Quality of Care at the End of Life for Nursing Home Residents With and Without Dementia

Prior research has suggested nursing home (NH) residents with dementia receive suboptimal quality of care at the end of life (EOL). However, there has been no national-level empirical study comparing EOL care for residents with and without dementia. This study examined the within-facility differences in EOL care, and explored longitudinal and cross-sectional variations in EOL care between residents with and without dementia in US NHs.

Study population included 628,031 long-term decedents from 13,719 NHs who died during CY2005-2007. Databases consisted of the Medicare Denominator File, hospital inpatient claims, hospice claims, MDS, and OSCAR.

For each outcome measure (in-hospital death and hospice use), we fit a series of logistic regression models with individual characteristics and facility fixed-effects.

Overall, 61.3% of decedents had dementia, 16.4% died in hospital (14.4% and 19.8% with and without dementia, respectively), and 34% used hospice services (34.4% and 33.4% with and without dementia, respectively). While prevalence of in-hospital deaths remained virtually un-changed from 2005 through 2007, hospice use increased from 31.94% in 2005 to 33.58% in 2006, and then to 36.08 in 2007 (from 31.83% to 34.01% to 36.83% for residents with dementia, and from 32.01 to 32.88% to 34.89% for those without dementia). Within the same facility, residents with dementia were significantly less likely to die in hospital (OR=0.78, CI: 0.75-0.81) and more likely to use hospice care (OR=1.08, CI: 1.05-1.12).

Based on the EOL outcome measures we studied, decedents with dementia received higher quality EOL care compared to decedents without dementia in the same NH.