A national study of the location of death for older persons with dementia

Abstract

Objectives: To describe where older Americans with dementia die and to compare the state health system factors related to the location of dementia-related deaths with those of cancer and all other conditions in this population.

Design: Cross-sectional study.

Setting: United States in 2001.

Participants: All persons aged 65 and older who died from nontraumatic causes in the United States during 2001. The underlying cause of death, as determined by the National Center for Health Statistics from death certificate data, was used to categorize subjects into three decedent groups: dementia (n=88,523, 5.1%), cancer (n=389,754, 22.4%), and all other conditions (n=1,256,873, 72.5%).

Measurements: Site of death was identified as the hospital, home, nursing home, or another location as recorded on the death certificate. In each state, the proportion of hospital deaths in the three decedent groups was categorized as high, medium, and low based on terciles of the national distribution. Using multivariate analyses, associations between state-level variables and the proportion of deaths occurring in the hospital in each state were examined.

Results: The majority of dementia-related deaths in the United States occurred in nursing homes (66.9%). In contrast, most older persons with cancer died at home (37.8%) or in the hospital (35.4%). The hospital was the most common site of death for all other conditions (52.2%). The state-specific proportion of dementia-related deaths occurring in hospitals varied from 5.0% to 37.0% across the nation and was in the highest tercile in 18 states. Hospital death rates related to cancer and all other conditions were in the highest tercile in 14 of these 18 states. After multivariate adjustment, states with a greater number of hospital beds per 1,000 persons and a lower percentage of decedents aged 85 and older were more likely to have a higher proportion of hospital deaths in all three decedent groups. Additionally, in the dementia cohort, states with fewer nursing home beds had a greater proportion of hospital deaths.

Conclusion: The majority of older Americans whose underlying cause of death is attributable to dementia on their death certificate die in nursing homes. State-level factors, including the availability of hospital and nursing home beds and the age of decedents in the population, explain, in part, the wide state-to-state variability in the proportion of dementia-related deaths occurring in the hospital.