How Wrong Can It Be? The Impact of Using Inappropriate Adult-Specific Value Set for EQ-5D-Y in Cost-Effectiveness Estimates

OBJECTIVES: EQ-5D-Y is commonly used in clinical trials to measure health-related quality of life in interventions involving children and/or young people (CYP). Utility scores derived from EQ-5D-Y are employed to evaluate the cost-effectiveness of the interventions. However, it is frequently observed that the EQ-5D adult-specific value set is used to generate utility scores for EQ-5D-Y, despite recommendations against this practice from the instrument developer. This study aims to investigate the impact of using adult-specific vs. youth-specific value sets for EQ-5D-Y on cost-effectiveness estimates.

METHODS: Data were obtained from a randomized control trial in England, which investigated the effectiveness and cost-effectiveness of LEGO®-based therapy compared to usual support in 248 CYP aged 7-15 years with autism spectrum disorder. Proxy-version EQ-5D-Y-3L data were collected. In the original trial analysis, the UK adult-specific value set was utilized due to the absence of a UK youth-specific value set. To assess the impact of value sets used on QALY gains, the Dutch and Germany value sets (both adult and youth) were applied to the original trial data.

RESULTS: Compared to the original results using the UK adult value set (QALYs gained of 0.009, 95% CI: −0.008 to 0.028), the estimated QALY gains were 0.015 (95% CI: -0.022 to 0.055) and 0.039 (95% CI: -0.057 to 0.143) when applying Dutch adult- and youth-specific value sets, respectively. Similarly, the QALY gains were 0.003 (95% CI: -0.031 to 0.038) and 0.037 (95% CI: -0.072 to 0.148) when using Germany adult- and youth-specific value sets, respectively. A substantial two to twelve-fold difference was observed between the use of youth- and adult-specific value sets.

CONCLUSIONS: This study demonstrates a substantial difference in QALY gain estimations when employing adult-specific versus youth-specific value sets for EQ-5D-Y. The findings strongly support against the use of adult-specific value set for EQ-5D-Y.