EORTC QLU-C10D was similarly valid and sensitive as EQ-5D-5L but more responsive to cancer patients' health deterioration

Abstract

Objectives: The psychometric performance of the Quality of Life Utility-Core 10 Dimensions (QLU-C10D), a recently developed disease-specific preference-based measure (PBM) for patients with cancer, is not well understood yet. This study aimed to compare the construct validity, sensitivity, and responsiveness of QLU-C10D with that of 5-level EQ-5D (EQ-5D-5L), a generic PBM. Study Design and Setting: We recruited patients with cancer from outpatient clinics of a tertiary cancer hospital then interviewed them face-to-face at two consecutive clinic visits using both QLU-C10D and EQ-5D-5L questionnaires. Construct validity was assessed through known-group comparisons and correlation analysis, agreement of utility scores was examined using intraclass correlation coefficient (ICC), and sensitivity and responsiveness compared using effect sizes (ESs) derived from known-group comparisons and standardized response means (SRMs) derived from within-group comparisons, respectively. Results: We surveyed 626 patients; 280 of whom also completed the follow-up survey. Mean baseline QLU-C10D and EQ-5D-5L utility scores were 0.799 (SD: 0.224) and 0.916 (SD: 0.156), respectively. Both utility scores demonstrated known-groups and convergent/discriminant validity and their agreement was moderate (ICC: 0.60). EQ-5D-5L's ES was slightly higher than QLU-C10D's ES for cancer stage, whereas QLU-C10D's ES was higher for Eastern Cooperative Oncology Group. The SRM of QLU-C10D was considerably higher than that of EQ-5D-5L for deteriorated patients (−0.43 vs −0.02), whereas their SRMs were similar for the improved patients. Conclusion: Although QLU-C10D and EQ-5D-5L utility scores are both valid in patients with cancer, they may not be used interchangeably, and the QLU-C10D scores appear to be more responsive to deterioration in health status than the EQ-5D-5Lscores.