Does Physician Peer Effect Exist? Evidence from New Drug Prescriptions
2009/06/23
研討會日期 : 2009-06-23
時間 : 14:30
主講人 : 周欣怡副教授
地點 : B110
演講者簡介 : 周欣怡教授為Ph.D. in Economics，Duke University (1999)。現為Lehigh University經濟系副教授。其主要研究領域為健康經濟學、應用計量與應用個體。
演講摘要 : We investigate how and why a physician’s prescription of a new drug (the second-generation antipsychotics or SGA) is influenced by his or her colleagues at the same hospital during common working hours. We offer a theoretical framework that emphasizes the information externality through social learning to explain why peer effects among physicians’ prescriptions of a new drug should be expected. Thus, the main contribution of this paper is to use this framework to examine the existence of learning-based peer effects, a topic that is largely ignored in the past literature. We use a unique data set that includes the universe of antipsychotic medication use in Taiwan from 1997 to 2004. The longitudinal information on antipsychotic medication use by each physician-patient pair in each hospital allows us to identify the peer effects by controlling for time-invariant unobserved heterogeneities. Using the first-difference method that mitigates the simultaneity problem, we find that an increase by one hundred percentage points in the SGA prescription ratio of a physician’s same-hospital colleagues working in the same month can be transferred to approximately one percentage point increase in the physician’s SGA prescription ratio. We also find strong and consistent evidence in support of such learning-based peer effect across various settings. The monthly peer effect appears small; however, its cumulative effect over time is not trivial. Based on our empirical findings that the peer effect is persistent, we find that approximately 62% of the increase of SGA prescription between 1997 and 2004 inTaiwan could be explained by the social multiplier effect that depends on the peer effect.
