Realities in cost-effectiveness analyses: a study of castration-resistant prostate cancer patients using a medical claims database

Previous cost-effectiveness analyses (CEAs) of abiraterone for castration-resistant prostate cancer (CRPC) patients have not shown favorable results for this new drug. These CEAs were generally conducted based on models used in clinical trials, where comparisons were made with patients given placebos. However, details on any other therapies provided to the comparison groups were not analyzed. These additional therapies should be considered when conducting CEAs to ensure better applications to clinical practice and policymaking. The objective of this study was to elucidate the actual therapies provided to CRPC patients using real-world claims data. We obtained anonymized computerized health care claims data of Japanese prostate cancer patients from the Japan Medical Data Center. This database comprises data from more than 2.5 million insured persons aged below 75 years from over 50 companies between January 2005 and June 2013. From among the prostate cancer patients, we identified CRPC patients as those who had been administered docetaxel, and further investigated their treatments and health care costs. Health care costs were estimated using a regression model accounting for variations in inpatient care, chemotherapies, death, and age. We identified 2138 prostate cancer patients, 36 of whom had been administered docetaxel. We excluded patients diagnosed with other cancers, resulting in a final sample of 18 cases. Of these, 66.7 % were administered other types of chemotherapy, which had not been considered in the control groups in most previous CEAs. We estimated mean health care costs for CRPC to be approximately US$952 per patient per month, and found that these costs were significantly affected by inpatient care and chemotherapy use. Actual therapies include a variety of treatments for CRPC patients, including various types of chemotherapy. Our study estimated health care costs based on real-world claims data. This study contributes to future CEAs by not only providing an estimate of health care costs for these patients, but also demonstrating that the actual therapies provided to comparison groups should be considered when conducting CEAs.