Methods of analysing health-related quality of life data for economic evaluation

Abstract: There are several preference-based measures (PBM) available to elicit HRQoL values for cost-effectiveness analysis (e.g. EQ-5D, HUI etc). The comparability of cost-effectiveness studies using different PBM's has been questioned and it is for this reason that NICE has specified a preference for the estimation of HRQoL using one measure, EQ-5D, to ensure the comparability of evidence for decision-making purposes.

Despite having a clear framework, there are a number of challenges to NICE's chosen methodology. One is a lack of relevant data. Clinical trials are typically a key source of evidence for cost-effectiveness analysis but evidence suggests that it is rare to find that HRQoL data are used from clinical trials. Instead, relevant data are derived from a range of other sources, including the published literature and observational studies, and synthesized within a decision-modelling framework.

In this seminar, I will discuss the different methods for analyzing HRQoL for use in cost-effectiveness analysis with a specific focus upon mapping methods used for linking intermediate health outcomes to EQ-5D estimates. I will explore the relative merits of the different methodologies in terms of their assumptions about the distribution of HRQoL data and predictive performance. Finally, I will consider the implications of the different methods of analysis in terms of the impact upon uncertainty in a decision model.