The candidate is a fellowship-trained stroke neurologist dedicated to a career in health services research. This application proposes a comprehensive program to facilitate the candidate's development into an independent investigator in patient-oriented research. The candidate's career goal is to become a leader in stroke health services research by combining expertise in clinical medicine, survey and epidemiological analysis, and health economics, to better understand the social and economic impact of stroke. The University of Michigan has outstanding clinical and health services research resources to support this proposal. The training component of this proposal includes formal graduate education in healthcare economics and survey methodology. The goal of the research proposal is to develop and refine methods for the population-based study of the direct and informal caregiving costs of stroke in the elderly. The Health and Retirement Study (HRS), a nationally representative, NIH-funded longitudinal study of the health and economic implications of aging, will be the main source of data for this proposal. The extensive data collected for the HRS will be used for the following Specific Aims.
Aim 1 - To develop methods for determining the annual direct medical costs and informal caregiving costs for elderly individuals in the US.
Aim 2 - To determine the relationship between stroke and annual direct medical costs and informal caregiving costs.
Aim 3 - To develop longitudinal models to predict time to institutionalization or death following stroke, and for estimating and predicting lifetime direct medical costs and informal caregiving costs associated with stroke. The proposed analysis will use all waves of the HRS data from 1992 through 2004. Annual direct medical costs will be calculated using HRS-linked Medicare claims data, gross costing, and self-reported medical expenditures. Informal caregiving costs will be calculated using self-reported hours of caregiving. The relationship between stroke and the above costs will be analyzed using a two-part regression model. Longitudinal models for predicting time to institutionalization or death will be developed using regression and proportional hazards analysis, and stroke-associated lifetime costs will be estimated using Markov modeling. The proposal will yield, for the first time, nationally representative estimates of the total economic burden of stroke, which will then be available for use by health policy decision makers.
