The purpose of this study was to compare characteristics of nursing home residents who died to those who remained alive during a two year period and to describe trends in end-stage identification and hospice use. We conducted a secondary analysis of minimum data set (MDS) data for 45,069 residents in 103 homes located in 12 states.  Most recent MDS assessments were matched with discharge data from the same time period to identify residents who died.

The sample was 68.9% female and 92.1% Caucasian with a 19.05% mortality rate (8,584 residents died). Residents who died were significantly more likely to be male, white, older, and widowed. Of those who died, 21.8% were identified as end-stage and 15% received hospice care prior to death. Decedents who were coded as end-stage prior to death were more likely to be non-white, married, and younger. Those who received hospice prior to death were more likely to be non-white, female, divorced, and younger. Each additional year of age reduced the odds of end-stage identification by 1.6% and reduced the odds of receiving hospice care by 1.0%. Therefore, the odds of a 90 y.o. resident receiving hospice care were 20% less than the odds of a 70 y.o. resident receiving hospice care. Hospice use varied widely among diagnostic groups from 26.4% of the decedents with cancer to 14.0% of those who died from congestive heart failure. Implications: The dying trajectory is frequently undocumented and many residents who could benefit from hospice do not receive it. Older residents and those with terminal illnesses other than cancer are most at risk.