subjects. However, it is not a method free of biases. We have shown that the Ratio Bias can have a dramatic influence on the VOSL estimated using this method. The CV-SG method should be improved in order to provide more consistent estimate of the VOSL.

0286 Healthcare network: implications for Italian local health authorities Piraino N. (Politecnico di Milano, Department of Management, Economics and Industrial Engineering, Milano, Italy) Masella C., Piraino N. The Italian Health Care Sector is characterized by a number of reforms that have deeply modified its managerial and organizational settings. Currently the Italian healthcare system is highly decentralized: there are many types of healthcare organizations and different coordination mechanisms (hierarchy, market and network) are implemented. The aim of the paper is to analyse different types of network implemented in the Italian healthcare sector and to assess the impact of coordination mechanisms adopted on local health authorities, from an organizational and managerial point of view. On the basis of the theory of network-based organizations, two case studies are conducted considering two main units of analysis: 1) the network and 2) the local health authorities. The scope is to assess empirically how the network affects the managerial practices of the local authorities, for example in terms of new models of governance and accountability arrangements. The healthcare networks analysed are two pilot experiences, carried out in Regione Marche (ASUR-network) and Regione Toscana (Società della Salute), and the local health authorities involved are, respectively, ASURMarche and Azienda USL 10 di Firenze. The research could provide a useful contribution for health institutions to improve their performances and to understand the effects of regional health policy, particularly in the light of the latest changes in Italian Healthcare system.

0604 Public Health Economics in The Netherlands Polder J. (National Institute for Public Health and the Environment (RIVM), Bilthoven, Netherlands,) Klazinga N. The contribution of health economics to the public health debate is steadily increasing. In this presentation the Dutch approach will be highlighted. We will start with descriptive studies on health expenditure and population health. Cost-of-illness figures will we presented and the role of the last year of life will be discussed, as well as the influence of determinants (like hypertension and obesity) and unhealthy behavior (smoking, physical inactivity, drinking too much, eating unhealthy things). In the week previous to this conference, the new edition of the Dutch Public Health Forecasts report will be launched. In this presentation the main conclusions will be summarized from an economic perspective.

0130 Ageing impact on elasticity of Health Care Expenditure Polistena B. (University of Rome Tor Vergata, Rome, Italy) Ratti M., Spandonaro F. Forewordstudies, using cross-section and panel data, compare the evolution of the health care expenditures in different countries. Common factors in these studies are the specification of the variables explaining

health care expenses, as well as the evaluation of the elasticity of health expenditures to per capita GDP.pioneering results obtained by Newhouse (1977) and those obtained by Abel Smith in their researches on a sample of OCSE countries, show an elasticity greater than 1, while Hitiris e Posnett (1992), introducing demographic variables in the analysis, evaluate an elasticity very close to 1.of the study study analyses the impact of the introduction of demographic factors into the prediction models of the incidence of the health care expenditure on GDP, in a fast ageing country like Italy.context of this analysis is somehow peculiar, because of the redistributive effects induced by the funding rules of regional Health System in Italy solidarity mechanisms tend to invert the relation between the GDP share absorbed by health care, and average per capita GDP.and Methodologyanalysis is conducted on socio economic expenditure data at level of the Italian regions, through regression models on panel data.

0169 Health Care Cost Analysis for Heart Failure Patients in Marche Region Politi C. (Azienda Sanitaria Unica Regionale, Regione Marche, Ancona, Italy) Deales A., Marcobelli A., Cicchitelli F., Stanislao F. Objectives: To estimate health care costs for heart failure (HF) patients in Marche region in 2004 through record linkage of administrative and demographic databases. To update and complete a published early version of the study. Methods: Patients hospitalised in 2001-2004 with HF as primary or secondary diagnosis and those with co-payment exemption due to HF are identified. Their health care consumptions in 2004 are retrieved by record linkage of administrative databases; costs are estimated by DRGs, outpatient tariffs and drug gross expenditure. Total costs, average costs per patient, consumptions patterns by age and sex are analysed. Results: 21.071 patients out of 1,5 million inhabitants in Marche region are recorded with HF (1,4% prevalence rate). In 2004 total health care costs are estimated around 41.4 million € (1,7% of the regional public health expenditure). Hospital care, drugs and outpatient care cover respectively 80%, 18% and 2% of the total costs. The average cost per patient is 1.964 €. Older patients register on average lower unit costs, in particular for drugs and outpatient services. Males record higher average costs per patient than females. Conclusions: Findings confirm the relevance of HF and highlight different patterns of access to health care according to age and sex. Work is in progress to extend the analysis to other Italian regions.

0216 Health Services Access and Utilization in Brazil Porto S. (Escola Nacional de Saúde Pública, FIOCRUZ, National School of Public Health, Rio de Janeiro, Brazil) Brazilian health reform stablished the implementation of social and economic policies in order to guarantee universal and igualitarian acess to health services.results of the household survey that has been accomplished in 2003 showed that there still are important inequalities in the access and utilization of health ser vices in Brazil, but that there are improvements in several indicators between 1998 and 2003, such as Percentage of the population which uses regularly one health service; Percentage of persons which had at least one medical consultation in the previous year; Percentage of persons that demanded some kind of health service in the previous 15 days.survey also showed a decline at the percentage of persons that evaluated their health status as bad or very bad.analysis of these data showed that the modification of the health ser vices delivery model and of its financing mechanisms contributed to the improvement of access conditions. 1998 primary health care program has been prioritized and it has been stablished that its

financing should reach a situation of expenditures per capita equality. There have also been created incentives for the development of the “Family Health Program” (Programa de Saúde da Família) and community health agents.paper discuss the contribution of primary health services to the improvement of health services access and utilization indicators.

0366 Early life effects on Individual Mortality Portrait F. (Vrije Universiteit, Amsterdam, Netherlands) Portrait M. Lindeboom G.B. On average being born in a recession increases the mortality rate later in life. We examined the role of nutrition and of exposure to diseases during childhood in explaining this result. We have individual data records from Dutch registers of birth, marriage, and death certificates, covering an observation window of unprecedented size (1812–2000). These are merged with historical data on regional agricultural prices and mortality indicators, and with national macro-economic outcomes. We correct for secular changes over time and other mortality determinants. We estimate duration models where the individual’s mortality rate depends on current conditions, conditions early in life, individual characteristics including land owning status and interactions terms. The results indicate that exposure to disease during infancy increases mortality at all ages and that farmers benefit from higher agricultural prices.

0490 Cost of quality and non-quality in care provided by healthcare organisations – Current situation and proposals Préaubert-Hayes N. (HAS/ Haute autorité de Santé (French national authority for health), Saint-Denis la Plaine Cedex, France) Saillour-Glénisson F., Préaubert-Hayes N. Purpose: To make professionals in healthcare organisations (HCOs) aware of the economic dimension of quality, and to provide information on the cost of quality and non-quality. Methods: Systematic search on databases, useful websites and the grey literature (1990–2003). A critical literature review report was produced and discussed with a working group at 4 meetings. Results: 1-Frequency of deficiencies of quality of care. The data available for French HCOs pinpoint the most common areas of non-quality that represent potential savings through improvements in quality of care: adverse events, unjustified interventions and hospital stays. 2-Cost of deficiencies. There are very few data for French hospitals. Foreign studies have shown that the deficiency-related excess costs vary according to type of deficiency. The most expensive are: postoperative infections bacteraemia, pneumonia, drug-related adverse events, hospitalisations and inappropriate prescribing. 3- Positive impact of interventions to prevent deficiencies. Studies published in France and abroad agree that the following may be useful: quality improvement programmes, prevention of infections, and cost rationalisation programmes. Looking ahead: To remedy the lack of information and to ensure the permanence of quality functions in HCO’s, studies must be carried out to measure the cost of the main deficiencies, the cost of investing in quality.

0588 Health Systems and Financing Challenges in New Accession Countries Preker A. (World Bank) A group of ECA countries have recently been admitted to the EU and others are lined up to join. Although the EU charter has few specific proEur J Health Econom Suppl 1 · 2006

| S79

visions on health broader European legislation have a significant impact on the health sector of the new accession countries. This paper reviews the course of health care reforms in the accession countries and challenges faced by the new members in modernizing their health systems and health care financing.

0284 Surgical volume-outcome effects Raikou M. (London School of Economics, London, UK) McGuire A. There is a widely held view that as surgical volume increases there is an improvement in outcome. There are two possible explanations for this effect; a learning by doing effect and a selective referral effect, whereby hospitals with better outcomes receive higher referrals. The empirical literature on the subject is mixed however. In this paper a competing risks model, with shared frailty to control for unobservable patient heterogeneity, is developed using longitudinal data over the period 1989 to 2002 for primary hip replacement (approx. 50,000 incident cases) and primary knee replacement (approx. 20,000 incident cases). The multiple outcomes are defined as within hospital mortality, 30-day mortality, 1-year mortality and re-admission for revision. Fixed effects are used to control for selective referral effects, allowing analysis to concentrate on the learning by doing hypothesis. By linking the patient level data to another data set containing hospital resource information it is possible to control for varying surgical input levels. A small but significant learning by doing effect is calculated for both procedures implying that the volume-outcome relationship is dominated by the selective referral effect. A squared volume effect is found to be significant indicating an upper bound to the learning by doing effect.

0115 Analysis of the ageing impact of pharmaceutical expenditure in Italy Ratti M. (University of Rome Tor Vergata, Roma, Italy) Polistena B., Spandonaro F. ForewordItaly, according to ISTAT the elderly people will constitute the 34% of the total inhabitants in 2050. patient’s age is very important in the allocation of per capita pharmaceutical public expenditure from 38,7 euro for child until four year progressively to climb up to 585,9 euro for the patients with age above than the 75 years. In total the population with more than 65 year include 60% of the expenditure and prescription in Italy.of the studyaim of the work is to estimate the ageing impact on the Italian pharmaceutical public and private expenditure. multivaried forecast model has been construct, on the regional panel data. The led analysis to put in evidence the critical aspect to the ageing, besides to evaluate the effect of the pharmaceutical and industrial policies. In particular, we evaluate the effects of the recent regulation referred to the pharmaceutical sector like different system of price reduction, introduction copayment and limitation to the prescriptions, whose effect are to be discuss.

0345 The long run productivity development in Finnish primary care and its correlates Räty T. (Government Institute for Economic Research (VATT), Helsinki, Finland) Luoma K., Aaltonen J., Järviö M.L. We measure productivity changes of primary care in Finland as a ratio of key ser vices produced and real operating costs. It is necessary to

S80 |

Eur J Health Econom Suppl 1 · 2006

improve this measure in order to slow down the expected future growth in total health care expenditure. Our data cover all the health centres that produced both inpatient and outpatient care in 1997–2003. We use the newly developed techniques to correct asymptotic bias in non-parametric efficiency scores and bootstrap the confidence intervals for the explanatory model parameter estimates. The bias estimate of efficiency scores is on average 2.8%, the amount with which the regular scores overestimate efficiency. The average productivity declined 13.7% over the whole period; the result is insensitive to estimated bias. Even if the standard parametric confidence intervals do not generally apply when efficiency scores are regressed, our bootstrapped inter vals are almost equal to parametric ones. Of the correlates used the increased municipal taxes accounted for 3% of the productivity increase. The correlates, that are expected to decrease the need of primary care ser vices, have had negative impact on productivity, implying that health centres have not been able to adjust their expenditures correspondingly. The organisational changes that have taken place within primary care have not resulted in the desired productivity improvements.

0376 Full economic evaluations of diagnostic strategies: a survey of publications identified using the NHS Economic Evaluation Database (NHS EED) Redekop K. (Erasmus MC, Institute for Medical Technology Assessment, Rotterdam, Netherlands) Objective: To conduct a literature sur vey of published full economic evaluations (EEs) of diagnostic strategies and consider implications for the future. Methods: Publications from 1995–2004 were identified using the NHS EED database (York, UK). Various study characteristics were examined and changes over time were tested. Results: 612 EEs were identified, involving various diseases (most frequent: neoplasms (27%), digestive diseases (16%), bacterial infections/ mycoses (13%)). A slight increase in cost-utility analyses was seen (6% (1995–99), 13% (2000–2004), p=0.010). Amongst 391 studies examining only diagnostic interventions, benefit was based primarily on diagnostic/clinical outcomes (43%), and seldom on QALYs (9%). Modelling was frequent (usually decision models), and use increased over time (34% (1995-99), 44% (2000– 2004), p=0.006). Discussion: Most economic evaluations of diagnostic strategies still base benefit on diagnostic or clinical outcomes. While incorporation of health outcomes and modelling is favourable, complexity of the study increases, necessitating greater vigilance about the quality of the study. The NHS EED and sister projects (EURONHEED, CODECS) help not only to identify relevant publications but also to appraise their quality. Autonomous developments (STARD statement for reporting diagnostic studies, systematic reviews) should also improve quality of publications and provide better estimates of diagnostic accuracy.

0245 The CareQol Brouwer W. (Department of Health Policy and Management University Rotterdam, Netherlands) Redekop K., Brouwer WBF., Exel NJA., Gorp B. The societal perspective in economic evaluations dictates that costs and effects of informal care are included in the analyses. However, this incorporation depends on practically applicable, reliable and valid methods to register the impact of informal care. Wethe conceptualisation and a first test of the CareQol instrument, aimed at measuring carerelated quality of life in informal caregivers. The instrument combines the information density of a burden instrument (encompassing seven important burden dimensions) with acomponent (a VAS scale for hap-

piness). instrument was tested in a Dutch sample of heterogeneous caregivers (n=175) approached through regional caregiver support centres. This first test describes the feasibility as well asand clinical validity of the CareQol instrument. The seven burden dimensions related well with differences in VAS scores. In all instances, the average CareQol-VAS scores decreased as the severity of problems increased. Multivariate analyses showed that the seven burden dimensions explained 37% to 43% of the variation in CareQol-VAS scores, depending on the model used. The CareQol seems a promising new instrument to register the impact of informal caregivers in economic evaluations.

impact on costs, versus 40% of patients with lung cancer with a reduction in mean cost of more than 500 €. Discussion – Conclusion: The use of PET scan for radiotherapy decision-making seems more effective for lung cancer than for Hodgkin’s disease, both in terms of costs and outcomes. This result might help policy makers for prioritisation.

0507

Atun R.A. (Imperial College London, Centre for Health Management, London, UK) Dantas A., Gurol-Urganci I.

An assessment of the ability of contingent valuation (CV) to elicit patients’ preferences for the organisation of home care: the case of blood transfusion (BT).

0591 Impact of NICE on uptake and diffusion of new drugs: empirical study

Objective: The purpose of this pilot study was to test the ability of CV to elicit cancer patients’ preferences for BT, either at home or in the hospital. Methods: All patients receiving BT were enrolled in a prospective study, in 2002. In face to face interviews, there were asked if a future BT was necessary, (1) whether there would prefer to receive it at home or in the hospital, (2) what they would be willing to pay (WTP) and (3) willing to accept (WTA), for their preferred transfusion place. A bidding algorithm with two different first bids was used. Various patients’ data was collected: previous experience of cancer care, personal characteristics, quality of life... Results: Forty-four patients were enrolled and forty accepted to participate. Of 40 patients, 32 (80%) expressed a preference for home BT. On the sample of 32 patients, no protest response was expressed for WTP, whereas most WTA responses were protest responses. WTP was related to some socio-demographic characteristics (household income, distance from hospital, stage of disease...). We also observed a first bid bias. Conclusion: The use of WTP seems appropriate for eliciting patients’ preferences in this context, contrary to WTA. These results should be confirmed on a larger sample of patients who are being enrolled.

Background: Health technology assessment is increasingly used in health systems to inform reimbursement decisions and guide healthcare professionals about clinical and cost effectiveness of new drugs and technologies. However, few empirical studies exist to demonstrate the effect of organizations such as NICE on the uptake and diffusion of innovations. Available studies have focused on the extent of and problems regarding the implementation of NICE guidance, using a variety of techniques such as descriptive analysis, before-and-after studies, interrupted time series and surveys. There is a need to explore whether NICE guidance has succeeded or had a deleterious effect in improving patient access to cost-effective and innovative drugs. Methods: The impact of NICE guidance on the number of prescriptions for drugs in seven drug classes (ACE Inhibitors, Alzheimer’s, Anti-aggregants, Beta-blockers, Cox II Inhibitors, Diabetes, Obesity) were analysed using panel data regression methods, controlling for cost per unit and various demand side characteristics, such as disease prevalence and number of hospital consultations. Quarterly data covers England, for the period April 2000 to March 2005. Results: Analysis shows that prescription levels are influenced by prevalence levels, respective health utilisation measures for the disease categories and a linear time trend. The impact of NICE guidance is insignificant. Conclusions: NICE inter ventions have limited success in increasing prescribing levels for cost-effective drugs. Prescribing for cost-effective drugs in England strictly follows demand.

0608

0326

Economic assessment of innovation in radiotherapy: impact of FDG PET scan on treatment decision and target volume in lung cancer and Hodgkin’s disease

A comparison of the resource use, costs and health outcomes of endovascular coiling versus surgical clipping of subarachnoid aneurysms in the UK

Remonnay R. (GRESAC Centre Leon Berard, Lyon, France) Morelle M., Giammarile F., Carrie C., Carrere MO.

Rivero-Arias O. (Health Economics Research Centre, University of Oxford, United Kingdom) Wolstenholme J., Gray A., Molyneux A., Kerr R.

Remonnay R. (GRESAC Centre Leon Berard, Lyon, France) Morelle M., Devaux Y., Carrere MO.

