Estimating health care costs for advanced metastatic bladder cancer patients who are lost to follow-up: Applications of the Kaplan-Meier sampling average (KMSA).

21 Background: There is an increasing need to evaluate the quality and cost of care within oncology. Estimating health care costs for patients who are lost to follow-up or censored often results in underestimating costs, biasing comparative assessments. This study compared traditional costing approaches with that of KMSA methods for patients treated for metastatic bladder cancer. Methods: Advanced metastatic bladder cancer patients receiving second line chemotherapy were identified from the Truven Health MarketScan Claims Databases (04/2011 to 09/2014). All-cause health care costs represent the total dollars paid to all providers of care regardless of diagnosis and were inflated to 2014 USD. Total, medical, pharmacy, and site-of-care costs from treatment initiation to death, end-of-enrollment, or end-of-study period were compared using the KMSA and traditional methods. In KMSA, the follow-up period is divided into small uniform intervals; total cost in each interval is then multiplied by the individual patient’s survival probability at the beginning of each interval and then summed across the intervals. Results: The censoring rate of patients’ follow-up was 49%. The mean per patient total health care captured by traditional method was only 72% of the total cost estimated by KMSA method or $32,930 less per patient. Overall, underestimates ranged from 28% to 41% with pharmacy, ER, and lab service costs underestimated the largest when using the traditional method. Conclusions: The KMSA produced estimates more reflective of actual patient-level health care costs in this claims-based observational cancer study compared to traditional methods by accounting for the effect of censoring. Researchers should consider KMSA when evaluating and comparing costs of care in the presence of variable follow-up.[Table: see text]