OBJECTIVES: Measures of health and well-being are critical to capturing the cancer patient journey. For economic evaluation, measures that generate preference-based scores (PBMs) are most useful and have emerged based on cancer measures, QLU-C10D and FACT-8U, as well as the generic EQ-HWB and CS-Base. This study goal was to compare the ability of different prominent and emerging measures of health and well-being (EQ-5D-5L, EQ-HWB, FACT-8U, QLU-C10D, and CS-Base) to capture the impact of cancer based on status and functioning.

METHODS: A cross-sectional panel of cancer patients (capped at 70 per cancer type; 40% provided proof of diagnosis) completed an online survey of generic and disease-specific measures. Cluster analysis explored predictive ability of each measure to classify patients based on clinical anchors of severity using: 1) ECOG status; 2) cancer progression; and 3) EQ-VAS-based scores (by quintiles). Each measure was compared to other measures, identifying four or five clusters based on response structure, then ordered by severity of index value or level-sum score. Agreement between classification of patients by severity (e.g. ECOG) and the clustering by a measure, e.g. EQ-5D was assessed using interclass correlation coefficients.

RESULTS: 520 patients completed the survey; 40% were diagnosed with stage III or IV cancer. The CS-Base tended to have higher concordance across the 3 criteria (ECOG, VAS, cancer status) which suggested broader content coverage, followed by the EQ-HWB, EQ-5D-5L, then the QLU-C10D and FACT-8U. The exception for the EQ-5D-5L with ECOG (43%), which is a functional ability driven measure.

CONCLUSIONS: Newer measures such as the EQ-HWB and CS-Base had content that aligned better with clinical measures such as cancer status and functional performance compared to cancer-specific PRO measures such as the QLU-10D and FACT-8D. In selecting among PBMs, one should carefully consider ability to capture cancer-related outcomes of interest.