Abstract
BACKGROUND: The Japanese government is promoting physician-led home visits as well as end-of-life care at home. However, the proportion of deaths occurring at home has remained unchanged for the past 20 years. OBJECTIVES: To report the cumulative incidence of deaths at home and to explore the factors associated with deaths at home versus other places, mainly hospitals. METHODS: This was a multicenter prospective cohort study in a primary care setting. We enrolled patients aged ≥65 years who had started to receive regular visits by family physicians from 13 facilities in and around Tokyo between February 1, 2013 and January 31, 2016. Patients were followed-up until January 31, 2017. The primary outcome measures were mortality rate and cumulative incidence of deaths at home. RESULTS: We enrolled 762 patients. Of 368 deaths, 133 occurred in the patient's home. The mortality rates at home were 137.6/1000 person-years (95% confidence interval 116.1-163.1). In cumulative incidence function, the longer duration of care at home lowers the likelihood of death at home. Multivariable multinomial logistic models showed that younger age and higher Barthel Index score reduced the likelihood of deaths at home, while receiving oxygen therapy and the presence of a full-time caregiver increased the likelihood of deaths at home relative to deaths at other locations. CONCLUSIONS: Of deceased patients, only one-third died in patients' homes. We found several factors associated with deaths at home, which appeared to reflect the readiness of patients and their families for death. Geriatr Gerontol Int 2022; 22: 1005-1012.