Essays in Health Economics

Abstract

This dissertation investigates whether healthcare decisions and health behaviors are affected by institutions and incentives, focusing on the legal system (Chapter 1), health insurance coverage (Chapter 2), and soda taxes (Chapter 3).

Chapter 1 investigates to what extent physician decisions to abandon ineffective treatment practices are affected by medical malpractice standard of care definitions. In some states the standard of care requires doctors to follow customary practice of the community in which they practice and in others physicians must adhere to national customs. Local and national practice can and often do differ. Legal scholars hypothesize that local standards of care reduce the incentive for physicians to keep abreast of medical advances, slowing the adoption of new treatments and the de-adoption of ineffective ones. This chapter analyzes state court cases to categorize state standard of care definitions as based on local or national custom. Next the chapter examines the effects of these definitions on patient care, focusing on the physician's decision to discontinue the use of vertebroplasty - a surgical procedure to alleviate pain after vertebral fractures - after two influential studies questioned its effectiveness. I find that while de-adoption occurred rapidly in all states regardless of the legal standard of care, rural areas reduced vertebroplasty use by less in locality states than they would have had a national standard of care applied.

Chapter 2 (joint with Helen Levy) examines the effect of Medicaid insurance coverage on healthcare utilization for seriously ill patients who are hospitalized after seeking care in an emergency room. Two channels exist through which Medicaid coverage may affect healthcare use: on the extensive margin more people may gain Medicaid coverage, and because they are now insured they may be more likely to seek care. On the intensive margin, conditional on seeking care, insurance coverage through Medicaid may affect “treatment intensity,” or “how much” patients are treated. Focusing on non-deferrable admissions allows us to estimate intensive margin effects without the confounding effects of selection into treatment and changes in patient composition. We find that the 2014 Medicaid Expansions increased the share of patients covered by Medicaid, which was partially offset by a decrease in the share of patients with private coverage. We find no statistically significant effects of Medicaid coverage on treatment intensity or on mortality. The coefficients were imprecisely estimated because the analysis does not separately identify effects from the coverage gain channel and the crowd-out channel, and these effects likely operate in opposing directions: positive and negative, respectively.

The last chapter evaluates the effects of sugar-sweetened beverage taxes on prices and consumption, comparing taxes in Berkeley and Philadelphia within the same study and using the same methods so that measured differences can be more easily attributed to local supplier and consumer responses rather than to differences in methodology. In Berkeley (Philadelphia), 6% (57%) of the tax on regular soda was passed on to consumers and in Philadelphia 67% of the tax was passed through for diet soda. In both cities pass-through declines with beverage size. Consumption of regular soda in Berkeley (Philadelphia) decreased by 7.6% (28%) and in Philadelphia consumption of diet soda decreased by 33%. In Philadelphia I find evidence of cross-border shopping: stores in neighboring regions lowered prices and purchases of soda from these stores increased.