Context: PET scan is an innovative imaging tool. Associated with tomodensitometry (TDM), it allows a better definition of the radiation treatment, compared with TDM only. Objective: To assess the effects of Pet scan on resource allocation (costs and savings) and on the choice of the following treatment. Method: In 2004 and 2005, 209 patients were enrolled (97 patients with Hodgkin’s disease and 112 with non small cell lung cancer) in a national, prospective, non randomised study conducted in 8 hospitals. Two treatment decisions made on the basis of TDM only or on PET scan, were compared. The direct medical cost of using PET scan was assessed by micro-costing, using data collected from specific questionnaires. The costs of new exams and the costs and savings associated with changes in the chosen treatment were calculated on the basis of reimbursement rates. Results: The mean cost of using PET scan, corresponding to an extra cost, was approximately 800 € (50% for FDG). Treatments were modified for only 10% of patients with Hodgkin’s disease with a minor

Background: The International Subarachnoid Aneurysm Trial (ISAT), a multinational clinical trial including eight European countries, has recently shown that endovascular coiling yields better health outcomes at one year than neurosurgical clipping for the treatment of ruptured intracranial aneurysms. However, data on cost-effectiveness have not been reported for any of the countries participating in the trial. Objective: To perform an economic evaluation of endovascular coiling compared to neurosurgical clipping using data collected alongside ISAT for the UK setting. Methods: 2143 patients in 11 countries with ruptured intracranial aneurysms were randomised to coiling (n=1073) or clipping (n=1070). Direct health services resource use and employment-related data were collected prospectively alongside the study. Multilevel modelling was used to analyse resource use data because of its international nature. A detailed costing analysis was performed in the sub-sample of 1644 UK patients to validate the overall results achieved by the multilevel modEur J Health Econom Suppl 1 · 2006

| S81

el. Effectiveness data will be expressed as life years and quality adjusted-life years. The economic evaluation will be presented as within trial analysis up to two years follow-up. Results: The endovascular intervention was more expensive than neurosurgery due to the costs of coils. However, these costs were offset by the longer inpatient length of stay of patients in the neurosurgery group. The mean (SD) overall total cost per patient at 24 months followup was estimated to be £16199 (£14635) and £17643 (£17711) in the endovascular and neurosurgery groups respectively, a non significant difference of –£1444 with associated 95% non parametric confidence interval from –£3191 to £389. These results will be presented with effectiveness data to provide within trial incremental analysis estimates at 24 months follow-up. Conclusion: Endovascular coiling appears to have slightly but not statistically significant lower costs than neurosurgery at 24 months follow-up. The final within trial economic analysis will provide useful information for decision makers in UK as well as other countries in Europe.

0351 Evaluation: Static or dynamic? Does it matter how we model the cost-effectiveness of chlamydia screening? Roberts T. (Health Economics Facility of Birmingham, UK) Barton P., Robinson S., Bryan S., Low N. The appropriateness of different modelling approaches to use in economic evaluations of infectious diseases such as chlamydia is becoming increasingly controversial. In a recent systematic review we found that the majority of economic evaluations on Chlamydia screening used decision analytic models, referred to as ‘static’ because they assume a constant force of infection. Very few studies had used a transmission ‘dynamic’ model, incorporating the effects on screening outcomes of ongoing transmission of infection in the population. The issue is controversial: some experts argue that static models give biased results of the impact of screening, while others believe that a simple static approach is adequate, avoiding the complexities of dynamic modelling. We compared the results of economic evaluations using identical data in both dynamic and static models. The Chlamydia Screening Studies (ClaSS) project in the UK investigated population screening and generated empirical data. We evaluated the cost-effectiveness of this screening approach using discrete event simulation: a dynamic model: In a parallel evaluation we are now conducting a static decision analysis, which uses identical data as far as possible. The full extent of the differences, if any, in the results of the two approaches will be presented and discussed in July 2006.

0348 Test-retest Reliability of Time Trade off and Person Trade Off Robinson S. (Health Economics Facility of Birmingham, UK) Bryan S., Freeman T. Economic analysis is increasingly being employed in formal resource allocation decision-making processes in health care. The consequence is that the methods being employed by economic analysts are increasingly subject to close scrutiny. One such area of methodology concerns the instruments used to elicit preferences for various health states for use in the construction of quality-adjusted life years (QALYs). There are a number of techniques, which can be used to elicit preferences with different techniques producing different results. The objective of this study is to explore the test-retest reliability of two techniques Time Trade Off (TTO) and Person Trade Off (PTO). study conducts a review of the literature on test-retest reliability of health state valuation techniques. It also reports the results of an empirical study which analysed the test-retest reliability of both TTO and PTO valuations collected by a general pop-

S82 |

Eur J Health Econom Suppl 1 · 2006

ulation postal survey. total of 798 respondents returned questionnaires. The intra class correlation coefficients ranged from 0.32-0.84 for TTO and, 0.17 : 0.82 for PTO, with the majority of coefficients being >0.50. The health states with low coefficients varied between techniques. The reliability coefficients varied between techniques and health states, with the TTO technique tending to produce higher coefficients.

0349 A qualitative investigation to explore the validity of the visual analogue scale, time trade off and person trade off techniques Robinson S. (Health Economics Facility of Birmingham, UK) Freeman T., Bryan S. There are a number of techniques that can be used to elicit individual’s health state preferences, with different techniques often yielding different results. There has been very little research around the cognitive process which respondents undertake in order to reach their valuations. this study, qualitative methods were used to provide a greater understanding of the process individuals use and the experience they face when participating in health state valuation exercises. Semi structured interviews were undertaken with respondents who had participated in an earlier health state valuation exercise which involved valuing health states using the Visual Analogue Scale (VAS), Time Trade Off (TTO) and Person Trade of Techniques. interviews explored a number of areas that are relevant to assessing the validity of the instrument and process of elicitation. The PTO exercise was generally referred to as the most difficult of the three exercises. Respondents experienced more difficulty valuing the lives of others, as in PTO, rather than valuing time for themselves i.e. life years in TTO. There was evidence that some respondents were employing heuristics (cognitive shortcuts) in order to simplify the decision making process. The implications of these and other findings were also explored.

0487 Do members of the public wish to give more weight to some QALYs than others?: Results of pilot studies. Robinson A. (University of East Anglia, Norwich, UK) Baker R., Bateman I., Donaldson C., Loomes G., Pinto J.L., Ryan M., Shackley P., Smith R., Sugden B. The National Institute for Health and Clinical Excellence (NICE) makes recommendations to the NHS on the adoption (or continuation) of therapies. This requires judgements about whether the values/weights attached to gains in quality and length of life should vary according to the characteristics of the patients receiving them, and how much the NHS should be spending (at the margin) to achieve such gains. This paper discusses an ongoing research project that sets out to inform such judgements. One aim of the project is to estimate the relative weights to be attached to a quality adjusted life year (QALY) according to the type of gain (for example, life extending versus quality of life enhancing) as well as the characteristics of the recipients of these gains. We report the results of pilot studies carried out using both ‘matching’- or person trade off- and discrete choice approaches prior to conducting valuation studies in representative population samples. Results show that members of the public can understand complex information concerning QALY gains when those gains are represented diagrammatically and are willing to differentiate between ‘types’ of QALYs. We also report the progress of the main study, involving 600 members of the public, due to commence in spring 2006.

0539 From a case mix adjusted budget allocation within a national hospital service into a US type Prospective Payment System to create market competition? Rodrigues J.M. (University of Saint-Etienne, France) If the first DRG pilot tests were processed 21 years ago (1983) in France, the real implementation began only 7 years ago (1997) with a modest DRG/GHM adjusted budgeting based on redistribution of resources between and within France 22 regions. During the 2003 fall the French government passed a new funding act in the house forecasting a PPS like payment for public hospitals starting in 2004 and expected to reach 50 % of the total resources based on reimbursement to introduce market competition between public and private for profit hospitals using the French DRG system named GHM in 2012 The paper recalls the DRG/GHM budget allocation within the French national hospital service applied from 2007 and 2003, the approved new PPS system for the coming years and the background of the 2 contradictory health care reforms of 1996 and 2004. It addresses the organizational and cultural success factors for this new stage for the French DRG PMSI project by identifying the characters and processes of the French health care fragmentation and of political system supporting the social modernization theory of enlightened despotism proposed during the XVIII century by Voltaire for the King of Prussia, the Empress (Tsarina) of Russia and the King of France before the French revolution

0030 Are fixed-term jobs bad for your health? A comparison between Spain and Germany Romeu Gordo L. (Max Planck Institute for Human Development, Berlin, Germany) Gash V., Mertens A., Gordo L.R. In this paper we analyse the health effects of fixed-term contract status for men and women in West-Germany and Spain using panel data. This paper asks whether changes in the employment relationship, as a result of the liberalisation of employment law, have altered the positive health effects associated with employment (Goldsmith et al. 1996; Jahoda 1982). Using information on switches between unemployment and employment by contract type we analyze whether transitions to different contracts have different health effects. We find that unemployed workers show positive health effects at job acquisition, and also find the positive effect to be smaller for workers who obtain a fixed-term job. We also establish surprising differences by gender and country, with women less likely to report positive health effects on job acquisition in Spain.

0005 Efficiency versus quality in the Portuguese NHS Giraldes M.R. (General Directorate for Health Ministry of Health, Lisbon, Portugal) The objective of this paper is to present the methodology used in the evaluation of the Portuguese NHS in an efficiency and quality perspective. A composite Efficiency and Quality Indicator has been used, which included efficiency indicators related with the main hospital activities (inpatient care, outpatient care, day-hospital and emergency care) and with the auxiliary sections of clinic support and the hotel support services, weighted according to the relevance of its expenditure in total expenditure, and quality indicators. In a quality perspective two process indicators, the percentage of surgeries in ambulatory care and of cesarians in total deliveries and an outcome indicator, the number of episodes of inpatient care due to nosocomial infection in total days of inpatient

care, have been considered. A case-mix index has also been considered to correct all the indicators used in the composite indicator of efficiency. The composite Efficiency and Quality Indicator resulted from the mean of the Composite Indicator of Efficiency and the Composite Indicator of Quality, this last one in an inverse basis. The results should be compared only for hospitals within the same group of hospitals, with a similar number of beds and of treated patients. Alcobaça (Group I), Santo Tirso (Group II), Portimao (Group III), Santarém (Group IV), Universitary Hospital of Coimbra (Group V), and the Gama Pinto Institute (Group VI) are the most efficient and with better quality. A similar methodology has been applied to the health centres.

0432 The impact of EU anti discrimination regulation on private supplementary health insurance in Germany. Will unisex tariffs become mandatory? Rothgang H. (Centre for Social Policy Research, University of Bremen, Germany) The EU Council Directive 2004/113/EC of 13 December 2004, which has to be implemented into national law within 3 years, rules that from December 21st 2006 onwards, sex as a factor in the calculation of premiums is only permitted where the use of sex is a determining factor in the assessment of risk based on relevant and accurate actuarial and statistical data. It is up to insurance industry to prove this. Moreover, any sex differentiation of premiums based on pregnancy and maternity is forbidden. this paper for the first time respective calculations for Germany are presented, using cost profile data from both private and public health insurance and life tables used by private insurance companies. As the calculations demonstrate, higher premiums for women can be justified actuarial. However, sex-based differences in costs vanish if costs of pregnancy and maternity are excluded. Thus, unisex tariffs become necessary. tariffs, however, bear dangers with respect to cream skimming by insurance companies and adverse selection behaviour by men. Thus, a risk equalisation scheme is suggested which guarantees both, equal premiums for men and women to pay, but sex-specific premiums for insurance companies to receive.

0134 Health, human capital and the intergenerational transmission of poverty Rungo P. (A Coruna University, A Coruna, Spain) Currais L., Rivera B. This Paper develops a theory of fertility and choice between investment in education and health-related spending. Fertility differential are generated by the assumption that parents’ productivity as teachers increases with their own human capital. The existence of a fixed time-cost of rearing a child determines a trade-off between quantity and quality of children. Quality (human capital) depends both on education and physical status, which may be controlled by parents through health-related spending (nutrition). Taking into account the effects of health on human capital, the model shows that a minimum physical status is necessary before parents optimally choose to invest in education. Furthermore, healthrelated spending above certain level, even with no investment in education, may lead an economy to a high human capital and low fertility equilibrium. The model also generates a poverty trap related with a low level of human capital and a consequent insufficient expenditure in nutrition and health-related spending.also test the role of health status in the intergenerational transmission of poverty. We use the National Health Spanish Survey in order to link the health status of children and their future income with the health status and human capital of parents. We find strong evidence that support the theoretical findings.

Eur J Health Econom Suppl 1 · 2006

| S83

0418 Individual health expenditure and income Ruz Torres R. (Department of Applied Economics, Université Libre de Bruxelles, Brussels, Belgium) The current Belgian risk adjustment formula is based on individual data. Different socio-economic status variables are included. Because it was not available for every individual, income is not, income is generally seen as an indicator of ability to address the need for health expenditure, and as a factor in influencing health status. The hypothesis, therefore, is that income and health expenditures are related. use cross-sectional individual data for 1998 to explore the relationship between income and health expenditures. Analysis of the residuals of the regression used in the risk adjustment scheme reveals a non linear relationship with income. The current risk adjustment model’s predictions underestimate health expenditures for low income individuals. The model is also less precise for low income individuals, suggesting that morbidity related variables are missing.a risk adjustment setting, these findings imply incentives for the insurers to risk select against low income individuals. The current risk adjustment model needs improvement, by refining the socio-economic variables included, and by adding diagnostic-related variables.a risk adjustment setting, these findings imply incentives for the insurers to risk select against low income individuals. The current risk adjustment model needs improvement, by refining the socio-economic variables included, and by adding diagnostic-related variables.

0090 Accounting for differences between EQ-5D and SF-6D scores in a sample of the general population Sach T. (University of Nottingham, University Park, Nottingham, United Kingdom) Barton GR. Objective: To identify when, and to what extent, the EQ-5D and SF-6D give different utility estimates. Methods: Information on six personal characteristics (age, gender, ethnicity, smoking status, body mass index and occupational skill level), ten health conditions (back pain, hip pain, knee pain, heart disease, stroke, asthma, cancer, diabetes, rheumatoid arthritis and osteoarthritis), the EQ-5D and SF-6D was requested from 2770 patients aged ≥45 years in one general practice. Regression analysis was used to estimate variation in the difference between the EQ-5D and SF-6D scores, the six personal characteristics and ten health conditions acted as explanatory variables. Results: The EQ-5D estimated the mean utility of patients in the best health state (with the most favourable personal characteristics and none of the ten health conditions) to be 0.096 higher than the SF-6D. Conversely, compared to the SF-6D, the EQ-5D estimated the mean loss in utility associated with back pain, hip pain, knee pain and osteoarthritis to be significantly (p<0.01) greater, and the mean utility of less healthy patients to be up to 0.274 lower overall. Conclusion: There is the potential for the EQ-5D and SF-6D to give quite different utility estimates, particularly for individuals at either end of the health spectrum.

0208 Costs and benefits of nursing home care Sadiraj K. (SEO Economic Research of Amsterdam, Netherlands) Kok L., Stevens J., Gameren E., Woittiez I. The paper describes a cost-benefit analyses of nursing home as opposed to elderly living in their own home. The analysis is based on surveys of

S84 |

Eur J Health Econom Suppl 1 · 2006

elderly living at home and elderly living in a nursing home. The survey contains detailed information on the physical and mental disabilities of the elderly as well as demographic and social characteristics. The costbenefit analysis requires that the two groups have similar characteristics. We achieve this by first, estimating a model that explains the probability whether an elderly lives at home or in a nursing home, given individual’s (health, social and demographic) characteristics. Then we use this model to predict for each elderly his/her preferred state. Elderly that actually lived in a nursing home but were predicted by the model as living at home is the comparison group to the elderly living at home. We calculate for these two (comparable) groups public health costs related to living in a nursing home or at home, costs for the elderly themselves (including differences in happiness) and other social costs like rent subsidies and informal care. We conclude that nursing home care is much more expensive for the society as a whole then the costs made for elderly living at home. For the elderly individuals there is no significant difference in costs between the two groups.

0332 Predicting hospital length of stay and treatment cost under capacity constraint Saka O. (King’s College London, Division of Health and Social Care Research, London, UK) McGuire A., Bonet M.J., Rudd A., Wofe C. Length of stay is commonly taken adopted as a performance measure within the hospital sector. However length of stay may be an endogenous variable influenced by, for example, capacity constraints and discharge policies. This paper models uses patient level data on stroke in a survival model to estimate the conditional length of stay in a stroke unit. The model analyses the influence of the responsiveness of Social Ser vices and/or primary care in making arrangements for discharge and the fluctuations in the number of beds available in the stroke unit on the probability of discharge. The influence of a number of factors, including changes in discharge factors (capacity, early discharge policy, bed transfers, social service responsiveness and mortality rates) on the cost of running services and the outcomes is also assessed. It is noted that the probability of discharge is correlated with factors beyond the control of the hospital, such that discharge policy may be considered partly exogenous. There appears to be a wider variation in discharge probability the closer a patient comes to end of hospital stay, reflecting an increasing uncertainty of placement at the individual level for specific case types.

0015 Technical Efficiency, Economies od Scale, Ownership and Management in The Brazilian Hospital System Sampaio de Sousa M.C. (Universidade de Brasília, Brazil) Proite A. In this paper, we estimate DEA (Data Envelopment Analysis) technical efficiency scores for 1170 Brazilian hospitals included in the SUS (Central Health System) using a recently proposed method that combines bootstrap and jackknife resampling to eliminate the influence of outliers and possible measurement and recording errors in the data. We use the variable returns to scale variants of the DEA method. After computing the efficiency scores, we use quantile regression, to investigate the determinants of those scores. Our results confirm the importance of the scale effect (measured by the total number of services) over the technical efficiency of this sector, controlled by the average permanence in the hospital, its average costs and human capital variables. Regarding hospital’s management, we highlight the negative effects of non-profit organizations over the efficiency, which differs from the empirical literature. Excessive specialization also has a negative effect over the efficien-

cy scores, pointing out the existence of an optimal mix between specialization and generalization hospitals’ characteristics. As for property, private hospitals have its performance negatively affected for those units over the 30th efficiency quantile, suggesting that market failure effects are important in this sector.

