How Efficient is the Market for Physician Referrals?

Abstract
Primary Care Physicians (PCPs) hold significant sway over a patient’s health care decisions, particularly with regard to the choice of specialist. But to what extent do PCPs efficiently learn about specialist quality, if at all? In this paper, we study PCP-to-specialist referral networks and the responsiveness of PCPs to negative quality signals. Using the population of orthopedic procedures for Medicare fee-for-service beneficiaries from 2008 through 2018, we estimate a learning model in which PCPs update beliefs on specialist quality based on the outcomes of their referred patients. Preliminary results show that PCPs are responsive to negative outcomes from specialists, but the magnitude of this response is relatively small. As such, there is potential for large welfare gains if PCPs can improve the efficiency with which they adjust their referral patterns.