OBJECTIVES: To develop a mapping algorithm for converting the prostate cancer specific Health-Related Quality of Life (HRQoL) instrument FACT-P (Functional Assessment of Cancer Therapy–Prostate) to the preference-based EQ-5D (Euro-QoL 5D) instrument for measuring health status. METHODS: Data were obtained from the phase 3 placebo-controlled AFFIRM trial of enzalutamide-proposed INN (MDV3100) in men with metastatic castration-resistant prostate cancer previously treated with docetaxel-based chemotherapy. EQ-5D and FACT-P data were collected for a subset of patients at baseline and throughout the study until treatment discontinuation. We compared three statistical mapping techniques to estimate patients’ EQ-5D index scores determined using the UK-tariff: (a) linear regression estimated by generalized estimating equation (GEE) algorithms; (b) two-part model (TPM) combining logistic and linear regression estimated by GEE algorithms; (c) separate mapping algorithms for patients with poor health defined as FACT-P ≤78. To select the best model specification, four different sets of explanatory variables were compared. The models were fitted to the full dataset and cross-validated using a 10-fold in-sample cross-validation. The variance explained by the model was assessed by the marginal R2. Model performance was assessed by comparing predicted and observed mean EQ-5D scores, the mean absolute error (MAE) and the root mean squared Error (RMSE). RESULTS: Values for both FACT-P and EQ-5D were available for 234 patients. The TPM model including the FACT-P sub-domain scores and demographic variables was the best-performing model (marginal R2 = 0.689) providing the most accurate predictions (MAE = 0.125; RMSE = 0.170). The physical well-being and prostate cancer specific subscales in the logistic part and functional and emotional well-being subscales in the linear regression part had the highest explanatory value. CONCLUSIONS: The developed algorithms for mapping FACT-P to EQ-5D enable the calculation of appropriate preference-based HRQoL scores for use in cost-effectiveness analyses when EQ-5D data are missing or inadequate