The pattern of use of long-term care services among African-American elderly differs from that of white Americans. Most notably, African-Americans are significantly under-represented in nursing homes. No studies have sufficiently assessed the extent to which these differences are due to differences in socioeconomic status or other independent factors. Furthermore, no studies have demonstrated the impact of this difference in the use of long-term care services on the survival of black elderly. This study will analyze the differences between black and non-Hispanic white elderly in their patterns of use of long-term care services after they become significantly impaired. It will focus on the contribution of socio-economic factors to these observed black-white differences and the relationship between the types of long-term care services used and survival. The analysis will be based on an economic model which integrates the determinants of three sets of long-term care outcomes: 1) living arrangements, 2) use of community long-term care services, and 3) mortality. This economic model will serve as a framework for understanding the behavior of families in response to significant functional impairment in a parent and the relationship of this response to mortality. It will provide an economic perspective to understanding the relative role of race as a proxy for socio-economic status in the use of long-term care services. Household and family size and composition, financial resources, and opportunity costs will be studied as potential factors explaining the different patterns in the use of long-term care services for black and white elderly. We will estimate equations for the hazard of nursing home entry, the choice of community living arrangements, the use of community long-term care services, and the mortality of a population of elderly parents of respondents to the Panel Survey of Income Dynamics. The effects of race will be assessed by the comparison of two analytic approaches. First, we will model the effect of race by entering race as a dummy variable in the equations for each outcome. Second, we will estimate separate equations for blacks and whites and compare the patterns of significant explanatory variables in the two sets of equations. Statistical techniques will correct for the competing risks of death and nursing home admission and the non-random selection of parents who become significantly impaired.
