With global super-aging, understanding the longitudinal changes in activities of daily living (ADLs) among older adults is critical for maintaining their functional independence, particularly in later life. However, evidence of longitudinal ADL patterns in Asian octogenarians remains limited. This study examined 6-year ADL transition patterns among community-dwelling older adults in Japan aged 85 years or above using latent transition analysis (LTA) to identify dynamic changes and associated factors. The present study used data from the Tokyo Oldest Old Survey on Total Health study, a prospective cohort of 538 participants aged 85 years or older. ADL was assessed using the Barthel index at baseline, 3 years, and 6 years. Through LTA, latent clusters at each time point were identified, and transition probabilities across clusters were estimated. Cluster characteristics and transitions were examined relative to demographic, physical, cognitive, lifestyle, and biological factors. Random forest analysis was used to identify predictors of longitudinal ADL decline. Three ADL clusters were identified: "Robust," "Partially Impaired," and "Impaired." Over six years, transitions to worse functional states or mortality were common, particularly among those in the "Partially Impaired" and "Impaired" clusters. Participants in the "Robust" cluster who maintained functional independence demonstrated better physical and cognitive functions, healthier lifestyles, and optimal nutritional and inflammatory profiles. Random forest analysis revealed "Stairs" and "Dressing" as critical predictors of overall ADL deterioration. The results contribute to the understanding of aging-related functional trajectories and offer actionable insights for healthcare planning and intervention development. This study was approved by the ethics committees of the Keio University School of Medicine (no. 20070047), and written informed consent was obtained from all participants. The TOOTH study was registered in the University Hospital Medical Information Network Clinical Trial Registry under ID UMIN000001842.
Osawa et al. (Tue,) studied this question.