0420 Social Capital and Health in England. Sassi F. (Department of Social Policy, London School of Economics and Political Science, London, United Kingdom) A growing body of literature suggests that aspects of social capital are strongly associated with health and healthy lifestyles. However, largescale studies are often hindered by the limited availability of valid and relevant measures of social capital.pooled 2000-03 Health Sur vey for England (HSE) data we assessed the impact of four dimensions of social capital (trust, social support, individual and community participation in membership organisations) on a range of health indicators. developed a multi-level logistic model, accounting for the contribution of individual and area-level variables (including deprivation, income distribution and residential segregation) on perceived health, psycho-social health, limiting long-standing illness, and cardiovascular disease. A novel employment-based index of social capital (derived from Annual Business Inquiry data) was used to assess community participation in membership organisations. Other social capital indicators were drawn directly from HSE.social capital appears more strongly related to health than community social capital. In both cases, however, the relationship is not linear and health appears to improve first and then decline slightly as social capital reaches higher levels. This may be consistent with an inverse relationship between “bonding” and “bridging” social capital at such high levels.

0198 Financing Health Care in Germany Sauerland D. (WHL Graduate School of Business and Economics, Lahr, Germany) The ongoing political discussion about health care reforms to ensure a sustainable financing mode for health expenditures is focused on two proposals: a statutory insurance with flat rate premiums vs. a statutory insurance with income related premiums. Objectives: This paper offers a projection of the future development of health care expenditures in the Statutory Health Insurance system. It outlines the financing burden for the insured in case of the two proposals discussed. Methods: OLS regressions are used to find out the determinants of real per capita SHI expenditures. Based on the best long-run estimation, the development of future expenditures is forecasted. Results: The forecast shows an increase in expenditures to about 500 bill. € in the year 2040. The forecasted flat rate premium goes up to about 700 € per capita. In case of income related premiums, an annual increase in income of about 4 percent is necessary to ensure stable premium rates. Conclusion: The driving determinants of SHI expenditures, i.e. a higher income, an ageing population and the technical progress, are not changed by the different ways of financing discussed. However, an ongoing economic development with low GDP growth rates will make the financing burden even higher.

0143 The Role of Community Social Capital in the Reducing the Prevalence of Serious Mental Illness Scheffler R. (University of California at Berkeley, USA) Brown T.T. We show that lagged community social capital is strongly and inversely related to the prevalence of serious mental illness and also exhibits strong diminishing returns. We estimate a mental health production function, which controls for sex, age, race/ethnicity, marital status, education, and individual social capital, as well as unobserved area-level heterogeneity using three years (1999–2001) of U.S. data on 48,222 adults. The presence of mental health is measured using a valid and reliable indicator of serious mental illness the Kessler K6. Our measure of community social capital is a new validated ecological measure of community social capital, the Petris Social Capital Index. We estimate that an increase of 0.4% in community social capital in selected communities would reduce the number of individuals with serious mental illness in the U.S. by approximatley 29,000.

0398 Selection or incentive? An analysis of the effect of choice of health insurance deductibles on physician visits using matching techniques Schellhorn M. (GSF – National Research Center for Environment and Health, Institute of Health Economics and Health Care Management, IGM, Neuherberg, Germany) Gerfin M. In Switzerland, basic health insurance is mandatory and each individual is insured separately. The insurance premium varies by region of residence but is independent of income and risk. The insured face a minimal annual deductible. Annually, they are given a choice of higher deductibles to reduce their insurance premium by a regulated percentage. The choice of a higher deductible sets incentives for a more cautious utilization of health services. Clearly, the choice is made based on expected health service utilization. In this paper, the effect of the choice of a higher than the minimal deductible on physician visits is analyzed. To disentangle incentive from selection effects we employ propensity score matching techniques for the choice of deductible to control for the endogeneity of this choice. We differentiate between the effects of the deductible on the probability and frequency of primary care physician and specialist visits. We use data from the 2002 Swiss Health Survey which contains information on the relevant health, health care utilization and insurance status variables for the current and previous year allowing us to adequately model the choice of deductible and subsequent physician visits. The main finding is that about two thirds of the observed lower utilization for individuals with a high insurance deductible is caused by selection effects. Incentive effects have a bigger impact on the probability of visits than on their (conditional) frequency. Preliminary findings indicate no significant difference between primary care physician and specialist visits.

0307 The Global Health Financing Environment Schieber G. (World Bank, Washington D.C., USA) Langenbrunner J. This presentation assesses the global health financing situation and its implications for health reforms for the low and middle income countries in the Europe Central Asia (ECA) Region. The huge global inequities in disease burden versus health spending, the difficult tradeoffs faced by countries in terms of providing universal access to an essenEur J Health Econom Suppl 1 · 2006

| S85

tial package of ser vices as well as financial protection to their populations, and the critical fiscal and macroeconomic dimensions of scaling up health systems are discussed. First, the global health financing situation is assessed in terms of disease burden and health expenditure patterns. Second, the importance of out-of-pocket spending, the need to improve risk pooling and methods for improving risk pooling are highlighted. Third, the importance of external assistance and the need to reform the international aid architecture are assessed. Fourth the importance of fiscal space in order to accommodate increased spending is analyzed along with methods to increase it. Fifth, the critical epidemiological, demographic, economic, and political dimensions for health financing reform policies in ECA low and middle income countries are discussed.

0156 Co-payments and risk adjustment Van de Voorde C. (Catholic University of Leuven, Belgium) Schokkaert E. Out-of-pocket payments by patients play an important role in the Belgian system of compulsory health insurance (a) They are set by the central regulator and differentiated for different health care items with the purpose of influencing patient behavior. This feature is neglected in the risk adjustment system. (b) Weaker social groups are partly exempted. A uniform maximum billing system has been introduced with an incomedependent ceiling. Sickness funds are fully compensated for the resulting higher costs.analyze these policy options using a database with individual observations for the Belgian population. Although the policy option with respect to (b) removes all incentives for the sickness funds to fight moral hazard among the weaker social groups, it can be defended on equity grounds. Differentiated co-payments are a relatively efficient way to improve accessibility, if they are fully compensated for in the risk adjustment system. At the same time, it could be advisable to give the funds more freedom to vary the co-payments for different health care items in order to stimulate behavioral changes by the patients. The resulting cost differences should not be compensated for through risk adjustment.

0442 The use of costing methodologies to determine prices for inpatient health services: a review of nine European countries Schreyögg J. (Department for Health Care Management, Institute for Technology and Management, Berlin University of Technology, Germany) Objectives: For economic evaluations as well as cross-country cost comparisons it is useful to know to which extent prices and actual costs differ in each country. This paper compares costing practices to determine prices for inpatient health ser vices between nine EU member states (Denmark, France, Germany, Hungary, Italy, the Netherlands, Poland, Spain and United Kingdom). It is investigated how costing information is derived to determine prices, how prices are set and to which extent prices reflect the actual costs incurred. Methods: Partners of the EUHealthBasket project in each country were asked to provide detailed descriptions and analyses of costing practices and price setting in each country. Results were compared according to defined criteria in order to reveal commonalities or discrepancies. Results: Although most countries use DRG-systems, costing methodologies largely differ regarding sample size, data quality and statistical approaches. This especially applies to the number of hospitals included in the sample to derive cost data varying from 8 in Italy to 221 in Germany. There is a trade-off between the representativeness of the data sample and the data quality determining the extent to which prices reflect the actual costs incurred.

S86 |

Eur J Health Econom Suppl 1 · 2006

Conclusion: Researchers should carefully distinguish between prices and costs and consider that there might be large differences between them depending on the country.

0465 The impact of co-payments on patient behaviour: evidence from a natural experiment in Germany Schreyögg J. (Department for Health Care Management, Institute for Technology and Management, Berlin University of Technology, Berlin, Germany) Grabka M., Busse R. Objectives: As part of the Statutory Health Insurance Modernization Act a co-payment of €10 per quarter for the first visit at a physician’s or a dentist’s office has been introduced with effect of 1st January 2004. The study investigates whether co-payments changed patient behaviour without discrimination of certain population groups. Methods: Data of the German Socio Economic Panel (SOEP) from 1995-2004 (n=20.821) is used for the study. Two logistic regression models are performed considering socioeconomic variables as well as the state of health. Two control groups are constructed and compared by independent t-tests. Results: Physician visits declined in the year 2004 compared to the year 2003 and before. The decline was significantly higher for insured affected by co-payments than for those which were not affected. However the share of those patients who at least had one physician visit in both years remained stable. The regression models point out that no discrimination of persons with low social status, disabilities or poor health could be observed. Conclusion: It seems plausible, that the introduction of co-payments has contributed to a reduction of physician visits without discrimination of certain groups.

0576 The european collaboration on assessment for reimbursement: achievements and futher challenges Schuurman A.R. (Head of the Reimbursement Department of the Health Care Insurance Board (CVZ), Brussels, Netherland) The are growing possibilities for European co-operation – and the necessity of such co-operation – in assessing drugs for reimbursement. The collaboration is between the national competent pharmaceutical authorities responsible for the assessment of (new) medicines for reimbursement and/or for definitions, pricing and reimbursement conditions. The main tasks for the assessment of medicines: Monitoring (new) drug assessments in all (30) countries, and exchanging information on these current assessments on a weekly basis. Exploring and developing the methodology and criteria for the assessment of drugs for reimbursement. Reviewing existing guidelines for assessing medicines for reimbursement and exploring the possibilities for harmonising these guidelines, starting with the area of clinical effectiveness. Discussing criteria for the re-assessment of medicines for reimbursement in the light of new information on clinical effectiveness and indications, and sharing information on the review of national reimbursement decisions. Discussing criteria for therapeutic equivalence and interchangeability. Collaborating on the post-listing reviews (for example, Vioxx, Celebrex) and adjusting indications. Discussing arrangements for Reference pricing and (therapeutic) clustering. How the results of an assessment will be used is of course the responsibility of the national authorities, according to the subsidiarity principle. The perspective of common, adequate assessments lowers the so called 4e hurdle to a small hickup.

0095

0598

HEE-GER: A systematic review of German health economic evaluations

Policy externalities – the health effects of agricultural, trade and foreign aid policies

Schwappach D.L.B. (University Witten-Herdecke, Witten, Germany) Boluarte T.

Scott G. (Department of Applied and International Economics, Massey University Wellington Campus, New Zealand) Scott H.

Objective: HEE-GER is a systematic literature review that analyzes health economic studies that assessed health care under country specific settings for Germany. We provide an overview of the characteristics of the reviewed evaluations and their quantity and quality. A central part of the study is the extraction of techniques and sources used to derive QALYs in published cost-utility studies. Methods: We searched generic and specific databases in April 2005 for relevant journal articles published between 1990-2004 using “high sensitivity” search strings. The references of retrieved articles and relevant German language journals were manually searched for additional material. German experts and recognized authors were asked to provide bibliographic data. Studies were included in the review if they fulfilled a set of inclusion criteria. The full texts of retrieved studies were assessed by two independent reviewers using a data extraction sheet. We collected details on study objective, inter vention type and disease category, design, methodology and background data. Data relating to health and preference measurement methodology were extracted in detail. Results: The systematic literature search initially identified 2,158 articles of which 734 studies were included in the critical appraisal. The results of the quantitative assessment of the characteristics of these studies are reported.

Background: Health, agricultural, trade and foreign aid policies are frequently formulated and implemented in isolation and by different people with dissimilar objectives and knowledge sets. Aims: To investigate the impact that agricultural support and protection of domestic production and associated trade and foreign aid policies may have on health outcomes and social welfare. Method: The literature is reviewed and a number of cases selected that illustrate actual and possible public policy linkages between health and agricultural related support and protection. Results and conclusions: Three groups of countries categorised by the degree of government intervention in agricultural markets and economic development were isolated. These groups of agricultural producers were; developed economy price setters, developed economy price takers, and developing economy price takers. Results suggest that the “isolationist” approach to agricultural policy development and implementation may result in unintended deleterious domestic and international health effects. Public policy making requires an integrated approach and informed tradeoffs between policy and political objectives.

0599

0189

The economic cost of treated and untreated sleep apnoea to New Zealand

Does relative pay influence shortages of doctors and nurses? Implications for the cross-national mobility of health professionals.

Scott H. (ScottEconomics, Wellington, New Zealand) Scott G., Mihaere K., Gander P.

Scott A. (Melbourne Institute of Applied Economic and Social Research, Melbourne, Australia) Elliott R., Skatun D., Ikenwilo D., Bell D., Roberts E.

Background: The consequences of sleep apnoea range from loss of productivity, co-morbid disease, injury and death. While there is a growing body of knowledge on the prevalence of sleep apnoea in New Zealand little is known about the economic cost effects. Prevalence and cost information provide essential components of the data required to conduct cost-effectiveness analysis and formulate appropriate policies. Aims: To investigate the economic cost of treated and untreated sleep apnoea. Method: The prevalence of sleep apnoea, the probability of each outcome and unit costs of each outcome was used to calculate the total economic and social cost of the condition. An outcome tree was developed where the population with sleep apnoea were tracked as to likely decisions, the usual treatment choices available in Wellington, New Zealand, and outcomes. A probability and cost were attached at each outcome event which enabled us to calculate the costs for the untreated and treated arms. Monte Carlo simulation was used to take account of uncertainty in the estimates of prevalence, probability of outcomes and unit costs. Results and conclusions: Interim results suggest that the costs of sleep apnoea and the differences in cost between treated and untreated individuals are such that treatment strategies and protocols should be developed.

The impact of pay on the recruitment and retention of health professionals has been cited as a key factor in motivating doctors and nurses to change jobs or alter their labour supply. This paper reports the results of a study in the UK NHS that examined the impact of pay differentials across geographical areas on vacancy rates for doctors and nurses in those areas. The econometric results show that relative pay is important for nurses but not for doctors. This suggests that nurses supply their labour to defined local labour markets, whereas doctors are more likely to work in a national labour market. The results have implications for the type of policies used to influence recruitment and retention. Nurses’ employers are likely to compete with other locally-based employers and so flexibility in setting pay is important in addressing shortages. However, the nature of competition in the doctors’ labour market is likely to be based on training opportunities and reputational concerns that require a different type of policy response. Different sets of factors are therefore likely to influence the cross-national mobility of doctors and nurses, including the level and flexibility of remuneration and the competitiveness of within country labour markets, in addition to other factors such as relative licensing and training requirements and quality of life.

0603 Regulating quality in health care – the issues facing the english national health service Scrivens E. (Health Care Standards Unit, Keele University, UK) The NHS in England is undergoing considerable structural change to introduce more locally accountable hospitals, greater patient choice, payment by results and local General Practioner based commissioning of Eur J Health Econom Suppl 1 · 2006

| S87

ser vices. There is a need to ensure that the reformed NHS provides incentives for service improvement, whilst at the same time maintaining consistency in access to ser vices and the safe delivery of ser vices. The UK government is committed to reducing bureaucracy and the demands for data collection, which is based upon associated requirements to reduce the overall costs of regulation, whilst increasing the accountability of health services.This requires a new approach to regulation, with universal national tariff setting to equalise the distribution of resources, whilst balancing a standardised concept of quality with incentives to encourage local innovation in service design and delivery. This paper examines the current costs of existing regulation, the issues that need to be considered in the design of such a regulatory system which must operate at reduced costs, and in particular the compares the costs and impact of different models of regulation using standards and quality assurance processes and their contribution to the improvement in regulation.

0378 Reporting Heterogeneity and Social Health Inequalities in France Sermet C. (IRDES, Paris, France) Debrand T., Jusot F. Sermet C. Measuring health is a crucial issue to understand social health inequalities. Do social variations in reported health status reflect differences in “true health” or are they due to reporting heterogeneity? Recent literature suggests that for the same “true” health status, reporting of health depends on age, gender, education, occupation, ethnicity and country. the “true” health status cannot be observed, we propose to approach it by a latent variable estimated by a multiple indicators multiple causes structural equations modelling (Shmueli, 2003). This latent variable is measured by the three health indicators promoted by the European Office of WHO self-assessed health, long-term illness and limited activities. Moreover, this modelling allows to distinguish the impact of socioeconomic characteristics on “true” health status from the reporting bias induced by socioeconomic characteristics on each health indicator.the 2002 French Health Sur vey, we estimate this model to compare the health inequalities with this latent health status indicator and with the three usual health indicators.

0053 Diffusion of medical innovations Serra-Sastre V. (LSE Health and Social Care, London, UK) Technological change has been identified as a leading factor explaining increasing health care expenditure in a number of OECD countries over the past few decades. This has generated a growing interest in the adoption and diffusion process of innovations in the health care sector. There have been few empirical studies analysing this process. The aim of this paper is to provide evidence on the adoption process of new drugs within the UK NHS primary care sector. Specifically, a panel data model is presented to consider the determinants of the diffusion of statins, a class of cholesterol-lowering drugs that protect against cardiovascular disease introduced into the UK in the early 1990s.diffusion process is examined at the micro level. It is hypothesized that, controlling for socioeconomic incentives and demographic characteristics of the market where potential adopters operate, characteristics of GP firms play a major role. The analysis provides evidence on the degree of substitution and expansion of statins over the period 1992-2004. The findings suggest that specific regulatory factors influence diffusion but that particular individual GP practice characteristics are of greater importance. The paper thus offers guidance on which factors to target if policy-makers wish to influence the diffusion process.

