Predictors of the Location of Death Among Patients Under Hospice Care

Background/Purpose: It has been consistently reported that the majority of people, including terminally-ill patients, prefer to die at home if possible. However, in 2015 the percentage of people who died at home was 29.7%, and those who died in hospital was 36.7%. In order to compare these statistics from the total population with hospice patients, this study analyzed patients who died in 2015 while receiving care from the largest hospice agency in the DC metro region and examined their location of death. This study also investigated factors associated with the location of death among hospice patients. The significance of the study includes the large sample size that allows a thorough analysis of hospice patients’ location of death by comparing their demographic/socioeconomic factors.

Methods: This study used routinely collected patient data in a hospice agency focusing on the data of those who died in 2015 and whose locations of death were recorded (N=4,299). The dependent variable was a location of death, which had five categories: assisted living facilities/hospice house, home (reference group), hospital, nursing home, and hospice inpatient unit. Independent variables were age at the time of death, sex, race, income, the duration of hospice care, and cancer diagnosis. Descriptive statistics were used to summarize chacteristics of diseased hospice patients. Multinomial logistic regression was conducted to identify factors associated with the location of death.

Results: A significantly smaller proportion of hospice patients (4.3%) died in the hospital. In contrast, 49.9% died at home, 19.1% in a hospice inpatient unit, 15.9% in a nursing home, and 10.8% in an assisted living facility, including a residential hospice house. Some of the highlights of the findings from multinomial logistic regression include that a cancer diagnosis (p<0.001), a longer stay in hospice care (p<0.001), living in a higher income neighborhood (p<0.001), an older age (p<0.001), and a person of color (p<0.001) are associated with lower likelihood of death in the hospital than at home.

Conclusions and Implications: Patients who die in the hospital tend to receive more intrusive, rigorous treatment than those who die anywhere else. Such aggressive treatment is often painful without offering much benefit to especially terminally-ill hospice patients. Previous studies also found the association between emergency room visits, hospitalization, and death in the hospital among hospice patients. The current study findings indicate that certain characteristics of hospice patients are vulnerable to hospital death, including non-cancer diagnosis, a shorter stay in hospice care, living in a lower income neighborhood, a younger age, and being white. It is important for hospice care providers to be aware of these characteristics and to improve communication with hospice patients and their families so that they will know what to do when sudden changes in symptoms happen to avoid unwanted consequences, such as hospitalization for end-of-life care and death in the hospital.