Conflict of Interest and Incentives in Health Care
This dissertation explores conflicts of interest and incentive schemes in the health care industry. Chapter 1 explores internal conflicts that a firm may unwittingly introduce itself. Even when perverse incentives do not exist, the introduction of additional activities or tasks may create a conflict at the operational level by undermining the firm‘s ability to perform its primary task. I will examine this question in the setting of kidney transplant centers by estimating the effect of diversification into liver transplants on the risk-adjusted mortality rates of kidney transplant recipients. The results suggest diversification in general increased the risk adjusted mortality rate, but this increase was largely offset for older patients who tend to have more comorbidities. Further, the effect of diversification was more detrimental for smaller centers. This suggests that a firm‘s patient (or customer) mix and characteristics may define what constitutes a conflict for the firm. Chapters 2 and 3 examine external conflicts of interest in the form of perverse incentives. Payments from pharmaceutical manufacturers to physicians typify the type of complex conflict of interest relationship firms in many industries face. Such payments may create value for hospitals in the form of samples, financing trainings, etc, but they may also capture value away by inducing physicians to prescribe more (and more expensive) drugs. Chapter 2 examines whether disclosure of such payments has any effect on physicians‘ prescribing behavior. I exploit a natural variation in the timing of disclosures for Pfizer, in which payments can be observed but physicians were not aware that they would be disclosed. This enables estimation of the effect of disclosure on physician behavior using a differences-in-differences approach. The results indicate that while the post-disclosure period was associated with a decline in the number of prescriptions overall, physicians whose names had been disclosed as having received payments actually slightly increased the rate at which they prescribe branded drugs. Chapter 3 explores whether there is concrete evidence that payments to physicians have a persuasive effect, rather than an informational effect or simply reflecting the physician‘s existing preferences. This analysis relies on another natural experiment, in which physicians had their payments cut, unrelated to any changes in the product offerings of the company paying them. I use this cut to demonstrate that these physicians significantly altered their prescribing behavior.