Association between place of death and regional variation in senility mortality in Japan: a nationwide ecological study

Abstract Background In recent years, senility mortality has increased substantially in Japan, accounting for 12.9% of all deaths in 2024. Although senility (International Classification of Diseases, Tenth Revision, code R54) is often regarded as a non-specific cause of death in international mortality statistics, its regional distribution and associated factors remain unclear. In this study, we examined regional variation in senility mortality and its association with place of death in Japan. Methods This nationwide ecological study included all 335 secondary medical areas in Japan. Standardized mortality ratios for senility were calculated separately for men and women using mortality data from 2020 to 2024. Global Moran’s I was used to evaluate spatial autocorrelation. Associations of place of death and regional contextual characteristics with senility mortality were examined using Spearman’s rank correlation analysis and negative binomial regression models adjusted for regional contextual characteristics. Results Substantial regional variation in senility mortality was observed across Japan. Significant spatial clustering was identified for both men (Moran’s I = 0.360, p = 0.001) and women (Moran’s I = 0.338, p = 0.001). Higher proportions of deaths occurring in long-term care facilities and at home were independently associated with higher senility mortality. Conversely, a higher proportion of deaths occurring in medical institutions was associated with lower senility mortality. These associations remained significant after adjustment for regional contextual characteristics. Conclusions Senility mortality exhibited substantial regional variation and significant spatial clustering across Japan. The consistent associations between senility mortality and place of death suggest that senility mortality may reflect regional population age structure and differences in end-of-life care and care delivery systems. These findings suggest that senility mortality may provide additional information about population age structure and end-of-life care in super-aged societies.