S88 |

Eur J Health Econom Suppl 1 · 2006

0044 Applying Behavioral Economics to Changing Health Behavior: the case of Weight-Loss Management Heshmat S. (University of Illinois at Springfield, USA) Rationale: Americans spend in the aggregate many billions of dollars each year trying to lose weight through dieting or exercise, indicating the nation’s desire to slim down. Despite documented short-term success, dieting has a very low success rates, most dieters regain their weight back within 3-5 years. Objectives: This paper presents some insights from behavioral economics to the various aspects of overeating, dieting, and the relapse as discussed in the weight-loss literature. The objective of this paper is to discuss why people fail to stick to their goal for eating healthy diet in order to lose weight. The understanding what drives these decisions is a critical part of health prevention. Methodology: The paper provides a review of past studies from the health economics and behavioral economics literatures. The purpose is not to provide an exhaustive summary of the vast literature relating to this issue, but rather to provide a framework in which to take a fresh look at a persistent public health challenge. Conclusions: Behavioral economics takes the premise that many of the factors that lead to relapse stem from faulty logic, and irrational assumptions, leading to inconsistent behavior. Dieting behavior appears to be prime candidate for benefiting from behavioral economics, such as how people deal with outcomes that are uncertain and how they discount delayed costs and benefits. The prevalence of these biases suggests that there is room for improvement. Similarly, the knowledge of these biases should help dieters to develop problem-solving skills in weight management. This paper demonstrates the role of behavioral economics in changing people behavior toward better eating habits, and suggests how to put that insight to work in programs that tackle weight loss program

0331 Optimal age grouping for capitation contracts Shmueli A. (Hebrew University of Public Health, Jerusalem, Israel) Bernice Oberman M.A, Zmora I. Age remains a central risk-adjuster in many systems. Most countries use 5- or 10-year age-cohorts, following the traditional practice in demographic population accounts (“conventional age grouping”). The objective of this study is to derive an optimal age grouping for capitation contracts, which will exhaust the information embedded in the data on medical cost by age, for the total population, for men and women. Cost data were collected for a sample of over 100,000 insurees of Clalit Health Services, the biggest sickness fund in Israel, in the year 2004. We use Classification And Regression Trees (CART) to separate out optimal age groups. Several measures are used to evaluate the resulting classifications in comparison with the conventional groupings. The main characteristics of the optimal grouping relative to the conventional one are the single-year group following birth (age 0-1), followed by relatively large groups at young ages, narrowing unequally with advanced age. The optimal grouping is different for men and women. Conventional age grouping for capitation contracts results in sub-optimal age grouping. Capitation contracts based on conventional grouping might have adverse incentives for risk selection.

0370 Private health expenditures in social healthcare systems Shmueli A. (Hebrew Universityof Public Health, Jerusalem, Israel) Bernardi B., Endveld M., Achdut L., Attilla V. Private out-of-pocket health spending is an expression of consumers’ sovereignty, trying to achieve better health facing the public package of benefits. However, since there is neither cross-subsidization nor solidarity in private out-of-pocket expenditure, it has always been an open social and equity issue in social healthcare systems. The present analysis uses national Family Expenditure Sur veys from 2000/1 to examine private health expenditures in a comparative three (UK, Italy and Israel) social healthcare systems context. It uses national age-based capitation rates to provide a reliable measure of the households’ health needs, allowing us to focus on horizontal and vertical (in)equity in the consumption of six types of private health expenditures: total health expenditure, voluntary insurance premiums, total excluding insurance, optical expenditure, expenditure on medical and dental care, and expenditure on medications. The cross-countries differences are related to the variety of the healthcare systems: an old NHS in the UK, a new NHS in Italy, and a competitive social insurance in Israel.

0339 Is waiting-time prioritisation welfare improving? Siciliani L. (Economics Department, University of York, UK) Gravelle H. Rationing by waiting time is commonly used in health care systems with zero or very low money prices. Some systems prioritise particular types of patient and offer them lower waiting times. We investigate whether prioritisation is welfare improving when expected benefits vary across patients and are the sum of observable and unobservable components. If the observable component is dichotomous (for example gender) prioritisation is welfare improving only if the demand elasticity of the group with higher expected benefit is lower (in absolute terms) than the demand elasticity of the other group. If the observable component is continuous (for example depends on age, which enters negatively in the utility function), waiting-time prioritisation is welfare improving if the mean age (or observable benefit) of the patients who are indifferent (between receiving the treatment or not) is higher than the mean age of the patients who receive the treatment. This is certainly the case if the joint density of observable and unobservable components is uniform or normal (it has no effect if density function is negative exponential). We also show that a more extreme form of prioritisation, where patients whose observable component is below a threshold are denied treatment, can further increase welfare. Keywords: waiting times, prioritisation, rationing.

0340 Optimal contracts for health services in the presence of waiting times and asymmetric information Siciliani L. (Economics Department, University of York, UK) This paper presents a model of optimal contracting for health services in the presence of excess demand and waiting times. We assume that i) hospitals differ in their demand for treatment; ii) potential demand is private information of the provider; iii) specialists can dump patients; iv) activity and waiting times are contractible; v) providers are semialtruistic.show that if differences in potential demand between hospitals are visible to the purchaser, then the optimal contract is for the purchaser to offer a transfer to the provider in exchange for the provision of the desired level of activity and waiting time, without leaving

any rent to the provider. Also a mix of implicit rationing (through waiting time) and explicit rationing (through dumping) may be optimal. Rationing by waiting alone induces excessive disutility for the patients, while rationing by dumping alone generates excessive disutility for specialists and excessive costs. the more realistic assumption of asymmetric information, a separating equilibrium exists when it is optimal for the purchaser to contract more activity and longer waiting times to those hospitals with higher demand. If disutility from dumping is high (low) compared to the degree of altruism, hospitals with low (high) demand gain informational rents. A pooling equilibrium may also occur.

0614 Report on UK work to measure healthcare output, and productivity, through an aggregate measure which incorporates some quality adjustments, with proposals for future improvements Simkins A. (Department of Health, London, UK) Report on UK work to measure healthcare output, and productivity, through an aggregate measure which incorporates some quality adjustments, with proposals for future improvements. Objective/Aims: The objective of this presentation is to present the principles and methods used by the UK Department of Health to improve the measure of NHS output used in the National Accounts and in ONS productivity measurement, following the Atkinson Report (ONS, January 2005) on measurement of government output and productivity. Figures published by DH in December 2005 and ONS in February 2006 have shown a rise in quality adjusted productivity, on certain assumptions, but some of these are controversial and further debate and development is required. Method: The DH paper, Accounting for Quality Change, combined a summary of research commissioned from York/NIESR (separate ECHE submission from Andrew Street) with proposals for other quality adjustments. The main elements were: Rate of mortality within 30 days of hospital admission; Estimated benefits of value from hospital treatment; Changes in average waiting time for elective treatment; Sur vey evidence of improved patient experience (access, involvement in decisions etc); Improved blood pressure control in primary care; Using a value weight rather than a cost weight to reflect lives saved by prescribing statins; Annual rise in value of health in line with GDP per head. Result: The combined methods added 2.7% a year, on average, to output growth. The largest items were the ‘value of health’ and the value weight for statins. However, there are large data gaps including lack of routine information about the benefits from treatment. Separate analysis by the Office for National Statistics (ECHE submission from Phillip Lee) which compared output growth with growth in deflated inputs, using the same quality adjustments as DH, gave a range of estimates of productivity change up to 1.6% a year – but with many qualifications on interpretation. Conclusion/discussion: Measurement of aggregate healthcare output and productivity, for National Accounts or linked purposes of public accountability, requires methods which take account of quality. Quality of healthcare has several different domains and there are challenging issues on measurement and interpretation. However, the paper demonstrates that assumptions can be made which enable partial information to be combined into an overall assessment of productivity change. Further development work is required, and would benefit from wider debate on concepts and approaches. It would be very beneficial if similar methods could be used in different countries, both to assist in methodological improvements and to allow for comparisons in the efficiency of different healthcare systems and the success of different approaches to improving their performance and productivity.

Eur J Health Econom Suppl 1 · 2006

| S89

0427 Comparing the properties and performance of generic, single index number HRQoL instruments Sintonen H. (University of Helsinki, Dept.of Public Health and FinOHTA, Helsinki, Finland)9 There are a few generic, single index number HRQoL instruments like 15D, AQoL, EQ-5D, HUI and SF-6D, but none of them can claim a gold standard status. To be suitable for QALY calculations, the scores they produce should reflect a credible trade-off between length and quality of life. There are some theoretical comparisons between the instruments, but quite few studies comparing their properties and performance e.g. in terms of discriminatory power, responsiveness to change and credibility of scores in QALY framework in different populations, patients groups and situations. The purpose of this session is to encourage such comparisons and to bring together empirical findings from them so that the potential users of the instruments would became more knowledgeable about the pros and cons of different instruments.

0322 Empirical analysis of changes in the demand for prental health services in rural areas when economic incentives are introduced. Sosa-Rubi S.G. (Centre for Health Economics, University of York, UK) We investigate the shift in the patterns of use of prenatal ser vices in rural areas after the implementation of a social programme named Progresa. Cash transfers in this programme require active participation by the recipient households in exchange for the benefits, such as a minimum of number of visits to prenatal services. Two-part model (TPM) is applied to show that with a certain degree of homogeneity among users and providers such as when a single episode of illness is recalled could be an appropriate model for studying the demand for prenatal health services. Good performance of TPM-negative binomial model was found in the sub-sample of pregnancies recalled before the initiation of Progresa programme. However due to changes in the distribution of the prenatal visits in the sub-sample influenced by Progresa, difficulties in identifying the dichotomy between zero and non-zero users were found. The increase of the number of visits to prenatal services and the reduction of the number of women without any approach to the health services during their pregnancy was found. This reveals that financial incentives granted by Progresa goes beyond the individual of family’s constraints to use health ser vices, allowing mothers in rural areas to access basic health services during their pregnancy.

0600 Colon-rectal cancer screening, diagnosis, and treatment guidelines Spandonaro F. (CEIS Centre for Health Economics and Management (CHEM), Faculty of Economics, University of Rome Tor Vergata, Rome, Italy) Ratti M. Until now, Italian guidelines have focused their attention predominantly on clinical/epidemiological issues. With the revision of “Colon-rectal cancer screening, diagnosis, and treatment guidelines”, ASSR and CHEM are aiming to improve the role of economic evaluation, and for this reason have formed a Research Group (composed of experts from the regions, Academic Institutions, and Industry). The Group developed a methodological economic evaluation framework usable for “Colon rectal cancer guidelines” and designed a general methodology applicable for future guidelines in Italy, comparing it with the best International experiences on this topic.

S90 |

Eur J Health Econom Suppl 1 · 2006

Furthermore we faced with problems deriving from the new federalistic shape of the Italian NHS; every assessment needs to consider regional specificity (incidentally note that the importance of ASSR is more and more increasing due to that) but, at the same time, must guarantee some uniformity and suitable mobility characteristics in the place of application. Furthermore, the Group applied the methodology in assessing the virtual colonoscopy technology. The aim is to support Italian regions in evaluating and comparing costs and output in health care activities.

0096 Economic Evaluations for Reimbursement and Pricing Decisions in France: Can Pharmacoeconomic Evaluation Methods be transferred to Medical Devices? Späth H.M. (ISPB – Faculté de Pharmacie, Lyon, France) Taboulet F., Carrere M.O. Background: Pharmacoeconomic evaluation activities have grown in recent years, but few economic evaluations (EE) have focused on Medical Devices (MD). Nevertheless, the French Authorities suggest that MD manufacturers include EE in reimbursement and pricing negotiations. This study addresses the barriers to conducting EE of MD, in comparison with pharmacoeconomic evaluations. Methods: First, we studied the differences between MD and drugs that impact on the completion of EE. Then, we analysed items of the French “Guidelines for EE of Health Care Technologies” that might be barriers to EE of MD, as compared to drugs. Results: The six most important barriers to EE of MD are related to: (1) the feasibility of clinical trials, (2) the difficulty to define a comparator, (3) the fact that MD effectiveness often depends on the operator; (4) the short life cycle of MD, (5) the professional culture of engineers, (6) costs of EE activities. At the conference, we will present an analysis of these barriers, and recommendations allowing to overcome them. Conclusion: We recommend setting up multidisciplinary groups of engineers, healthcare professionals and economists from the beginning of MD development; and defining MD to be evaluated in priority, on which EE methods should be tested before being applied to others.

0253 Reforming Long-term Care in Germany Spermann A. (Centre for European Economic Research (ZEW), Mannheim, Germany) Arntz M., Michaelis J. Given projections regarding the number of benefit recipients in the long-term care social insurance (LTC-SI) in Germany, cost pressures are expected to rise continuously. Thus, reforming the LTC-SI appears to be the only possible route in order to ensure current care levels for future generations of the frail elderly without strongly increasing the social contribution rates for the LTC-SI. Since, from the LTC-SI perspective, nursing home care is much more expensive than home care, inducing a higher share of benefit recipients to stay at home becomes a natural starting point in order to reduce cost pressures. Thus, one main objective of the reform of the LTC-SI is a higher degree of flexibility in the LTC-transfer system that allows for more stable home care arrangements. One way to introduce more flexibility to the current system is to allow for a matching transfer in addition to the currently available inkind transfers and lump-sum transfers. A matching transfer allows for an extended spectrum of care services and also provides a supportive case management. The paper looks at the matching transfer from a theoretical perspective using a simple household bargaining approach and presents preliminary findings from social experiments in seven sites in Germany.

0160 Health Care Reform in Ukraine Staat M. (University of Mannheim, Germany) Pilyavsky A. We analyze the technical efficiency and efficiency change for 193 community hospitals and polyclinics across Ukraine, for the years 1997 to 2001. These facilities are a subset of the central medical institutions in rural Ukraine; they have an identical function in the health system and share the same departmental structure. The data comprise the number of staff in the departments as well as the polyclinics, the number of patient days, inpatient and outpatient admissions as well as the number of surgical procedures, lab tests and x-rays performed. Finally, the number of deaths and deaths after surgery are used as quality proxies. We employ an output-oriented order-m estimator, a robust nonparametric technique, to assess the efficiency of health care providers and changes of their productivity. The efficiency scores are close to unity for hospitals whereas polyclinics are somewhat less efficient. The Malmquist-index exceeds unity for three out of four periods for both hospitals and polyclinics indicating improved productivity on average. Whereas progress for hospitals is in the range of 2% to 4.5%, the index indicates progress for polyclinics in excess of 10% for the final period. This is the only result indicating a substantial effect of the reforms taking place.

0151 The Economic Impact of Chronic Diseases Stanciole A.E. (Department of Economics and Related Studiesof York, UK) Abegunde D. Chronic diseases are a question of increasing relevance for developing countries. The processes of urbanization and industrialization in developing countries are associated with the increase of several risk factors, such as obesity, smoking and the lack of physical activity. Consequently the already high rates of morbidity and mortality of chronic diseases in developing countries are projected to increase even more in the coming years.paper explores the effect of chronic diseases on households using micro data from the Living Standards Measurement Studies (LSMS) of the World Bank for Brazil (1996), India (1997) and Russia (from 2000 to 2003). We estimate the marginal effects of chronic diseases using Generalized Linear Models, which control for the problems of high rates of non-response and skewness in the data.main conclusion of the paper is that chronic diseases are significantly associated with higher levels of health care expenditures and, to a lesser extent, with productivity losses reflected in lower labour supply and labour income. On the other hand, we cannot reject the hypothesis that households are able to insure non health consumption against chronic diseases. This leaves the question of how they manage to do so, since there is no evidence of increased transfers from formal systems of social insurance.

0112 The impact of cross-reference-pricing on pharmaceutical prices – manufacturers’ pricing strategies and price regulation Stargardt T. (Department of Health Care Management, Faculty of Economics and Management, Berlin, Germany) Schreyögg J. Objective: Several EU-countries are determining reimbursement prices of pharmaceuticals by cross-referencing to prices of foreign countries. Our objective is to quantify the cross-border spill-over effects of crossreference-pricing schemes on pharmaceutical prices in the former EU15 countries.

Methods: An analytical model was developed estimating the impact of pharmaceutical price changes in Germany on pharmaceutical prices in other countries in the former EU-15 using cross-reference-pricing. We differentiated between the direct impact (from referencing to Germany directly) and the indirect impact (from referencing to other countries that conduct their own cross-reference pricing schemes). Results: If the respective drug is marketed in all referenced countries a price reduction of € 1.00 in Germany will reduce maximum reimbursement prices in the Netherlands between € 0.27 and € 0.29. In Ireland, in addition to a direct impact between € 0.21 and € 0.23, referencing to the Netherlands expands the impact of a price reduction in Germany to a total impact between € 0.27 and € 0.29. In Austria, the German price reduction contributes to a total impact of € 0.15 a direct impact of € 0.07, while € 0.08 are due to referencing to the Netherlands and Ireland. In Italy the total impact is between € 0.33 and € 0.36. Conclusions: While these results might be favourable to the health care system in general, cross-border spill-over effects of policy interventions induce strong incentives for strategic product launches, lobbying activities and affects the efficiency of regulation as a whole.

