Abstract
BACKGROUND: Physical activity (PA) is a promising non-pharmacological intervention for this population. However, few studies have investigated their PA trajectories, influencing factors, and their relationship with health outcomes. AIMS: The aim was to identify latent trajectories in PA and their determinants in older adults with mild cognitive impairment (MCI) or dementia, as well as to assess the associations between PA trajectories and health outcomes based on the capability-opportunity-motivation behavior model. METHODS: This is a cohort study. Data were obtained from a national cohort study and included participants aged 60 years and older with MCI or dementia. PA trajectories were identified using group-based trajectory modelling. Multinomial logistic regression was conducted to identify the predictors of PA trajectories. Linear regression models were used to assess the associations between PA trajectories and health outcomes. This study adhered to the STROBE checklist for reporting. RESULTS: Three distinct PA trajectories were identified: high-decreasing and rebound class (9.34%), moderate-decreasing class (10.31%), and low-increasing class (80.34%). The logistic regression showed that age, sex, education level, body mass index, residence, depressive symptoms, mobility activities of daily life score, frequency of social activities score were PA predictors. Adjusting for sociodemographic variables, only the high-decreasing and rebound class remained significantly associated with worse self-rated health. DISCUSSION: This study revealed three PA trajectories among older adults with MCI/dementia. Besides sociodemographic variables, addressing physical function and mental health, providing social support are vital for promoting PA in this population.