Community supply of hospice: does wealth play a role?

Abstract

Hospice is underused by older, rural, and minority populations. Because local availability of hospice is an important predictor of use, we aimed to identify geographic variation in hospice supply and examine its community-level determinants, including wealth. This was an observational geographic study using the 2008 National Hospice and Palliative Care Organization's National Data Set and the 2,000 U.S. census data for 3,140 U.S. counties. Our outcome of interest was hospice supply defined according to the number of hospice programs (regardless of level of care) servicing each county. We used binomial multivariable regression to test the relationship between supply and log-transformed counts of population, African Americans, Hispanics, residents aged 65, and high school educated residents as well as area, median household income, and certificate of need status. Hospice availability varied greatly across the United States, with an unadjusted mean of 24.3 hospice programs servicing each county (standard deviation 19.7, range 0-160). After adjusting for all covariates, median household income (incidence rate ratio [IRR] 1.03, P < 0.001), population count (IRR 17.9, P < 0.001), count of African Americans (IRR 1.26, P < 0.001) and elderly adults (IRR 2.81, P < 0.001) positively predicted supply, whereas area (IRR 0.84, P < 0.001), certificate of need status (IRR 0.89, P < 0.001), count of Hispanics (IRR 0.86, P < 0.01), and high school educated (IRR 0.03, P < 0.001) negatively predicted supply. There is gross variation in hospice supply that can be explained by community wealth, population density, age, ethnicity, and race. To address disparate utilization of hospice, the relationship between wealth and availability will need to be better understood.