Estimating utility of health: Some methodological issues
Key Objectives
To develop and apply novel methods for modelling quality of life valuation using advanced econometric techniques. The strength of these techniques lies in their flexibility, in that the data informs the structure of findings, rather than being pre-defined by the analyst
Cost-effectiveness analysis of alternative healthcare interventions relies on having a measure of effectiveness, and many regard the quality adjusted life year (QALY) to be the current gold standard. In order to compute QALYs, we require a suitable system for describing a person’s health state, and a utility measure to value the quality of life associated with each possible state. There are a number of different health state descriptive systems, and the most commonly used one is the EQ-5D, under which health is decomposed into five dimensions and each dimension has three levels, resulting in a total of 243 (35) health states. We take a sample from these states and ask selected respondents from a target population to value these health states (thus providing something called utility scores). A regression model is then estimated and used to predict the utilities of all other health states.
In the last twenty years a large number of studies have been carried out to identify the best methodology on how to collect the EQ-5D valuation data. In contrast, there are very few studies on how to use these data to estimate utilities. In general the EQ-5D utility score is skewed, censored, hierarchical and noncontinuous. However, these features have been largely ignored by the existing economic valuation studies and very often a normal distribution assumption is adopted for the ease of estimation. This oversimplification is very likely to cause biased utility estimates and thus inaccurate cost-effectiveness analysis. Consequently, policy makers would make their decisions based on a fragile ground.
This project intends to fill in the gap and aims to identify and develop appropriate statistical tools that can accommodate the special features associated with the EQ-5D data. Moreover, we will use these better methods to analyse the Australian EQ-5D data collected through an NHMRC project conducted by the Centre for Health Economics Research and Evaluation (CHERE), and provide more accurate utility estimates to Australian health economists and policy makers.
Funding source
UTS Faculty of Business Grant
CHERE staff
Yuanyuan Gu, Richard Norman, Rosalie Viney
