Mapping Expanded Prostate Cancer Index Composite (EPIC) Questionnaire to EuroQoL-5D (EQ-5D) Utility Weights to Inform Economic Evaluations for Prostate Cancer

Abstract

OBJECTIVES: To develop a mapping algorithm to obtain EuroQoL-5D-3L (EQ5D) health utilities from Expanded Prostate Cancer Index Composite (EPIC) questionnaire. METHODS: This mapping study utilized baseline data from an international, multicenter, randomized controlled trial (NCT00331773) of patients with low-risk prostate cancer. Patient health-related quality-of-life (HRQoL) data were collected using EPIC and health utilities were obtained using EQ5D. Data were divided into an estimation sample (n=765, 70%) and a validation sample (n=327, 30%). The relationship between the instruments was estimated using ordinary least squares (OLS), Tobit, and two-part models. Five-fold cross-validation (in-sample) was used to compare the predictive performance of the estimated models. Final models were selected based on root mean square error (RMSE). OLS models using baseline cross-sectional data, combined data from all assessment periods, and random effects (RE) models that explicitly model the longitudinal nature of the data were estimated to compare predictive ability of algorithms derived from cross-sectional and longitudinal data. Longitudinal predictive performance of OLS models derived using baseline data was examined in the post-intervention data. RESULTS: A total of 565 patients in the estimation sample had complete information on both EPIC and EQ5D questionnaires at baseline. Mean observed EQ5D utility was 0.90±0.13 (range: 0.28-1) with 55% of patients in full health. Low to moderate correlations were found between EQ5D utility and urinary (r=0.38), bowel (r=0.34) and hormonal (r=0.55) domains of EPIC; sexual domain was weakly correlated (r=0.18) with EQ5D utility. OLS models outperformed their counterpart models for all pre-determined model specifications. The best model fit was: “EQ5D utility = 0.248541 + 0.000748*(Urinary Function) + 0.001134*(Urinary Bother) + 0.000968*(Hormonal Function) + 0.004404*(Hormonal Bother) – 0.376487*(Zubrod) + 0.003562*(Urinary Function*Zubrod)”; RMSE was 0.10462. When comparing cross-sectional vs. longitudinal data, a mapping algorithm obtained using combined EPIC subdomain data outperformed other model types. Mean absolute differences (MDs) between reported and predicted were low in general and decreased as the time of assessment increased. CONCLUSIONS: This study identified mapping algorithms to generate EQ5D utilities from EPIC domain or sub-domain scores, with satisfactory longitudinal predictive performance. The study results will help estimate quality-adjusted life-years in future economic evaluations of prostate cancer treatments.