Do Physicians Follow the Golden Rule? New Evidence on Imperfect Agency from Physicians’ Self-Prescriptions

Tuesday, June 12, 2018: 8:20 AM
Mountain Laurel - Garden Level (Emory Conference Center Hotel)

Presenter: Mariana Carrera

Co-Author: Niels Skipper

Discussant: Daniel W. Sacks


Designing incentive schemes for physicians requires an understanding of how physicians make tradeoffs between their patients’ benefit and their own self-interest. How physicians self-treat is informative about this agency problem. We use a unique dataset from Denmark on prescription drug claims for the entire population, including physicians’ prescriptions to themselves, to identify the weight that physicians put on a patient’s utility relative to their own.

In the Danish national healthcare systems, physicians are asked to be dual agents, acting primarily in the patient’s interest but also as stewards for the public insurer. We focus on statins—a class of cholesterol-lowering drugs that contains commonly prescribed older drugs as well as newer drugs that have been proven to be more potent. Guidelines advocate that all patients start with an inexpensive older drug and only move to newer drugs afterwards, if necessary. However, we document that when treating themselves, physicians are far more likely to start directly with a newer drug in violation of these guidelines. Also, they increase their own use of the newer drug in response to new evidence about its clinical efficacy. These differences persist when we use individual-level data on income, education, and health characteristics to compare physicians against other high-earning professionals of similar health status.

We use maximum likelihood methods to estimate a random utility model in which the drug costs paid by the government affect the physician’s utility in the same way, whether they are writing a prescription for themselves or for a patient. If they discount patient utility relative to their own, then they are more willing to trade-off government costs for clinical benefit when they themselves receive the clinical benefit. We find evidence that physicians weight patient clinical benefit less than their own, and that this weight is larger when risk factors are present that make cholesterol-reduction more important.