Abstract / Summary
Abstract Background Multimorbidity is linked to disability and mortality in later life, but functional trajectories vary among older adults with similar disease burden. We examined whether regular exercise, social participation, good sleep, and a cumulative score were associated with later ADL/IADL limitation or death among Chinese older adults with multimorbidity. Methods Using the 2008–2018 Chinese Longitudinal Healthy Longevity Survey, we constructed rolling person-period observations for three intervals. Respondents were eligible if, at interval start, they were aged 65 years or older, had ≥ 2 chronic conditions, and reported no limitations in activities of daily living (ADL) or instrumental activities of daily living (IADL). The outcome was incident ADL/IADL limitation among survivors or death before the next wave. Modified Poisson GEE models estimated adjusted risk ratios (RRs). Sensitivity analyses adjusted for health indicators, applied inverse-probability weighting (IPW) for follow-up availability, used logistic regression, and excluded mobility items closely related to exercise. Results The analytic file included 1,241 person-intervals from 1,045 respondents; 696 adverse transitions occurred. Regular exercise was associated with lower risk of the composite adverse transition (RR 0.87, 95% CI 0.78–0.96; P = 0.007). Model-standardized risks were 52.5% with regular exercise and 60.5% without it (risk difference − 8.0 percentage points, 95% CI − 13.8 to − 2.2). The association remained after health-reserve adjustment (RR 0.89), IPW (RR 0.87), and exclusion of walking, carrying, and crouching/standing items (RR 0.84). Social participation, sleep, and the resource score were not independently associated with the outcome. Associations were concentrated in IADL and community-function transitions, and interval-specific estimates suggested possible temporal heterogeneity. Conclusions Among Chinese older adults with multimorbidity who were functionally independent at interval start, regular exercise was associated with a modestly lower risk of ADL/IADL limitation or death by the next survey wave. The association remained after measured health-reserve adjustment, IPW, and exclusion of mobility items closely related to exercise. These observational findings support further study of exercise as a potential component of functional-maintenance strategies in this population, while causal interpretation remains limited.