0009 International comparison of formal and informal long term care Stevens J. (Social and Cultural Planning Office, Hauge, Netherlands) Pommer E., Gameren E., Woittiez I. We analyse the differences in the amounts of formal and informal longterm care ser vices given in nine European countries. Informal care is defined as care by relatives, friends and neighbours. Formal care is defined as paid or professional care. The ser vices considered include nursing, personal care and domestic help. analysis is based on the Survey of Health, Ageing and Retirement in Europe (SHARE). In 2004 about 23.000 persons aged 50 or older were interviewed. The countries included in the analysis are The Netherlands, France, Germany, Austria, Denmark, Sweden, Spain, Italy and Greece. focus on elderly with impairments. First we construct a measure of impairments based on Mokken scale analyses, including scales for physical and mental impairments. Cross-tabulations provide indications of complementarity between formal and informal domestic help and substitution between formal and informal nursing care. a multinomial logit model we explain the type of formal care received. Preliminary results show that after correcting for age, impairments, received informal care, availability of informal care and country-dummies, substitution between informal and formal care varies between countries and type of care. There remains a significant difference in the amount of formal care supplied implying cultural and institutional differences between countries.

0434 The role of indirect effectiveness and cost-effectiveness comparisons of health technologies in health policy Stoykova B. (Health Economics Research Group, Brunel University, Uxbridge, UK) Muniswamy S., Buxton M. Background: In HTA, indirect comparisons of effectiveness and costeffectiveness of technologies often need to be made, for example where two competing drugs have been each compared against placebo/best supportive care, or where two new technologies have been developed in parallel and each has been compared to current practice. Methods: In this paper we present a case study comparing the cost-effectiveness of alternative tests for risk stratification of patients with suspected coronary artery disease based on two RCTs. The first RCT studied myocardial perfusion imaging, and the second RCT studied stress echocardiography, in each case comparing them with the current standard Eur J Health Econom Suppl 1 · 2006

| S91

in the UK of exercise electrocardiography (ETT), which has unsatisfactory sensitivity and specificity. Results: Mean costs to diagnosis (95% CI) in the ETT arms of the trials were significantly different: £490.44 (453.8–527.1) vs. £231.84 (192.3–271.38). Whilst both trials were comparable in key aspects of their design, comparative analyses of data from the common technology indicated major differences (e.g. rate of coronary angiography) confounding the indirect comparison of results, and these indicated the need to adjust the comparison focussing on a common comparable subset of patients. Conclusion: Indirect comparisons of costs and incremental cost-effectiveness ratios of alternative technologies need to be appropriately adjusted. We recommend the use of appropriate modelling techniques as the best available method to inform policy.

0344 Pharmaceutical Patents: Incentives for R&D or Marketing? Straume O.R. (Department of Economics, University of Minho, Braga, Portugal) Brekke K.R. We analyse how a patent-holding pharmaceutical firm may strategically use advertising of existing drugs to affect R&D investments in new (differentiated) drugs, and thereby affect the probability distribution of future market structures in the industry. Within a fairly general model framework, we derive exact conditions for advertising and R&D being substitute strategies for the incumbent firm and show that it may overinvest in advertising to reduce the incentive for an entrant to invest in R&D, thereby reducing the probability of a new product on the market. In a more specific setting of informative advertising, we show that such overinvestment incentives are always present, and that more generous patent protection implies that a larger share of the patent rent is spent on marketing, relative to R&D.

0402 The different conceptions of the social cost of a risk factor: what is the theoretical and ethic background of each methodological option ? Béjean S. (Université de Bourgogne, Laboratoire d’Economie et de Gestion, Pole d’Economie et de Gestion, BP, Dijon, France) Taïeb H.S. Background: The evaluation of the social cost of a risk factor provides very valuable information for prevention policy decision-making. Several fields are concerned: road accidents, exposition to air pollution, exposition to risk factors at work. Objectives, assumptions: The objective is to produce a series of criteria that may help to classify surveys on social costs imputable to a risk factor according to the methodology used. Methodology: A review of literature of (1) surveys focused on methodological and ethic problems in cost-benefit analyses and (2) sur veys modeling the social cost of a specific risk factor. Results: Three types of criteria : The type of costs included in the sur vey (direct, indirect, intangible costs) depends on the point of view of the analysis (public health insurance system, enterprises, society as a whole). As regards indirect costs, the Friction costs theory (Koopmanschap 1992) raises debatable ethic questions. The value of lost lives imputable to the risk factor : is the price of an individual different during occupational life and retired life ? on what basis can we evaluate the monetary value of life (average or mean wage, per capita GDP) ? what are the different choices for the actualisation of the value of a lost life ? Has the delay between the occurrence of a risk factor and its consequences been taken into account in the sur vey ? The results of social cost evaluations can be very sensitive to long term health effect of risk factors (Chanel & Vergnaud, 2004).

0503 0508 Health: a vital investment for economic development and poverty reduction in Eastern Europe and Central Asia Suhrcke M. (World Health Organization, European Office for Investment for Health & Development, Venice, Italy) McKee M., Rocco L. The recent health record of several countries of Eastern Europe and Central Asia (ECA) is unfavourable by many standards, and some of the countries have seen life expectancy declining in recent years. In light of the relatively poor health status and given the apparent low policy priority on health in the region, this report seeks to demonstrate that by assigning a comparatively low importance to health, the countries are paying a prohibitive price in terms of foregone economic development. The report represents the first comprehensive effort to produce new evidence on the economic impact of (ill) health in ECA countries, both on the microeconomic (i.e. individual & household) as well as on the macroeconomic level. Results demonstrate that ill-health, in particular in the area of non-communicable disease and injuries, has left a significant negative impact on economic outcomes in ECA countries. Reducing the burden of disease is projected to contribute significantly to sustained future economic growth in the countries. We also analyse the role that factors beyond a narrow interpretation of “healthcare” have played (and could play) in promoting health in the region. In particular we quantitatively assess the role of social capital and of the quality of governance.

S92 |

Eur J Health Econom Suppl 1 · 2006

Divergent National Strategies for an EU-Health Care Policy and their Results Sundmacher T. (Universität Duisburg-Essen, Campus Duisburg, Fachbereich Betriebswirtschaftslehre, Lehrgebiet Mikroökonomik, Duisburg, Germany) Divergent National Strategies for an EU-Health Care Policy and their Results An increasing influence of EU activities on the national health care policy is perceptible. Indeed, this influence is still relatively low. For the future, different results are possible. The question arises, which strategies European governments will choose and which health care policy at EU level results from this. Some recognizable strategy patterns are: 1. Preser vation. Minimum conversion of the present European change impulses, stress to the national competences, prevention strategy against EU activities; 2. Erosion. Conversion of change impulses leads to rejections which make wider (national) reform steps necessary, ’fellowpassengers’ at EU level, muddling through “Instrumentalization. Forcing of change impulses to stimulate own reform plans, ‚attentionrelated’ choice of delegation; 3. Economization. EU-health care policy for internalization of cross-border external effects and for the exploitation of scale effects, cost-cutting delegation strategy, relevant especially for smaller nations; 4. Excludation. Delegation of responsibilities for the (the reelection endangering) health care systems to Brussels (’system agricultural policy’), forcing of EU competence, no special contents, delegation for the reason of delegation Within the scope of a game theory model of the EU decision-making process these strategies are used and examined under variations of the gaming rules. Besides, learning processes are also taken into consideration.

0504 High Innovation Dynamics and the 4th Hurdle – no Necessary Contradiction? Sundmacher T. (Universität Duisburg-Essen, Fachbereich Betriebswirtschaftslehre, Lehrgebiet Mikroökonomik, Duisburg, Germany) An economic evaluation of phar maceutical reimbursement of ten counts as a bureaucratic innovation brake. An essential reason for this appraisal lies in the transmission of (earlier) experiences of the central, state owned registration procedure of pharmaceuticals (the first three hurdles). For the 4th hurdle an alternative solution is introduced. In this model a decentralized competitive arrangement is carried out as far as possible. To do so, a market failure analysis is carried out at different levels (in example for markets for (innovative) pharmaceuticals, for insurance markets and for economic evaluation markets). The introduction of a proving period before starting the 4th hurdle attack is another essential element of the new institutional setting. In this proving period there exist incentives for the manufacturers to close the collecting of evaluation data as quickly as possible. Furthermore such a 4th hurdle requires a separation of different targets of ‚the pharmaceutical policy’. R&D policy, health policy or consumer protection policy targets are in example mostly not congruent. Hence, building up on a systematization instruments are discussed for targets not being addressed by the health policy orientated 4th hurdle. Besides, these instruments should affect as little as possible the reaching of health policy targets.

0401 Activity Based Hospital Funding in a Cash-Limited System – Qualitative Analysis of ‘Payment by Results’ in England Sussex J. (Office of Health Economics, London, UK) Farrar S. Aims: ‘Payment by Results’ (PbR) is a major change in the system of funding hospitals in the National Health Service in England, whereby hospitals are remunerated according to the type and volume of activity they undertake. The aim of the research reported here, which is part of the national evaluation funded by the Department of Health, is to examine the process of implementing PbR and analyse its effects on local health economies. PbR is intended to incentivise the system to meet specific policy objectives and is likely to have perverse as well as intended effects. Methods and Data: Interview-based, qualitative analysis. A purposive sample of senior local NHS managers was contacted and inter viewed confidentially. Inter viewees were from Foundation and non-Foundation NHS Hospital Trusts, Primary Care Trusts and Strategic Health Authorities. Semi-structured inter views sought managers’ knowledge and expectations about the impact of PbR. Results and Conclusions: The interviews raised many interesting issues – some confirming, but others contrasting with, theoretical hypotheses as to the incentive effects of PbR and its impacts on hospital activity, efficiency and competition. The quantitative element of the evaluation will be used to test the robustness of these initial findings, some of which will be returned to in future rounds of qualitative analysis.

0320 Local differences in family doctor recruitment and retention and patterns of relative rewards Sutton M. (Health Economics Research Unit of Aberdeen, Scotland) Elliott R.F., Gravelle H., Hole A.R., Ma A., Morris S., Sibbald B., Skatun D. Access is a key objective of many health care systems. This requires a distribution of relative rewards that equalises the net advantages to health

care professionals of locating in different areas. The existence of chronic problems in recruiting family doctors to particular areas in England suggests a failure in the existing national pay structure. We analyse spatial variation in family doctor recruitment and retention and relative rewards across local areas in England. The analysis uses a unique micro dataset comprising several indicators of family doctor recruitment and retention problems (including vacancy rates, vacancy durations and job satisfaction), family doctor earnings drawn from tax records, observations of private sector pay and job characteristics at each location. The difference between the pay of family doctors and that of comparator professions influences area variations in recruitment and retention, suggesting the national pay system does not adequately compensate for differences in the cost-of-living and the disamenities of locating in particular areas. Features of family doctor jobs that have no private sector counterpart also help explain local recruitment and retention difficulties. These factors should be reflected in the national pay system to equalise net advantages across areas and encourage the migration of family doctors.

0069 Quasi-markets with fixed prices – quality-competition in elderly care Svensson M. (IHE, The Swedish institute for health ec, Lund, Sweden) Edebalk P.G. ‘Consumer choice’ in elderly care is a system in which the user is free to choose one of at least two providers, one being run by the local authorities. The user, following an assessment of his/her needs, receives a voucher, which is a document stating that the chosen provider is entitled to compensation out of public funds. Prices in Sweden are fixed and equal for all providers. ‘Consumer choice’ is a multifaceted concept. The competition is dependent on how the consumer-choice system has been constructed. For instance, is there free entry to the market or not? Do the authority have a tendering process to approve the providers, or do the provider get authorization by the authority? There is a lack of knowledge as to how different consumer-choice models function, seen from a quality-oriented competitive perspective.study describes what competition with fixed prices in elderly care is about and how the providers act to compete. A central question is how the design of consumer-choice models influences the competition. The study is based on a sur vey in two municipalities in Sweden, with experience of different consumer choice systems in elderly care.

0020 Measuring the True Cost of Medical Procedures Szabó T. (Sanigest International, Bratislava, Slovakia) Cercone J., Száz K. One of the principal objectives of the current Slovak healthcare reform is to introduce a performance-based financing system. For this, it is necessary to measure the real costs of medical procedures. International experience in this respect includes individual costing per patient, average costs per hospital procedure and costs based on treatment protocol. the lack of data, none of these models could be implemented as-it-is. A new method has been proposed based on Activity Based Costing, applying an innovative allocation model that allows associating costs to different time episodes of a treatment, differentiating between fixed and variable costs, and costs related to service components. The model has been successfully tested in two Slovak hospitals for a total of six medical procedures.Slovakia, medical records contain detailed data on patient treatment; but only in paper form. Information was extracted by nurses into a pre-prepared data collection sheet. Sample size was pre-determined using a proxy for the expected cost deviation, which, ex-post, resulted to be surprisingly robust. A single database was established to analyze Eur J Health Econom Suppl 1 · 2006

| S93

the information and determine the estimated cost per procedure for each hospital.results show a cost variation of between 9 and 78 percent for the same procedure performed in different hospitals. The variation is attributed to a high degree of heterogeneity in treatment protocols. The ABC costing was used to demonstrate to Slovak hospital managers where efficiency gains are possible for individual procedures. The development of the allocation algorithm will allow for future replication of other procedures and in other facilities.

0458 Public Health Concerns on Irregular Migration in Europe Szilárd I. (International Organization for Migration, Brussels, Belgium) With the recent enlargement of the EU, Hungary, Poland and Slovakia now make-up the new EU Eastern border to Ukraine, Belarus, Romania (and Moldova) and provide an even more susceptible entry space for irregular migrants (both through smuggling and trafficking in person) than before. The composition of the irregular migrant flow, per country of origin, is very broad with migrants originating from Vietnam, China, Afghanistan, India, the Caucuses, Iraq, and Somalia etc. IOM estimates the volume of trafficking to the EU from Eastern Europe and the Balkans as high as 200 000 women and children a year. According to OSCE children make up more than 30% of all the trafficked victims. What is known in general about the migratory pattern and process shows that irregular migrants, either as independently having committed a criminal action (illegal border crossing), or as victims of the criminal act (trafficked persons) have, very likely during their routing/journey been exposed to a range of health-related problems, such as: poor living conditions – exposure to a wide range of diseases – physical violence – sexual exploitation and – psychological abuse. Providing appropriate health promotion and care services for trafficked persons is not only a humanitarian obligation, but also a public health concern for countries of origin, transit and destination alike. This problem does not merely appear in the context of spreading sexually transmitted infections (STIs) and ‘common’ infectious diseases, such as the – (re)-emerging problems of TB, – HIV/AIDS and of – Hepatitis B and C. Possible consequence of the demolished public health system in the majority of countries of origin: vaccine preventable diseases are spread to transit and destination countries where most physicians have not been confronted with these pathologies before. To achieve significant advances in this field, governments must harmonise their public health policies including border management and service provision, availability of specially trained practitioners, and data and information sharing.

0554 Brain-trade; the trends and costs of medical migration Eke E. (Semmelweis University, Health Services Management Training Centre, Budapest, Hungary) Sinko E., Szocska M. Aims: 1. Exploring brain-trade: what are the potential economic relations and consequences of brain-drain – general background – what factors and figures should be included and why? 2. What can and cannot be known about migration (meaning outflow) of Hungarian medical professionals since the EU accession of Hun-

S94 |

Eur J Health Econom Suppl 1 · 2006

gary? What do we know about migration tendencies? Introduction of an ongoing migration research project and its results. 3. What economic impacts can be estimated in relation with the currently seen and predictable migration trends among Hungarian health professionals? 4. What challenges health policy faces regarding migration? Methods: 1. Review of Literature 2. Consideration and exploration of relating economic factors 3. Analysis of official data regarding migration of young medical doctors 4. Introduction and analysis of own data regarding migration tendencies among medical students and young medical doctors 5. Analysis of potential economic relations of our results 6. How can/ should the findings be involved into health policy strategies Results: 1. Brain drain in health sector has been a long-lasting issue in the world, although it has attracted much attention in Hungary (and surrounding countries) since the EU accession of our country in May1 2004. The precise extent of migration of medical professionals and its consequences in health-care and health status of the population offer much to explore. The main challenge is to see the issue in an overall economic context. 2. Based on available literature, official data and own data we attempt to calculate the costs of migration of Hungarian health professionals, namely physicians. 3. According to our study and assumptions we would like to explore what health policy and economic aspects should be considered and what recommendations can be made for intervention. Conclusion: Migration and its aspects must be seen, explored and handled in the scope of overall economic context. Brain drain also emerges as a workforce market issue in the form of Brain Trade, so the economics of this trend should also be investigated.

0025 The Effect of Poverty on the Health of Newborn Children: Evidence from Germany Tamm M. (RWI Essen, Germany) A high risk of children living in poor families to attain poor outcomes in various fields is well documented. We provide new insights into health outcomes of newborn children in Germany and their relation to poverty status and try to identify pathways of the effect of poverty on health.data consists of 565 newborn children and reports several health outcomes (low birth weight, preterm birth, small height at birth, small head circumference at birth, and any confirmed disorders) as well as the number of doctor visits and visits to a hospital during the first three months after birth, which were due to health problems and not due to regular medical check-ups. The data is based on a completely new questionnaire supplementing the German Socio-Economic Panel (GSOEP) in 2003 and 2004.econometric implementation is based on complementary log-log models for the binary health outcomes and on hurdle and zero inflated negative binomial models for the number of doctor visits and visits to a hospital.findings indicate that there is no significant relation between poverty status and most health outcomes of newborn children. However, preterm birth is significantly more likely by more than 20 percentage points among poor mothers. Furthermore, mother_health status has a significant effect on a couple of aspects of child health, at least mother_health during pregnancy. I.e. there is clear indication of intergenerational transmission of health factors to newborn children.

