OBJECTIVES: There is growing interest in using condition-specific preference measures, including the European Organization for Research and Treatment of Cancer Quality of Life Utility Measure-Core 10 Dimensions (EORTC QLU-C10D), to inform economic evaluations of cancer treatments. This research assessed the implications of using utility indices based on the 3-level EQ-5D (EQ-5D-3L), a mapping of the EQ-5D-3L to the 5-level EQ-5D (EQ-5D-5L), and the QLU-C10D and compared the psychometric properties of the indices.

METHODS: Utilities for progression-related states were estimated for 8 phase 3 trials of nivolumab±ipilimumab using the EQ-5D-3L UK and US value sets, mapped EQ-5D-5L English value set, and QLU-C10D UK and Australian value sets. These were entered into UK cost-effectiveness models to derive quality-adjusted life years (QALYs) for treatments. Psychometric properties of the indices were assessed using pooled trial data.

RESULTS: The mapped EQ-5D-5L English index yielded greater incremental QALYs for nivolumab±ipilimumab vs comparators than the EQ-5D-3L UK and QLU-C10D UK indices. Absolute differences in incremental QALYs ranged from 0.01-0.16 (mapped EQ-5D-5L English vs EQ-5D-3L UK), 0.04-0.31 (mapped EQ-5D-5L English vs QLU-C10D UK), and 0.02-0.15 (EQ-5D-3L UK vs QLU-C10D UK). All indices differentiated groups defined by baseline performance status (PS), EQ-5D visual analogue scale (VAS) score, and cancer stage. Similarly, all indices detected meaningful change in tumour response, VAS score, and PS. Effect sizes were generally similar in interpretation, though the QLU-C10D UK index was more sensitive to worsened EQ-5D VAS score and PS, while the mapped EQ-5D-5L English index was less sensitive to improved tumour response.

CONCLUSIONS: On average, the mapped EQ-5D-5L English index and QLU-C10D UK index yielded 6% more and 2% fewer incremental QALYs, respectively, than the EQ-5D-3L UK index. The lower QALY yields of the QLU-C10D were balanced by slightly greater validity and responsiveness. The implications of these findings for use of the EQ-5D-5L and QLU-C10D are discussed.