Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

High health care expenditures and an escalating number of elderly in western European countries force policymakers to consider what kind of treatments and interventions that can be offered by the public health care system. Decisions and interventions have to be justified both from a medical perspective and from an economic point of view. The importance of economic evaluation of stroke related care and rehabilitation for instance will increase in the future due to the increasing number of elderly, and thus an increase in stroke cases. However, cost utility analyses (CUA) are by many still considered controversial. One of the controversies is how health related quality of life (HRQoL) is measured. This is the first study that compares the performance of EQ-5D and 15D in measuring HRQoL in post stroke patients. The main objective of the paper is to analyse whether 15D agrees sufficiently with EQ-5D, and to discuss which impact the choice of MAU instrument will have in a cost utility analyse. The paper gives an exhaustive presentation and a head to head comparison of 15D and EQ-5D. EQ-5D is considered as the criterion measure while validity is tested. The degree of agreement between the two instruments is assessed, and their practicality investigated. Clinical indicators, health status measures, and socioeconomic/-demographic factors are used to investigate whether and when the utility scores obtained with the two measures are likely to differ. We also employ two different value sets for EQ-5D in order to see whether that got an impact of the results of the analysis. We use a cost example to demonstrate the instruments’ influence on a CUA. Norwegian post stroke patients are uniquely recruited for this study via one of the largest university hospitals in Norway. The data are collected by a postal questionnaire. Both of the instruments are found practical to use and easy to handle, but there is a slightly higher completion rate in EQ-5D. We find evidence of strong ceiling effects in EQ-5D. Both instruments demonstrate good construct validity; they detect utility difference within sub groups equally good. We also found evidence of concurrent, convergent and discriminant utility in 15D. However, there is only evidence of fair agreement between the two instruments. 15D gives a significant higher average utility score than EQ-5D, independent of value set. There are also significant differences in the average utility score within EQ-5D when using the two different value sets. The choice of multi attribute utility instrument, as well as choice of value set must therefore be done with precaution; they can influence the incremental cost effectiveness ratio of an intervention.