0264 The Attributable Cost and Lenght of Hospital Stay of Central Line Associated Blood Stream Infection in Three Intensive Care Units and Infectious Diseases Department in a Teaching Clinic in Milan, Italy. A Prospective, Matched Analysis. Tarricone R. (Commerciale Luigi Bocconi – CERGAS (Centre for Research on Health and Social Care Management), Milan, Italy) Torbica A., Musigrave F., Franzetti F., Corona A., Raimondi F., Rosenthal V.D. The aim of the study was to estimate costs of CVC-associated BSI from a hospital perspective.2 -year prospective nested case-control study was undertaken at the three Intensive Care Units and Infectious Diseases Department in a teaching clinic in Milan. patients with BSI (cases) and 56 patients without BSI (controls) were matched for departments, length of stay (LOS), gender, age, and severity of illness score. LOS in the department was obtained prospectively on daily rounds. As to costs evaluation, they were estimated based upon the consumption of resources while in hospital. Unit costs have been calculated through a micro-costing approach.mean age of cases and controls was 64 and 70 years respectively.mean LOS of patients with and without BSI was 18 and 7.28 days respectively. The mean extra cost for overheads was € 7.722. Overall, the mean costs of patients with and without BSI were € 17.469 and € 8.279 respectively. The extra cost per case was € 9.191.the present study, patients with BSI had significant longer hospitalization, and greater cost. The present study supports the need for routine utilization of preventative interventions to reduce CVC-associated BSI as a means of reducing health care costs.

0132 The change in the doctor-nurse paradigm, some conditions and implications Tavares A.I. (University Autonoma of Barcelona, Spain) Doctor-nurse game was firstly identified by Leonard Stein in 1967. In this game, nurses were playing the role of doctor_third arm. Nurses weren_highly educated and their productivity was not very high either. In the 1990 Stein et al. went back to analyse this doctor-nurse game and noticed that nurses role had changed. They had become more educated, their productivity had increased and they had become more autonomous health professionals. We propose a formalization of this historical evolution using agency theory. We derive implications on the doctor-nurse incentives within a team and infere about the validity of some of the doctors and nurses claims.

0540 The first decade of case-mix systems in Italy Tedeschi P. (Bocconi University Department: Center for research in health care management, Milan, Italy) Case-mix systems have been introduced in the Italian national health system since 1994 after the American experience with DRGs and as part of an overall reform process which has decentralising decisional, organizational and operational power from State to Regions. While the challenges leading to a hospital prospective payment system were similar to other European Countries (e.g. increasing demand, raising costs), Italy adopted a strategy of regional adaptation of case-mixes and tariffs leading through time to increasing interregional differences. In particular the intervention will address the causes leading to the health policy shift towards DRGs, the goals, debates, results and controversies related to the health system evolution under case-mix systems. For sure the Italian case highlights that health financing based on case-mixes requires con-

trol tools (both internal and external to hospitals) in order to exploit the potential of activity related funding while managing the effects of partial economic conveniences (e.g. private hospital providing convenient care, but not necessarily complementary to public offer and needs). The future of case-mix systems in Italy appears to be connected with the capability of combining clinical, cost and quality in order to foster continuous improvement in hospital practice and organization.

0058 The Impact of Aging on Future Healthcare Expenditure Telser H. (University of Zurich, Switzerland) Steinmann L., Zweifel P. The impact of aging on healthcare expenditure (HCE) has been at the center of a prolonged debate. This paper purports to shed light on several issues. First, it presents new evidence on the relative importance of the two components of HCE that have been distinguished by Zweifel, Felder and Meier (1999), viz. the cost of morbidity and the cost of mortality (their red herring hypothesis claims that neglecting the mortality component results in excessive estimates of future growth of HCE). Second, it takes account of recent evidence suggesting that HCE does increase life expectancy, implying that time-to-death is an endogenous determinant of HCE. Third, it investigates the contribution of population aging to the future growth of HCE. For the case of Switzerland, it finds this contribution to be relatively small regardless of whether or not the cost of dying is accounted for, thus qualifying the red herring hypothesis.

0150 Analysis of migration impact on developed countries Testi A. (University of Genova, Department of Economics and Quantitative Methods, Genova, Italy) Ivaldi E., Tanfani E. In the last decade the movement of people to developed areas highlights new and emergent issues, i.e. how to respond to the health needs of migrants. Moreover, the official data about migration do not include the growing irregular or undocumented flows. A commonly used assumption is that irregular migration accounts for one-third to one-half of new entrants into developed countries, an increase of 20 per cent over the past ten years (IOM Report, 2003). In European countries growing attention is focused on understanding stategies, policies and programs aimed at facilitating social and economic integration of migrants in receiving areas. In this context a topical issue is to document the impact on access and use of health ser vices.reference to the Italian National Health Service the main objectives of this paper are i) to identify major problems related to the lack of homogeneous and reliable data collection, in particular related to the undocumented migrants; ii) to detect the main health services required by migrants (expressed demand); iii) to highlight the magnitude of resources needed to achieve a positive impact on the health of the migrants and to address the question of the best financing mechanism.

0486 Approaches to capital investment in the hospital sector in Europe Thompson C. (European Commission, DG SANCO, Luxembourg) McKee M. While much has been written about health care financing in Europe in recent years, discussion has almost entirely focused on revenue. In contrast, there has been remarkably little written on financing of capital investment in European health care systems. Yet major changes are Eur J Health Econom Suppl 1 · 2006

| S95

underway in several countries, in particular involving new forms of public-private partnerships (PPP). At the same time, there is growing recognition of the way in which the inherited structure of the health care delivery system constrains the system’s ability to adapt to changing circumstances. This presentation reports the results of a survey undertaken among key informants in the member states of the European Union to begin to ascertain existing practices and future plans in relation to hospital planning and financing amongst public and private not-forprofit hospitals. The locus of hospital planning decisions reflect the constitutional framework of the country involved, and thus the emphasis on national or local plans. There has been an expansion of private sector involvement, with four basic models identified: private loans direct to the hospital private loans to a regional health body a PPP where the private sector’s role is to build, design and operate the non-clinical functions of the hospital and, finally, a PPP, where the private sector’s involvement also includes management of the clinical functions of the hospital. It is too early to say whether these approaches will be more successful than the models they are replacing.

0033 Preferences for Financing Long-Term Care in Old Age Tipper A. (University of Aberdeen, Scotland) Coast J., Flynn T. Little is known about the public_preferences for financing their own long-term care (LTC) for when they are in old age. Therefore a discrete choice experiment was conducted with a random sample of the population to estimate utility values for the various attributes and associated levels of LTC financing. attributes for the experiment, developed from the literature on financing LTC, were risk over covering all expenditures, life span risk, care access, finance and quality of care. Two versions of the questionnaire were developed, the first using best worst scaling and the second using pairwise comparisons. questionnaires are being administered by post to a random smaple (n=2000) of the general public in early 2006 with the data analysis performed shortly afterwards. The results will begin to fill the gap in the literature about the decision making process that individuals pursue in financing their LTC. The utility values can be used to design a policy for financing LTC defined by the most preferential features of the experiment

abilities of a prolonged wait which indicates the presence of prioritization by physicians/hospitals. Cancellation of procedures/appointments is associated with a higher probability of a prolonged wait, particularly when the cancellation is initiated by the patient which indicates that the hospital/physician may prioritize those who experience cancellation for reasons beyond patient control. Experiencing barriers to access such as scheduling difficulties and availability also lead to a high probability of a prolonged wait. Other factors associated with prolonged waiting as a result of bottlenecks present in the system are the lack of a family physician and the requirement of an overnight hospital stay for a non-emergency surgical procedure.

0168 The use of Patient Reported Outcomes (PRO) and health economics for regulatory approval and reimbursement Tolley K. (Mapi Values, Macclesfield, UK) Market access for new pharmaceuticals, requires pharmaceutical companies to secure regulatory approval from the FDA (USA)/EMEA (Europe). Traditionally, this involved companies including clinical endpoints and pharmacovigilance data in phase III trials to show the efficacy/safety of their product against placebo. In recent years there has also been a move towards the use of patient reported outcome (PRO) endpoints (e.g. symptom scales, health related quality of life measures) that can be submitted to FDA/EMEA to support a regulatory claim. This has led to a need to ensure PRO measures are properly developed and validated. Beyond regulatory approval, market access for new pharmaceuticals requires reimbursement from national and local payers. Increasingly in Europe (and elsewhere) 4th hurdle agencies such as the Scottish Medicines Consortium, NICE, Committee for Pharmaceutical Help in the Netherlands are requiring evidence on cost-effectiveness/ budget impact to determine reimbursement status or reimbursement price. Health economic based arguments based on robust clinical trial evidence are requested by these agencies. The session will focus on comparing European and USA needs for PRO for regulatory approval and compare different countries needs for health economic data for drug reimbursement.

0559

0105

Rapid health technology assessment for pharmaceutical reimbursement and decision making in Europe

Determinants of Prolonged Waiting Times for Specialized Health Care Services in Canada

Tolley K. (Mapi Values, Cheshire, UK)

Tissaaratchy P. (Health Canada, Ottawa, Ontario, Canada)

Health economic evidence needs and the impact on market access of three influential European reimbursement agencies that appraise new drugs at launch: the Single Technology Appraisal (STA) of NICE in the UK, the Scottish Medicines Consortium in the UK and the Committee for Pharmaceutical Help in the Netherlands will be discussed. First an up-to-date status report of the STA process and differences from a full NICE appraisal will be presented, with case studies of the first drugs to undergo this process presented (i.e. docetaxal, paclitaxel and trastuzumab for early breast cancer). Thereafter, a brief comparison of the process of decision-making involved with each agency will be presented. Also the critical pharmacoeconomic evidence needs of these agencies, with a focus on the specific problems of choice of the appropriate comparator, the measurement of health outcomes etc. will be considered. This session will contain an interactive exercise to review case study examples of products that have been appraised (and for which information is publicly available). The impact of each agency on new product reimbursement and market access within the UK and The Netherlands, and more widely in Europe, will be discussed. Some recommendations on best practice in optimising pharmaceutical company submissions will be provided.

The objective of this study is to explore, using recent Canadian patient sur vey data for 2003 available in the Health Ser vices Access Sur vey (source Statistics Canada), determinants of prolonged waiting times for specialized non-emergency ser vices, namely surgical, specialist, and diagnostic. The logistic regression models presented in this study tell us that willingness to pay by patients for services for which private sector options exist, prioritization of procedures by physicians, and cancellations by both physicians and patients impact the length of waiting. This paper adds to the existing body of analytical literature by providing some insights into the impact of patient and physician behaviour on prolonged waiting. results indicate that the behaviour of patients and physicians does influence the length of the wait for a ser vice. For specialized ser vices for which private sector options exist, patients in the highest income quartile have the lowest probability of experiencing a prolonged wait indicating a higher degree of willingness to pay by such patients. For non-emergency diagnostic scans, patients obtaining scans in public institutions have a higher probability of experiencing a prolonged wait. For all specialized ser vices, diagnostic conditions associated with higher mortality rates are associated with the lowest prob-

S96 |

Eur J Health Econom Suppl 1 · 2006

0616 What does industry think about a “4th hurdle”? Towse A. (Office of Health Economics, London, UK) Issues to be discussed: 1. Role of economic evaluation is identifying value in different types of pricing and reimbursement systems. Is reference pricing an alternative to economic assessment? Who sets the price and should there be an explicit willingness to pay for a QALY? 2. Availability of information and handling uncertainty. Payers don’t want to reimburse products that are not cost-effective, but don’t have all of the information at launch. What is an efficient way to resolve this dilemma? How is new information best handled? 3. Coverage of HTA. Is is “all new medicines?” Is that the best use of HTA resources? 4. What is a “good practice” process for an HTA institutional assessment of value? For example, what role do manufacturer’s submissions have? Should there be an appeal process? And patient group involvement? 5. Extent of pan- European sharing of assessments. What is efficient and not efficient? Can Europe’s HTA bodies pool skills and effort or are costs, effectiveness, and willingness to pay too different across European countries?

0271 Programme Budgeting and Marginal Analysis (PBMA): a case study of the Greek context Tsourapas A. (Health Economics Facility, Health Services Management Centre, University of Birmingham, United Kingdom) Frew E. Background: PBMA is a systematic, explicit priority setting toolkit, aiming to assist decision makers in identifying the most efficient use of resources. It can facilitate priority setting by encouraging participants to identify decision-making criteria and combine these with the best available evidence. To date this technique has not been used in Greece. Methods: Literature review and tape-recorded semi-structured interviews with senior medical and government decision makers from Greece, to determine the acceptability of PBMA in this context. Results: The literature demonstrates that for the PBMA framework to work well it is important to have active involvement, commitment and cooperation from all participants, especially senior members of an organisation. Although involvement and commitment were apparent among some interviewees as a means of obtaining more transparent priority setting, lack of co-operation between hospital departments would make application of PBMA difficult. Furthermore, recent reforms introducing regional health authorities (RHA) and future reforms introducing ‘local budgets’ in the RHAs could facilitate the adoption of PBMA, but time is needed for the RHA to gear to the new situation that ‘local budgets’ will set. Conclusion: PBMA may enhance decision-making in Greece, but the cultural context necessary for success does not currently seem to be present.

0145 Health status in adulthood according to family and social background : intergenerational transmission of income-related health inequalities Tubeuf S. (IRDES, Paris, France) Trannoy A., Rochaix L., Bouvard L. Objectives: The objective of this paper is to explore the hypothesis of an intergenerational transmission of health inequalities on French data, particularly in terms of equal opportunities.

Summary: Research into the determinants of inequalities (both in terms of health status and health care utilisation) has shown the importance of income and social characteristics (measured by social class and education levels) – Wagstaff et al. (2001), Kunst et al. (2000), Van Doorslaer and Koolman (2005)–. Another line of research into the determinants of income-related health inequalities has considered the whole life cycle. Wadsworth (1986) has shown the influence of health in childhood on career prospects and the impact of living conditions in childhood, even in utero, on health status in adulthood (Wadsworth (1999)). This analysis according to which family and social backgrounds in childhood strongly influence health in adulthood has not yet been carried out on individual French data. In this study, we explore the hypothesis of an intergenerational transmission of health inequalities. First, we build distributions of health status conditional on the family and social background, as measured by parents’ socioeconomic status. We then compare these conditional distributions of health in order to appreciate differences between the French situation and an ideal of equal opportunity. We apply the Lefranc-Pistolesi-Trannoy (2004) methodology based on nonparametric analyses of the distributions of health in stochastic term of predominance to the 2004 Sur vey on Health, Ageing and Retirement in Europe –SHARE. We then compare the French situation in terms unequal opportunities for health to the 11 other Europeans countries in the survey.

0277 Reforms of the National Healthcare System and Dynamics of the Hospital Efficiency Turati G. (University of Turin, Faculty of Economics, Torino, Italy) Piacenza M., Vannoni D. In this paper we estimate cost efficiency and technological characteristics of public hospitals in Piedmont – a Region of the Northern Italy – over the period from 2000 to 2004. During the Nineties, different regulatory reforms aiming at controlling health expenditure affected the Italian NHS, ranging from the introduction of a new reimbursement scheme for hospitals (DRG-based mechanism) to the implementation of fiscal federalism. We argue that all these different reforms rendered producers more prone to control costs, in order to be able to meet financial constraints. Our empirical findings support this hypothesis: average cost inefficiency scores obtained estimating a Translog stochastic cost frontier by the Battese and Coelli (1992) methodology show a decrease from 2000 to 2004, while the adoption of more effective technologies shifted upward the cost frontier. We also find significant economies of scale not exploited by producers.

0046 A project portfolio approach to prioritize innovative technology investments in healthcare organizations Turra E. (APSS, Trento, Italy) Germana H.R. In healthcare organizations and hospitals there is an increasing demand for utilizing healthcare and information technologies in the management of change and innovation. Faced with the complexity of that effort and the inflexibility of the capital budgets, combined with the ongoing operational needs, a systematic approach has been adopted by the Azienda Provinciale per i Servizi Sanitari (APSS) in Trento for the selective funding of technology enabled innovation activities. For this, APSS established project portfolio management capabilities to assist management in making the decision on how to allocate the discretionary side of the budgetary resources : that is, the portion of the budget not already committed to previous decisions. To guide the selection process, a formal review is periodically performed to analyise and selecting those projects, among the several ones which are proposed by the various departEur J Health Econom Suppl 1 · 2006

| S97

ments, each competing for granting a share of the capital budget. The goal is to discover those proposed innovation initiatives which makes a greater contribution to the company objectives based on explicit organization priorities and evalutation criteria. This is also to avoid that existing committed programs, backed by existing internal constituencies : would be favoured over value-adding and innovative technology initiatives, unprotected by committed stakeholders.

0129 Health Fund Competition in Germany Uebelmesser S. (CES, University of Munich, Germany) Becker B. In the 1990s, competition among public health funds was introduced in Germany. Even though legally forbidden, there is anecdotal evidence that funds engage in activities to selectively recruit enrolees. Empirically, it is difficult to separate direct risk selection of sickness funds from switching which – due to switching costs – might be biased towards the young and healthy. In this paper, we want to directly address this question by focusing on the advertisement strategy of sickness funds. So far, this has not been possible due to lack of data. We use a new dataset of all ads placed by public sickness funds in all German newspapers and magazines between 1990 and 2003 as well as manually collected ads placed in the German magazine Stern between 1992 and 2003. The analysis of the advertisement strategies provides us with evidence to draw first conclusions about the effects of more competition. Introducing more competition was intended to increase the efficiency of the health funds thus leading to lower contribution rates. We find that the advertisement activities in general as well as the content of the ads placed in the Stern magazine are affected by the most relevant institutional changes. Our content analysis, however, seems to indicate that the changed content cannot be taken as clear evidence of risk selection. This points towards a (more) important role played by the decision of the sickness funds about where to advertise for an analysis of risk selection.

