Measuring the cost-effectiveness (CE) of therapies treating metastatic colorectal cancer (mCRC).
405 Background: Traditional CE approaches often produce results that appear to be at odds with patients' (pts) behavior. Patients with terminal disease often value gains in survival much more highly than is suggested by standard CE estimates. In the context of metastatic cancer, CE classifies as cost-ineffective many therapies that pts appear to value highly. This study assesses whether CE analyses conflict with estimates of pt value for mCRC therapy. We infer pt value from the willingness to pay (WTP) of cancer pts for therapy out-of-pocket (OOP), and compare these estimates to the value implied by traditional CE. In contrast to traditional CE approaches, this study uses real-world data on pts' own WTP for therapy OOP. Methods: Revealed preference analysis is performed on retrospective claims data, including the pt OOP spending component, to infer the value metastatic cancer pts (N=13,938) place on 29 different therapies for CRC, breast, lung or head and neck cancer. The primary outcome measure is the average value placed by pts on therapy. Secondary outcomes are: utilization of therapy, total expenditures on cancer drugs, and the price elasticity of demand for therapy. Results: Utilization of therapy displays a negative but inelastic relationship with OOP spending on drugs (the estimated price elasticity is -0.007, with p<0.01). The estimated annual WTP OOP for treatment of mCRC specifically was $207,555, approximately 19 times the average annual cost of therapy ($10,775). Approximately 99% of mCRC pts value the therapy more than the cost. The analysis implies that the average mCRC pt values a statistical life year at approximately $282,000. In comparison, traditional CE analyses employed by regulators assign much lower values for a statistical life-year; for example, the UK's NICE uses a threshold of <$50,000 per year. Conclusions: Traditional economic valuations of medical treatment perform poorly in the terminal care setting, because they fail to recognize that pts place higher value on care provided at the end of life. As a result, these methods vastly understate the CE of treatment for mCRC and other terminal diseases. Use of pts' own OOP spending behavior appears to produce different conclusions for the value of therapy. [Table: see text]