Skip to main content

3125 steps to perfect health

Research Team: Stephanie Thomas1, Feng Xie2
CIC Specialist: Dan Marrable

A non-parametric approach to developing the EQ-5D-5L value set

The EQ-5D-5L is a method for evaluating individual preferences for different health states. The method is frequently used in health technology assessments to compare the relative health impacts of health care interventions in the course of health care system evaluations. A key component of the EQ-5D-5L is that it provides analysts with numerical values for 3125 unique health states. To develop these values a population sample is asked to value a subset of health states and the values from these collected states are modelled in order to infer the values of the states which were not sampled. We compare the impact of different modelling strategies on the resulting numerical values. Using observations from the Canadian population collected with the Time-Trade-off method, and working with the team at CIC, we find that modern (and computationally intensive) non-parametric regression techniques offer substantial improvements in model performance and predictive precision. Our new model results in significant changes to the set of numerical values and therefore may have important implications for health system policy.

  1., Department of Finance, Economics and Property, Faculty of Business and Law, Curtin University and Department of Economics, McMaster University
  2., Health Research Methods, Evidence and Impact, McMaster University and Program for Health Economics and Outcome Measures, Hamilton, Canada