0155 Qualifying the Relationship between Health Care Expenditure and Health Ulmann P. (CNAM, Paris, France) Nixon J. Previous empirical research by the authors confirmed that, for the European Union, increases in health care expenditure are significantly associated with large improvements in infant mortality, but only marginally for life expectancy. The findings, however, were attached to a number of caveats concerned with model misspecification issues and data availability, consistent with many previous studies in this field.The aim of the present paper is to report the findings of updated and more reliable analyses using the latest OECD health data, augmented by explanatory variables derived from other macroeconomic datasets. The study was extended to capture the influence of a) health care system typology, b) share of public health expenditure, c) share of primary care expenditures, d) primary prevention, e) educational care expenditures, f) the characteristics of health care financing. Finally, the diminishing influence of health expenditures on health in developed countries was also evaluated. This is work in progress and detailed results are not yet available. The results, however, will increase the validity of our previous findings and are expected to confirm a statistically significant association between health care expenditure and health outcomes, and enhance our understanding of the influence of other explanatory variables that explain this relationship.

S98 |

Eur J Health Econom Suppl 1 · 2006

0158 Comparing lifetime health care costs of major risk factors: smoking versus obesity Van Baal P. (RIVM, Bilthoven, Nederlands) Feenstra T., Polder J., Boshuizen H., Hoogenveen R. Estimates of health expenditure attributable to risk factors are usually based on cross sectional analyses and do not take into account health care costs of competing diseases that might result from changes in lifestyle. Using a dynamic population model, we calculated incidences, prevalences and health care costs of some major chronic diseases conditional on risk factors. We compared lifetime healthcare costs of a cohort smokers and a cohort of obese people to those of a cohort of ‘healthy’ people. Average disease costs per patient were linked with disease prevalence’s estimated with the model. At all ages, average healthcare costs were lowest for ‘healthy’ people. Until age 65, average healthcare costs were highest for obese people. At older ages, smokers were most expensive. However, since healthcare costs increased rapidly at older ages and life expectancy is longest for healthy people and shortest for smokers, total lifetime costs were highest for the cohort of healthy people and lowest for the cohort of smokers. Relating the difference in life expectancy to the difference in healthcare costs reveals that healthcare costs per life year gained are lower for prevention of obesity (€1,000) than for prevention of smoking (€ 5,000).

0567 Organization and Financing of ESRD Care in a Mixed Universal Insurance Van Biesen W. (Renal division, Dept of Internal Medicine, University Hospital Ghent, Belgium) Lameire N., Vanholder R. Belgian health care is organized as a mixed private/public system: compulsory contributions and reimbursements are organized by the state, but the providers are private. There is a fee-for-ser vice system with fixed reimbursement rates for each medical inter vention. There used to be only a few restrictions on the services provided, but this is rapidly changing due to the increasing costs of health care. For ESRD, the fee-for-service system discourages the development of preventive strategies, early referral to the nephrology unit, and the use of home based therapies like peritoneal dialysis. The aging of the general population is reflected in the rapidly increasing number of very old dialysis patients who tend to have multiple infirmities requiring more complicated care. Although the current system delivers an acceptable cost-benefit ratio with excellent standards of care, one may question whether unrestricted access will remain sustainable in view of the high cost of dialysis. If restrictions on access become necessary, how they should be developed and by whom will be the pertinent questions. Nephrologists should be aware of financial and organizational issues to raise their voices in this discussion.

0099 Universal mandatory private health insurance in the Netherlands Van de Ven W. (Erasmus University, Rotterdam, Netherlands) Prinsze F.J., Bruijn D., Schut F.T. Since 1 January 2006 all citizens in the Netherlands are obliged to buy private health insurance. There will be an annual open enrolment, community-rated premiums per insurer, and a risk-equalization system based on age, gender, disability, diagnostic cost groups, and pharmacy-

based cost groups.this paper we focus on supplementary health insurance as a tool for risk selection. For supplementary insurance the insurers are free in premium setting and underwriting practices. If an insurer rejects an applicant for the supplementary health insurance, this consumer most likely will buy the mandatory basic health insurance from another insurer. results indicate that there is a clear similarity between the questions in the health questionnaires that insurers use for applicants of supplementary health insurance and the subgroups of consumers who generate a predictable loss for the basic insurance. By the answers in the health questionnaires insurers can easily divide their applicants in two groups about one third with an average predictable loss for the basic insurance of about 800 euro per year and the others with an average predictable profit of about 350 euro per year. Therefore risk selection via the supplementary health insurance is a very easy and profitable strategy.

0107 Using open-ended follow-up questions to test the validity of DCE estimates Van der Pol M. (University of Aberdeen, UK) Shiell A., Au F., Tough S., Johnston D. The Discrete Choice Experiment (DCE) is becoming increasingly popular within health economics. However, concerns have been raised over the validity of the estimates produced by a DCE. This is investigated in this paper by testing the consistency between the estimates obtained from a DCE with those directly obtained using open-ended questions’s preferences for community perinatal care were elicited. Face to face interviews were held with 292 women. Consistency between the methods is explored by analysing the DCE data according to the optimal scenario identified in the open-ended question. levels of consistency were achieved when examining whether participants chose the package in the DCE identified as their optimal package in the direct open-ended question. However, the mean WTP derived from the open-ended CV question is substantially lower than the mean derived from the DCE. The results did suggest that individuals who are willing to pay relatively more in the open-ended CV question were also willing to make relatively larger trade-offs in the DCE. Research into the different biases that are present and may cause the difference in the WTP estimates continues to be an important line of enquiry.

0108 Is risk attitude in health context specific? Van der Pol M. (University of Aberdeen, UK) Ruggeri M. Introductionis an ongoing debate about which health valuation method is superior. One of the arguments that is used in favour of the Standard Gamble (SG) methods is that it incorporates individuals’ preferences for risk. This is seen as desirable because most medical decisions are made under conditions of certainty. However, most uncertainty is not of the ‘life or death’ form as present in the SG. The argument therefore only holds if individuals’ risk attitude is constant and does not depend on the kind of health risk. There is evidence in the psychology literature that risk attitude varies across different domains and situations. This has not been robustly tested using health outcomes. aim of this study is to examine whether individuals’ risk attitude for fatal changes in health differ from their risk attitude for non-fatal changes in health. ‘risk attitude is measured through the use of SG exercises. Two types of health outcome are used fatal changes (0 versus 5 years; and 5 versus 15 years) and nonfatal changes (amount of time spent in ill-health). In total, 102 university students filled out the web-based questionnaire.collection has just been completed. Results will be available soon.

0312 Care for a break? An investigation of informal caregivers’ attitudes toward respite care Van Exel J. (Erasmus MC, Institute for Medical Technology Assessment (iMTA) & Department of Health Policy and Management (iBMG), Rotterdam, Netherlands) Brouwer W.B.F. Public awareness is growing that informal caregivers should be supported in their valuable and indispensable effort to provide care to someone in their social environment. Respite care was shown be an effective way to alleviate caregiver burden and to reduce the adverse (health) effects from substantial burden. Yet, little is known about caregivers’ needs and desires regarding respite care. conducted a Q-methodological study and found three distinct groups of caregivers informal caregivers who “need and ask for respite care”, those who “need but won’t ask for respite care”, and those who “do not need respite care”. Caregivers in the first two groups experience substantial burden, while those in the third group enjoy sufficient support and appear to manage pretty well. Caregivers in the second and third group derive considerable “process utility” from care giving. On balance, caregivers in the first two groups would sometimes rather have someone else take over their task, but for caregivers in the second group this desire is frustrated by care recipient resistance against respite. respite care funds should be targeted at caregivers in the first two groups, while in the second group effort should be directed to both caregiver and care recipient.

0467 The demand and use for support and respite care for informal caregivers Van Exel J. (Erasmus MC, Institute for Medical Technology Assessment (iMTA) & Department of Health Policy and Management (iBMG), Rotterdam, Netherlands) We analysed the preferences for and use of support and respite care in a sample of informal caregivers. Almost 80% of the respondents desire support or respite care in general, some time off is preferred by 40% of the respondents. The results show that caregiver characteristics, care recipient characteristics, elements of the caregiving situation and institutional variables determine the desire for support and respite care. Especially the subjective burden of caregiving is important, whereas the number of caregiving tasks and the time invested (objective burden) hardly affect the desire for support and respite care. Caregivers are reasonably well informed about existing respite services and about one third of caregivers made use of respite care. These caregivers experienced substantial burden of caregiving. Most respite care facilities reach the caregivers most in need of support, but not all caregivers in need make use of respite care. More research is needed into the cost-effectiveness of respite care. References Koopmanschap MA, van Exel NJA, van den Bos GAM, van den Berg B, Brouwer WBF. The desire for support and respite care: preferences of Dutch informal caregivers. Health Policy 2004;68:309-20. Van Exel J, Moree M, Koopmanschap M, Schreuder Goedheijt T, Brouwer W. Respite care: an explorative study of demand and use in Dutch informal caregivers. Health Policy (in press)

0078 The length of waiting lists for long-term care in the Netherlands Van Gameren E. (Social and Cultural Planning Office, Netherlands) Since many years, the long-term care sector in the Netherlands has been characterised by waiting lists. Regular counts have been held of the number of people waiting for the various services, ranging from home care Eur J Health Econom Suppl 1 · 2006

| S99

to admission in a nursing home. In this paper the length of the waiting lists in 32 regions (per inhabitant aged 75 or more) in October 2003 is confronted with supply indicators such as the capacity of nursing homes and availability of personnel in a region, demand indicators as the numbers of elderly above age 85 and of single-living elderly, and indicators of the organisation of the care in the region. Demand and supply factors are generally positively correlated, indicating that a higher demand goes with a higher supply. Supply factors generally negatively correlate with the length of the waiting lists, more supply gives shorter waiting lists. For some of the demand factors a (wrong) negative sign is found. It remains complex to determine the driving forces behind the waiting lists. The most apparent outcome is that the waiting lists for the various services are highly interconnected, because some of the services are given as a temporary care solution to people waiting for other services.

0137 Effects of risk equalization on premium rebates for voluntary deductibles in social health insurance Van Kleef R. (Erasmus University Rotterdam, Netherlands) Beck K., Ven W. P.M.M, Vliet R. C.J.A. Without risk equalization among insurers, premium rebates for voluntary deductibles in health insurance can be very high due to self-selection. With risk equalization, the effect of self-selection is smaller since risk equalization (partly) corrects for differences in risk between the consumers who choose a deductible (the relatively healthy) and those who do not (the relatively unhealthy). As a result, these differences cannot be (fully) reflected in the premium rebate. study, which is based on data from a Swiss sickness fund, shows that without risk equalization the premium rebate for a voluntary deductible of SFr. 1,270 can (theoretically) be 86 percent relative to the actuarially fair premium for full coverage. With risk equalization based on age, gender, region and medical history, the maximum premium rebate drops to 40 percent. However, this premium rebate still exceeds the deductible amount (SFr. 1,270). So, even with comprehensive risk equalization, there still is a substantial effect of selfselection on the premium rebate, resulting in non-optimal risk solidarity. Including the level of deductible as a risk factor in risk equalization could optimize risk solidarity, but would also lead to lower rebates and consequently to a smaller number of consumers choosing a deductible.

0028 The social costs and benefits of prior authorization Varkevisser M. (Erasmus University Rotterdam / ECRi, Netherlands) Dijkgraaf E., Van der Geest S. In the Netherlands reimbursement of new, costly medications aimed at treating particular medical conditions is restricted. For some of these drugs prior authorization (PA) is compulsory, while for others health insurers are free to impose PA. To control drug costs, patients increasingly have to seek approval from their insurer. We have identified the potential social costs and benefits associated with PA for “special” drugs. Economic rational insurers use PA when private benefits exceed private costs. PA, however, is a labor intensive administrative procedure that raises costs for all affected parties. It may also reduce patients’ access to beneficial drugs. Especially when the restrictions for reimbursement that are imposed by the government lag behind clinical research. In itself PA may be useful to encourage appropriate use of costly drugs. Because its substantial administrative costs and its potential adverse effects on health outcomes, however, the consequences for all affected parties (both health insurers, physicians and patients) should be taken into account. It is unlikely that insurers will act in this way out of their own accord. We argue that within the current Dutch system it is therefore not ensured that PA will only be used when social benefits outweigh social costs.

S100 |

Eur J Health Econom Suppl 1 · 2006

0029 Patients’ hospital choice in the Netherlands Varkevisser M. (Erasmus University Rotterdam / iBMG, Netherlands) Schut E. We empirically examine patients’ hospital choice in the Netherlands, using a unique dataset with individual patient level hospital utilisation data for the year 2003 from a large Dutch health insurer. If the propensity to travel differs between different types of patients, then the use of current patient migration data to define geographic hospital markets, as commonly applied in antitrust cases, may lead to incorrect conclusions. First, we analyse patients’ decisions to bypass the nearest hospital. We show that despite the absence of both financial incentives and observable quality differences between hospitals a sizeable percentage of patients was not admitted to the nearest hospital. Second, we estimate a logit model of individual patient hospital choice. Our approach focuses on the substitutability of hospitals and is based on a random utility model in which the indirect utility of a patient who visits a specific hospital is a function of personal attributes, hospital attributes and attributes specific to a given patient-hospital combination (e.g. travel time). Our preliminary estimation results indicate that although travel time and hospital waiting time performance are important explanatory variables much of the travel for hospital care occurs for idiosyncratic reasons such as perceived quality differences or personal preferences.

0194 Cost-effectiveness analysis of the screening of hereditary hemocromatosis: How many alternatives are possible? Vicente A. (Servicio Aragonés de la Salud, Aragón, Spain) Garuz R., Misiego A., Ortega M. Introduction: Hereditary Hemochromatosis (HH) is a genetic disease with an autosomic recessive pattern that causes an excessive gastrointestinal absorption of iron, especially hepatic. Prevalence of the disease in Spain is between 1/1000 and 3/1000 but it is still not well defined, as well as its genetic expressivity. Early detection is basic to stop the development of an irreversible disease. Screening is performed by means of seric iron studies: Transferrin Saturation Index (TSI) and plasmatic ferritin. If high values are found, a confirmation genetic test is performed (not requiring liver biopsy). We have undertaken a cost-effectiveness test to determine the best screening strategy: Opportunist screening of men > 30 years or screening of close relatives of HH patients (siblings & offspring). Material and methods: The utilized costs have been the direct; Genetic test, analytical determinations, the phlebotomy and the HH’s complications, specially give them cirrhosis. They have been the measure of effect utilized the years of life earned (AVG) and to estimate the life expectancy in the general population in people and with HH, the same way that the benefits utilized Markov’s models themselves. The data of mortality come from the I.N.E and the suppositions of the model (the HH’s prevalence, probability of development HH) are extracted of the international literature. The adopted perspective matched her social and the annual rate of discount of the 5 %. Results: The best choice cost cash the strategy went from screening in brothers of people affected of HH (with an incremental cost of 2,281 euros for AVG). Screening’s strategy in children of people affected of HH (with an incremental cost of 18,838 euros for AVG). Screening’s strategy oportunistic in males of over 30 years (with an incremental cost of 23,794 euros for AVG). The results of the sensitivity analysis demonstrate consistency in a general way for all of the intervals of moral values of the assumed variables. Conclusion: The screening to detect homozygotes for mutation c282y of the gene HFE in first-degree relatives (brothers) with HH is the best choice cost cash. Our recommendation is of expectant prudence to the

hour to implement screening’s programs of the HH in males of general population without symptoms in our National Health System.

0296 Dynamic behavior of diagnostic imaging services providers under a fee-for-service reimbursement Vieira M. (HPA-Hospital Particular de Almada, Lisboa, Portugal) Barros P.P. We aim at identifying both supply and demand side determinants of expenditures growth with diagnostic imaging ser vices. In the Portuguese market, prices are exogenously fixed under a fee-for-ser vice regime, therefore private providers compete in other dimensions, and trends towards higher concentration in the supply side could be associated with higher growth in total demand. Our data is consistent with the common perception of a general trend in expenditures increasing with medical technology, due to a growing utilization of diagnostic technology, mostly linked to a rising share of more costly technology. The increase in annual invoice amounts is not evenly distributed across the sample of providers observed. Classifying providers according to their annual production, show us that those within intermediate classes have bigger annual positive variation of production, suggesting that they are the main drivers of expense increase from the supply side. Providers dynamic revealed towards a mix of increase of scale and scope is the result of a survival requisite derived from the competition among diagnostic services companies. The distinction between demand and supply side determinants is crucial to assess whether we witness increased satisfaction of unmet needs, or whether some sort of induced demand for these services exist.

0057 Is there a threshold effect in the relationship between income and health ? Vuilleumier M. (Institut de Recherches Economiques, Neuchâtel, Switzerland) Many empirical studies have investigated the relationship between income and health. They found a significant logarithmic association which reflects the effect of decreasing marginal returns of materialistic resources on health. However, the income and health relationship can also be seen as a succession of different linear regimes separated by thresholds. In that case, we observe a linear effect of income on health in each regimes, but with a decreasing effect in higher regimes. making the assumption that the income and health association can be divided into two different regimes, the aim of this study is to detect an empirical threshold on Swiss data using Quandt methodology. This methodology proposes a maximum likelihood procedure to detect the switching point into a linear regression system obeying two separate regimes. We also compute a likelihood function of a recursive regression model which depends only on the threshold. Finally, we derive a confidence interval using bootstrap simulation to test robustness of the threshold detected. use the “Self Assessed Health” (SAH) and an equivalised measure of household income. Data are grouped into ninety income categories of same length and we estimate the health mean in each category. Results: We find that the threshold which proposed the best adjustment of data is the 2900-3000 CHF income category. Effect of income on health is clearly higher below the threshold because of the significant difference between the income coefficients of each regimes. However, the result of the bootstrap estimation show that the threshold is not robust. The confidence interval is too large for claiming that there is a threshold effect in the relationship between income and health.)

