Cost-Utility Analysis Using EQ-5D-5L Data: Does How the Utilities Are Derived Matter?

Abstract

Objectives: To explore how the use of EQ-5D-5L value set and crosswalk from EQ-5D-5L to EQ-5D-3L (and use of 3L value set) would affect cost-effectiveness analysis results for England and six other countries (Canada, the Netherlands, China, Japan, South Korea, and Singapore).

Methods: Individual-level utilities derived from primary 5L data using both value set (5L) and crosswalk (c5L) approaches were applied to three Markov models assessing the cost-effectiveness of hemodialysis (HD) and peritoneal dialysis (PD) for end-stage renal disease (ESRD) patients to estimate incremental quality-adjusted life years (QALYs). The mathematic functions between incremental QALY and utility were derived.

Results: 5L- and c5L-based incremental QALYs were similar in the model for non-diabetic patients (range: 1.910-2.149, 1.922-2.121). 5L tends to generate more incremental QALYs than c5L in the model for diabetic patients (range: 1.454-1.633, 1.365-1.568) but fewer incremental QALYs in the model for all ESRD patients (range: 0.290-0.480, 0.315-0.493). In all models, 5L (c5L) generated more incremental QALYs when Chinese (South Korean) value sets were used. The largest and smallest differences in 5L- and c5L-based incremental QALYs were observed when Chinese and Dutch value sets were used. Incremental QALYs was a positive linear function of both utility of PD and difference in utilities of HD and PD.

Conclusions: The value set and crosswalk approaches may not be used interchangeably in economic evaluation when EQ-5D-5L data are used to estimate utilities. Results of cost-effectiveness analysis using Markov models may be affected by both absolute utilities and their differences.

Keywords: EQ-5D-3L; EQ-5D-5L; cost-utility analysis; crosswalk; value set.