Abstract:

This dissertation presents three empirical essays in health economics. In my first essay, I examine how cancer in the family affects an individual's demand for health insurance and cancer screenings. This question is central to the debate about the impact of genetic knowledge on the medical system. For some individuals, cancer in the family reveals information regarding genetic endowment (information effects), which will in turn influence economic behavior. In addition to revealing individual-specific information, cancer in the family might also raise awareness of general cancer risk and the disutility from developing cancer (salience effects). Using data from the 2000 U.S. National Health Insurance Survey, I test for the effect of cancer incidence in the family (parents or siblings) on insurance demand and cancer screening. Furthermore, I use information on the type of cancer to differentiate information from salience effects. In the case of screening, I find that information effects from the incidence of genetic cancers in the family increase the rate of screening by 11 to 47%; that the salience effects are approximately a third of the genetic cancer effects. In the case of insurance, I find that individuals with a family history of cancer are 5% more likely to be insured via both the information and salience channels. These findings have implications for how people use information regarding their future health status and for the empirical research on asymmetric information in insurance markets.

My second and third papers focus on the relationship between insurance type and the provision of preventive care. The economics literature on managed care and supply-side incentives in medicine has generally emphasized constraining the use of services. However, many valuable (and cost-effective) medical services are used sub-optimally and managed care was initially designed to encourage the use of these services. A wide variety of preventive services fall into this category, including preventive cancer screenings, beta-blocker use after acute myocardial infarction and monitoring conditional on diabetes. In these papers (second paper joint with Grant Miller), I provide new evidence on how health insurance arrangements influence preventive cancer screening and care following diagnosis of AMI and diabetes. We find that at recommended ages the probability of being screened increases by two to six percentage points more (or 14% to 70% more) in health maintenance organizations (HMO) than in indemnity plans, preferred-provider organizations (PPO), or point-of-service (POS) plans. I also find large differences across plans in secondary prevention of diabetes but no difference in secondary prevention of acute myocardial infarction. HMOs appear to achieve these results by contracting selectively with more efficient health care providers.