Essays in Health Economics

Abstract

In my dissertation, I study physician behavior and the role of physician styles. Chapter 1 establishes quasi-experimental evidence that physician styles (and not the sorting of patients) drive roughly 70% of variation in health care usage in U.S. Medicare after controlling for patient observables. The natural experiment relies on changes in the referral network of doctors for quasi-random assignment of patients to providers. Chapter 2 extends the investigation using the same natural experiment to study the effect of more health care on patient health. I find that more intensive medical care has minimal impact on patient mortality and hospitalization rates. However, I also show that patients who quasi-randomly receive more care initially also use more care in the longer-run as well. These two chapters establish that physician styles exist and that more intensive doctors do not appear to produce better patient outcomes. Chapter 3 studies how physician styles adapt to a negative income shock. In response to a nation-wide -5% change in Medicare reimbursement for physician services in 2002, doctors did not appear to dramatically adjust their behavior, suggesting no evidence of a short-run income effect by physicians.