Validating generic preference-base measures of health in mental health populations and estimating mapping functions

The NHS has to make some tough choices about what services to provide. One way to make these choices is to examine the value for money of services by looking at the costs and benefits they provide. A major problem for the NHS is how to measure the benefits of a service. One measure of benefit that is becoming increasingly used is the quality adjusted life year (QALYs), which measures the gain in the quality of life experienced by the users of services.

The calculation of QALYs requires some means of valuing the quality of life of people in receipt of services. Current ?generic? measures of health have been developed mainly for people with physical problems. This proposal aims to critically review two leading measures for calculating QALYs. It includes interviews with service users about their quality of life and how well these measures cover those aspects of their life. This study will also look at how well these measures have performed in statistical terms in previous studies. This study hopes to help improve the methods used by researchers in assessing the effectiveness of treatments.

Mental health services form a major part of health care in the UK and should be subject to the same rigours of cost effectiveness assessment as much as other services. However, there is a general recognition that the generic preference-based measures of health used for generating Quality Adjusted Life Years (QALYs), such as the EQ-5D (as preferred by NICE) and SF-6D, may not be appropriate for some mental disorders and even where they are appropriate they are often not used. This limits the role of economic evaluation in mental health.

Part 1 of the proposal will examine whether or not the low take-up of generic measures in mental health research is justified in terms of their psychometric properties of two of the most widely used measures (i.e. EQ-5D and SF-6D) including feasibility, reliability, validity (content, face and construct) and responsiveness in mental health populations. This is addressed using a combination of qualitative and quantitative methods including: a review of the qualitative evidence on the domains of quality of life raised by mental health service users; qualitative interviews with mental health service users in order to examine the coverage and relevance of two generic measures (EQ-5D and SF-6D); a review of the literature on the psychometric properties of these generic measures; and a quantitative examination of the psychometric properties of these measures using existing individual level data sets. Part 2 seeks to address the low use of generic measures by estimating mapping functions between widely used condition-specific measures and generic measures in order to facilitate evidence synthesis and cost- effectiveness analysis.