The relative influence of clinical information, financial incentives, and other factors on physician’s choice of chemotherapy

Abstract

6002 Background: Understanding the effect of financial incentives for specialty pharmaceuticals on physician choice has been a priority for policy makers. This research seeks to identify the extent to which published clinical information and financial incentives affect a physician’s choice of chemotherapy along with other factors. Methods: SEER-Medicare and Medispan pricing data were formed into a panel of 4,503 patients who were diagnosed with metastatic breast cancer and treated with chemotherapy from 1991 to 2002. A conditional logit discrete choice model was employed to analyze the effects of product attributes, including financial margin, volume of randomized controlled trial citations, FDA labeled indications, years since launch, generic status, and other covariates, on odds of choice of chemotherapy. A chemotherapy-level fixed effects specification was used to analyze choices of drug in patients’ first and subsequent drug administrations. Results: The natural log of clinical information was positive and significant (β = 1.270; p < 0.001); a one unit increase in log of citations increased odds of choice by 27 percent, all else constant. Physicians avoided generics (β = 0.703; p < 0.01), preferred on- to off-label drugs (β = 0.468; p < 0.01), and preferred older drugs to newer ones. Financial incentives for the first chemotherapy administration were positive and significant influences on choice of drug (β = 1.001; p < 0.001). A $100 increase in reimbursement per day increased the odds of choosing a drug by 10 percent. Preference for financial incentives, more clinical information, and on-label drugs declined in subsequent administrations. Conclusions: This research provides evidence that financial incentives have a positive but modest effect on the average physician’s choice of chemotherapy when treating metastatic breast cancer. New clinical information about a drug also had a positive effect on choice. [Table: see text]