The growth of hospice care in U.S. nursing homes

To inform efforts aimed at reducing Medicare hospice expenditures by describing the longitudinal use of hospice care in nursing homes (NHs) and examining how hospice provider growth is associated with use.

Longitudinal study using NH resident assessment (Minimum Data Set) and Medicare denominator and claims data for 1999 through 2006.

NHs in the 50 U.S. states and the District of Columbia.

Persons dying in U.S. NHs.

Medicare beneficiaries dying in NHs, receipt of NH hospice, and lengths of hospice stay were identified. The number of hospices providing care in NHs was also identified, and a panel data fixed-effect (within) regression analysis was used to examine how growth in providers affected hospice use.

Between 1999 and 2006, the number of hospices providing care in NHs rose from 1,850 to 2,768, and rates of NH hospice use more than doubled (from 14% to 33%). With this growth came a doubling of mean lengths of stay (from 46 to 93 days) and a 14% increase in the proportion of NH hospice decedents with noncancer diagnoses (69% in 1999 to 83% in 2006). Controlling for time trends, for every 10 new hospice providers within a state, there was an average state increase of 0.58% (95% confidence interval=0.383-0.782) in NH hospice use. Much state variation in NH hospice use and growth was observed.

Policy efforts to curb Medicare hospice expenditures (driven in part by provider growth) must consider the potentially negative effect of changes on access for dying (mostly noncancer) NH residents.