Welfare Effects of Physician-industry Interactions: Evidence From Patent Expiration

Abstract

Many transactions occur via expert advisors, especially in the healthcare sector, and as such, firms frequently implement strategies to influence these advisors. The efficiency of these interactions is an empirical question. Using data on physician-industry interactions and prescribing behavior during the entry of a major generic statin drug, we examine the causal effect and welfare implications of the most common type of interaction: meals. Guided by a theoretical model of endogenous meals, we develop an instrumental variables identification strategy and document reduced form evidence that these meals directly influence prescribing decisions. We find a significant degree of negative selection and primarily extensive margin effects, whereby firms target small meals to prescribers with an otherwise low propensity to use the target drug. Given this evidence, we estimate a structural model of drug choice that allows us to predict counterfactual outcomes in a world where these meals are banned. Results from these counterfactuals are in line with theoretical predictions that these interactions can offset efficiency losses due to pricing with market power. However, this offset does not appear large enough to justify their existence in this particular market.