Determinants of Medical Spending for the Elderly: Insurance, Patients, Providers

The goal of this research is to better understand the determinants of health care spending among the elderly in the United States. This is an important topic given the size and growth of the health care sector, and the pressure this places on public sector budgets. The elderly, as disproportionately large consumers of health care in general and publicly funded health care in particular, are a natural place to focus.
One natural cause of health care spending is moral hazard from health insurance; by reducing the cost of health care to the consumer, insurance may encourage more health care utilization. In previous work supported under the grant, we produced new analyses of moral hazard effects in health insurance, including an analysis of selection on moral hazard. A primary component of the current proposal is to further enrich our understanding of how health insurance affects health care spending by examining two specific areas: 1) heterogeneity - across types of people and types of health care - in the moral hazard effects of health insurance and 2) how the non-linear budget sets created by typical health insurance contracts affect the impact of health insurance contract design on annual spending and on spending over longer horizons. This work will be focused on the Medicare Part D prescription drug program. Together, the results will enable researchers
and policymakers to better model both aggregate and distributional impacts of alternative Medicare Part D consumer cost-sharing designs on prescription drug spending. The second part of the proposed renewal moves beyond the impact of insurance to examine the role of two other likely determinants of health care spending: patient demand and provider behavior. This work will draw primarily on the Medicare Parts A and B programs and focus on better understanding the determinants of the highly-publicized geographic variation in health care spending among Medicare beneficiaries. We will develop an empirical framework for estimating the contribution of patients (demand) and place (supply) in creating geographic variation in health care utilization. We will extend these methods to shed light on the factors that in turn contribute to patient demand and provider supply. On the demand side, we will examine the relative contributions of patient health and patient preferences over medical treatment, and in turn explore the role of
past experience in affecting patient preferences. On the supply side, we will examine, among other things, the role that the individual physician plays in affecting health care utilization, independent of other aspects of the health care environment (e.g. nature and type of hospitals and technology in the area). These analyses will increase our understanding of the underlying sources of geographic variation in health care practice, which is both of interest in its own right and a crucial, and natural first step for thinking about the likely impact of potential public policies aimed at improving the efficiency of health care delivery.