0223 Persons Without Health Insurance in Germany: Status Quo and Policy Options Walendzik A. (Institute for Health Care Management, University Duisburg/Essen, Germany) Greß S., Wasem J. Persons Without Health Insurance in Germany : Status Quo and Policy OptionsGermany more than 99 percent of the population is covered either by social health insurance or by alternative private health insurance. Nevertheless, since the early 1990s, structural changes on the labour market have led to a larger share of low-income self-employed persons and more low-income employees. At the same time, the legislator has restricted access to social health insurance. As consequence, the number of the uninsured has doubled to almost 200,000 persons in 2003. By using representative census data we were able to identify groups of persons who run a higher than average risk to become uninsured. These groups consist mainly of low-income self-employed and low-income employees, persons who got divorced, university students and foreigners. Available evidence from the US shows quite clearly that the uninsured have a higher risk to become severely ill or to die prematurely than persons with health insurance. We conclude that this can be avoided by introducing mandatory health insurance for the entire population. The same has been done recently in the Netherlands and ten years ago in Switzerland. Alternatively we suggest to improve access in social health insurance as well as in alternative private health insurance.

0279 Target payment and General Practitioners’ Service Activity in the UK NHS Wang Q. (University of York, UK) Maynard A. Background: In order to improve GPs performance, the target payment policy was introduced by the UK NHS in 1990. Target payment remunerates health care service providers only if they provide an agreed level of care. Since 1990, there are two levels of target for childhood immunisation (lower target 70% and higher target 90%) and cervical cytology (50% and 80%). These services were previously paid by fee-for-service (FFS) method. Under target payment method GP service activities (ST) were compared with those under FFS method (SF) when other factors that influence ST and SF were controlled. Hypotheses: ST will be increased to target levels; ST distribution will be bimodal while SF normal; Other factors will have lower influence on ST; ST will be increased more prominently for GPs who were previously with lower service activity than those with higher one. Data/Methods: Data at national and local level (e.g. general practice) were collected to test the hypotheses by using time series, panel data, and cross-sectional analyses. Results: ST has been increased to target levels; ST has not formed a bimodal distribution; The influence of other factors was not lower on ST; The service activity has been increased more prominently for GPs who were previously with lower service activity.

0612 Translating measurement into a national agenda; the first Dutch health care report, its impact and what comes next Westert G. (National Institute for Health and the Environment (RIVM) and Tilburg University, Bilthoven, Netherlands) Objective/Aims: The aim of the presentation is to report on the development of a national performance indicator framework for the Dutch Eur J Health Econom Suppl 1 · 2006

| S101

health system and to present findings from the first Dutch healthcare report. Method: The Dutch Ministry of Health initiated a project to assess and report the performance of healthcare quality, accessibility and cost at the national level in the Netherlands. RIVM prepared the report in two stages: 1 defining a performance indicator framework, in collaboration with OECD; 2 finding appropriate measures to report about quality, accessibility and cost. International and time comparisons were made were possible. Result: In general the Netherlands has a accessible health care system in 2004. Health care expenditures has risen substantially between 2000 and 2004, but in line with the trend in the EU-15. Holland ranks little above average in cost development in the EU-15 and OECD countries. Quality of health care is for most indicators above average, compared to other OECD countries. Capacity to improve is observed for sub-indicators in the following indicator domains: effectiveness of prevention, cure and care, patient safety and continuity of care. Conclusion/discussion: The first Dutch health care report has some strong and weak points. Strong is that the performance framework used has a solid international fundament (Arah et al, 2006). The set of indicators used is firmly rooted in present Dutch health policy. Measures and indicators not always match perfectly. Time and space comparisons were limited, because of (international) data availability and comparability. The Dutch Ministry of Health will use the report (after publication in May 2006) to take the pulse of Dutch health care in 2004 and to steer it’s future direction. A quest for quality is likely to emerge following it’s publication.

0333 Cost Effectiveness Analysis (CEA) and the local-level technology coverage decisions in England Williams I. (Health Services Management Centre, University of Birmingham, UK) Stirling B. In a context of rapid technological advances in healthcare and increasing demand for expensive treatments, local formulary committees are a key player in the management of scarce resources. However little is known about the information and processes these use when deciding on the inclusion of new treatments. We collated and analysed new technology request forms’ in order to gauge the types of information routinely requested by such committees. We then conducted case studies of four committees involving obser vation of meetings and inter views with committee members. In particular we sought to investigate the role and impact of CEA on committee determinations. We found that only a small minority of committees routinely request information on the cost effectiveness of a new technology and the case study committees rarely accessed such studies. A range of barriers to use of CEA were identified including lack of health economics expertise within the committee, doubts over the independence of analyses, and problems with implementation of study recommendations. There is a need to clarify the aims and functions of local level committees, and to consolidate their place within the broader political and financial context, if routine use of cost effectiveness analysis is to become a realistic aim.

0052 An estimate of the worldwide costs of dementia: results and methodological aspects Wimo A. (Neurotec, Harmanger, Sweden) Jönsson L., Winblad B. Dementia disorders are today considered to be a major cost driver in the health care and social systems not only in the advanced economies

S102 |

Eur J Health Econom Suppl 1 · 2006

but also in the developing countries. The main reason is the aging population worldwide and as a consequence, the increasing number of people suffering from dementia, in conjunction with the high level of care required by demented elderly.estimate of the worldwide costs of dementia is based on worldwide prevalence figures of dementia from different regions combined with cost of illness studies of dementia in key countries. For countries with no available cost data, costs were modelled based on the relationship between cost of illness and the gross domestic product (GDP). Two major cost components need to be included in such a cost model a) formal care costs in terms of direct medical costs and direct non medical costs and b) costs of informal care. the base option, the worldwide direct costs of dementia are estimated at 156 billion US$ and the costs of informal care to 92 billion US$, resulting in a total cost of 248 billion US$. of the assumptions include both uncertainties and also variability. In a sensitivity analysis different scenarios are investigated, resulting in a range of cost estimates. Incomplete demographic statistics, different ways to organize care as well as its financing, available resources, and particularly the way of measuring and costing informal care are examples of factors that contribute to the uncertainty. This study demonstrates that the worldwide costs are substantial, and particularly the expected increase of elderly people in the developing countries present a great challenge for social and health care systems.

0183 Paying For Long-Term Care for Older People in the United Kingdom Wittenberg R. (Personal Social Services Research Unit, London School of Economics, London, UK) Hancock R., Comas-Herrera A., Pickard L., Juarez-Guarcia A., King D., Malley J. The aim of this study is to make projections to 2050 of public and private expenditure on long term care services, and its distribution, under a wide range of options for reforming the system for funding long term care for older people in the United Kingdom. study involved linkage between macro- and micro- simulation models of long-term care financing. The former comprises cell-based modules for making projections of the numbers of service recipients and expenditures. The latter simulates the incomes and assets of future cohorts of older people and their ability to contribute toward care costs.research team have expanded their analyses of residential care charging and added analyses of home care charging. They have examined the impact on current and projected future public expenditure of a range of changes to the funding system. They have investigated the impact of the changes, together with the taxes needed to fund them, on different sections of the income distribution. key finding of this study is that removal of the means test for personal care would be costly and regressive. The same resources could be used more progressively by increasing the amount of income ignored in the means test.

0199 Understanding the reasons for the non recourse to the French Supplemental Public Health Coverage (called the CMU Complémentaire or CMUC) Wittwer J. (LEGOSé Paris Dauphine, Paris, France) Dufour-Kippelen S., Le Pen C., Legal A. This abstract features an analysis of the reasons for the non recourse to the CMUC, a means-tested French Public Health Program which provides an entirely free healthcare coverage. On the non recourse to public services are still very few in Europe. In France, the non recourse to the CMUC can be approximated to 25% by comparing the 4,7 million individuals benefiting today from CMUC, to the 6 million potential recipients initially envisionned when the 1999 law was voted.complementa-

ry methods allowing to confront objective and subjective data are used here. First, determinants of non resorting people are analysed by statistical and econometric methods exploiting the data from the National Survey on Health and National Health Insurance (ESPS) by the Research and Information Institute for Health Economics (IRDES). Then, face to face interviews of excluded people will be conducted in an organization devoted to welfare., the dynamic aspects will be approached thanks to the panel dimension of the survey and its retrospective questions.

0146 Socio-Economic, Demographic and Cognitive Risk Indicators of Dental Caries Amongst Adults in the Republic of Ireland Woods N. (University College Cork, Ireland) Objective: To identify the socio-economic, demographic and cognitive indicators of caries risk amongst adults in the Republic of Ireland. The data is based on a clinical examination undertaken in the National Survey Adult Oral Health between October 2000 and June 2002. Treatment need was estimated for 2,937 adults according to eligibility for dental services. The sample was weighted by gender, medical card status, and age. Methods: A high-risk caries prediction model was estimated for 16–24s, 35–44’s and those age 65 and over by logistic regression when the dependent variable (DMFS) was dichotomous, or by ordered logistic regression when the dependent variable was multichotomous. Goodness of fit was evaluated by estimating its sensitivity and specificity. Factors such as age, lower socio-economic status, oral hygiene behaviour, dental attendance, educational attainment, smoking, and unemployment were “high caries risk” indictors. “High need” indictors can be used to develop a population-based resource allocation formula. As some subgroups already have relatively good oral health, their capacity to benefit further is limited. A resource allocation formula could be developed, based on allocating resources according to the size of the problem (oral health need weighted in terms of the ‘capacity to benefit’ from care).

0126 Design of Efficient Health Care Systems Wranik D. (Dalhousie University – SPA, Halifax, Canada) Design of health care policy requires decision making with respect to choice of health care system characteristics, such as arrangements between patients, payers, and providers, public versus private insurance, remuneration of health care providers, gatekeeping arrangements, and sectoral resource allocation. These characteristics have been assessed from the perspective of health care system costs, but not from the perspective of health care system efficiency. Stochastic frontier analysis is used to assess the effects of health system characteristics on system efficiency, where characteristics are treated as environmental influences, which are neither inputs, nor outputs of the system. Results point toward the desirability of capitation or salary arrangements over fee for service arrangements, toward a shift of resources from acute care to chronic care sectors, and toward a greater share of public health care expenditure.

0275 Quality assessment of health systems: the connection between structure, process and outcomes of care within Germany Wuebker A. (WHL Graduate School of Business and Economics, Lahr, Germany)

Recently, there have been enormous efforts to measure the performance of different health systems (i.e. OECD, WHO) in order to get information about ways to improve the quality of a certain health system. However this research area is still in an early stage of development and there is much more research to be done. Objectives: The research plan outlines a framework to analyse the quality of a health system on the basis of the three quality dimensions introduced by Donabedian. It tries to develop a consistent line of arguments to clarify the logic of how structure and process measures affect outcome measures. To do so, it refers to selected diseases and tests the validity of this (theoretical) framework within an empirical analysis. Methods: As working basis serves the definition of quality developed by the Institute of Medicine (IOM) and the Donabedian approach that distinguishes between structure, process and outcome quality. In a first step the paper presents a consistent line of arguments to clarify the relation between those quality dimensions. The next step is to select a set of prevalent diseases to test the hypothesis, that good outcome indicators are dependent on good structure and good process indicators. Because of increasing financial shortages within health systems the study refers to the most costly medical conditions (i.e. heart diseases). This approach forms a link between medical and economic problems. Referring to evidence based medical knowledge measurable indicators are developed for each disease and the hypothesis mentioned above is tested. Results: Empirical analysis comes up with two results: First, the analysis confirms the predicted connection between the different dimensions of quality of care for some “Bundesländer”. Secondly, however, a good structure and a good process quality alone cannot guarantee a good outcome quality. Conclusion: In addition to the health system, there are other determinants (i.e. education, income per capita) that also affect the outcome quality of health care. Further empirical investigation of the influence of these factors on the outcome dimension could close the “explanation gap” and could deliver additional insights.

0500 Providing access for all to quality health and long-term care in a financially sustainable manner: the EU Open Method of Coordination at work Xavier A. (European Commission DG Employment, Unit E4 Social Protection, Pensions and Health Care, Brussels, Belgium) Par of the organised session by the European Commission: Policy choices for Europe- ageing and sustainability: “In 2004 the Council decided to extend the Open Method of Coordination (OMC) to health and long-term care. The OMC is an EU policy tool whereby EU countries exchange information and identify best practice while remaining fully responsible for the organisation of their services. It is used in the fields of Social Inclusion and Pensions. Consequently, in 2005, EU countries submitted to the Commission national statements regarding access, quality and financial sustainability in the fields of health and long-term care. The analysis of these statements identified common challenges faced by EU Member States in those fields, reviewed policy plans by individual countries and listed common key work areas where the OMC could be useful. These key areas include: ensuring access to care and reducing health inequalities, developing the provision of long-term care, promoting prevention, active life-styles and healthy ageing, promoting quality mechanisms, increasing patients’ choice and involvement, securing healthy financing to the system through efficient functioning of institutions, strengthening incentives to users and providers for rational resource use, and tackling staff shortages by developing human resources strategies. In 2006, EU countries will agree on the common policy objectives in these fields based on this reporting and analysis exercise.”

Quality assessment of health systems: the connection between structure, process and outcomes of care within Germany Eur J Health Econom Suppl 1 · 2006

| S103

0114 Modelling the Cost Effectiveness of EPO for Cancer Treatment-induced Anaemia Yao G.L. (University of Birmingham, UK.) Raftery J., Miners A., Hyde C., Wilson J. Background: Anaemia is a complication associated with chemotherapy. Whilst the use of erythropoietin (EPO) for the treatment of anaemia has been shown to decrease the risk of blood transfusion and increase haemoglobin (HB), its cost-effectiveness remains uncertain. Purpose: To evaluate the cost-effectiveness of EPO for the treatment of anaemia associated with chemotherapy compared with best standard care from the NHS perspective. Methods: A simulation model, incorporating sur vival, quality adjusted life years (QALYs) and costs, was developed. Patients were characterized by their baseline HB levels. Input values were derived from systematic reviews. The model has a 3-year time frame. The base case was 6 months chemotherapy treatment followed by an off-chemotherapy period. Results: The incremental cost-effectiveness ratio (ICER) was £119,247 per QALY gained with no survival gain to EPO. The ICER decreased to £31,119 when the most favourable survival gain was assumed. Results were sensitive to the length of timeframe and survival benefit of the treatment. Conclusions: EPO is effective in improving red blood cell transfusion requirements. Its impact on survival remains highly uncertain. If there is no impact on survival, it seems highly unlikely that EPO is a cost effective use of health care resources.

0131 The Long-term Cost-Effectiveness of Cardiac Resynchronization Therapy with or without an Implantable Cardioverter-Defibrillator Yao G.L. (University of Birmingham, UK) Freemantle N., Calvert M.J., Cleland J.G.F., Bryan S. Background: Cardiac Resynchronisation Therapy (CRT-P) is an effective treatment for patients with heart failure and cardiac dyssynchrony. Addition of an implantable-cardioverter defibrillator (ICD) function may further reduce the risk of sudden death. We assessed the cost-effectiveness of CRT-P and CRT-ICD compared to medical therapy (MT). Methods: A Markov model with Monte Carlo simulation to assess costs, life years and quality adjusted life years (QALYs) associated with CRT (± ICD) and MT was developed, based on a UK healthcare perspective. Model parameters were estimated from individual patient data from the CARE-HF trial. The additional sur vival benefit of an ICD was based on results from SCD-HeFT. The base case analysis was based on 10,000 individual life-time simulations assuming a battery life of 6 years. Result: The incremental cost-effectiveness of CRT-P compared with medical therapy £3,474 per QALY gained and £3,366 per life year gained. Data on the incremental cost-effectiveness of adding an ICD to CRT in patients at different risks of sudden death will be presented. Conclusions: Long-term treatment with CRT-P appears cost-effective compared to medical therapy. Compared to CRT-P, CRT-ICD was also cost-effective given a willingness to pay of £30000 per QALY.

0440 Setting Priorities for Managing Renal Failure in Ramallah Hospital (Palestine) Ziedan N. (University Palestinian Territory, Ramallah, Palestinian) Mataria A., Bate A., Donaldson C. This paper aims at describing a pioneering attempt to implement the Programme Budgeting and Marginal Analysis approach to decision-

S104 |

Eur J Health Econom Suppl 1 · 2006

making vis-á-vis managing renal failure at a main referral hospital : Ramallah Hospital : in the Occupied Palestinian Territory. State-of-theart standards in managing renal failure; e.g., kidney transplant, renal/ peritoneal home/hospital dialysis, and medication-based disease management were reviewed. The opportunity costs associated with the different options were estimated adopting a Ministry of Health and a societal perspectives. Health outcomes associated with the different treatment options were assessed using the EQ-5D Quality of Life instrument. Beside, programmes’ beneficiaries were also questioned about their willingness to pay values for benefiting from one option and not the others, using a marginal contingent valuation approach. Results form the costutility/benefit evaluations were then illustrated in a focus group discussion with an advisory panel composed of decision-makers and specialized health professionals. Discussions were used to inform the ‘list of priorities’ susceptible to enhance technical and allocative efficiencies in managing renal failure, while accounting for services growth and possible reductions. It was found that although renal dialysis would appear the option of choice from the MoH perspective, should a societal perspective be adopted benefits from kidney transplantation reveals to surpass transplant-associated costs.