Mortality differences in disabled older adults by place of care in Japan: nationwide 10-year results

Tomoko Ito, Mikiya Sato*, Hideto Takahashi, Chihiro Omori, Yuta Taniguchi, Xueying Jin, Taeko Watanabe, Haruko Noguchi, Nanako Tamiya

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


This longitudinal study aimed to compare 1-year mortality between users of home- and community-based services (HCBS) and residential facilities (nursing homes, group homes, and geriatric apartments) among non-hospitalized frail older adults in Japan. Using three nationwide data sources, we conducted a nationwide pooled cohort study of 1-year follow-up among certified users of long-term care insurance (LTCI) aged 65 years and older from 2007 through 2016 to compare 1-year mortality using a logistic regression model. Overall, compared to HCBS users, mortality was higher in residents in nursing homes and geriatric apartments but lower in group home residents. While mortality gradually increased over time among those in residential facilities, it remained at a level similar to that of HCBS users. Since 2006, Japan’s public health policy has been to increase end-of-life care in residential facilities. Our results indicate that this policy resulted in an increase in mortality in residential facilities, possibly due to accommodation of more severely ill people there, or a shift in their focus from transferring dying residents to hospitals to preserve the dignity of residents.

Original languageEnglish
Pages (from-to)542-559
Number of pages18
JournalJournal of Public Health Policy
Issue number4
Publication statusPublished - 2022 Dec


  • Long-term care
  • Mortality
  • Place of care
  • Residential aged care

ASJC Scopus subject areas

  • Health Policy
  • Public Health, Environmental and Occupational Health


Dive into the research topics of 'Mortality differences in disabled older adults by place of care in Japan: nationwide 10-year results'. Together they form a unique fingerprint.

Cite this