Risk of nursing home admission among older americans: does states' spending on home- and community-based services matter?

Abstract

Objective: States vary greatly in their support for home- and community-based services (HCBS) that are intended to help disabled seniors live in the community. This article examines how states' generosity in providing HCBS affects the risk of nursing home admission among older Americans and how family availability moderates such effects.

Methods: We conducted discrete time survival analysis of first long-term (90 or more days) nursing home admissions that occurred between 1995 and 2002, using Health and Retirement Study panel data from respondents born in 1923 or earlier.

Result: State HCBS effects were conditional on child availability among older Americans. Living in a state with higher HCBS expenditures was associated with lower risk of nursing home admission among childless seniors (p <.001). However, the association was not statistically significant among seniors with living children. Doubling state HCBS expenditures per person aged 65 or older would reduce the risk of nursing home admission among childless seniors by 35%.

Discussion: Results provided modest but important evidence supportive of increasing state investment in HCBS. Within-state allocation of HCBS resources, however, requires further research and careful consideration about fairness for individual seniors and their families as well as cost effectiveness.

Figure 1

Effects of state home- and community-based services (HCBS) expenditures on the hazard of nursing home admission by child availability. The x axis indicates a log scale of the level of states' annual expenditures on HCBS per person 65 years and older, ranging from $35 (the lowest level among the states in the sample) to $1,303 (the highest). The decile points of the log scale are indicated in exponentiated dollar amounts. The y axis indicates the estimated hazard of nursing home admission. The hazard was calculated for the “average” case based on the complementary log-log model presented in Table 4, fixing the values of independent variables at mean levels for continuous variables and 0 for dichotomous variables. “Child nearby” indicates whether a respondent had any child living within 10 miles. In 1995 interviews, 12.0% of respondents had no child, 28.9% had child(ren) living more than 10 miles away, and 59.1% lived with child(ren) or within 10 miles of child(ren).