Abstract
BACKGROUND: Previous research has suggested a likely reciprocal relationship between physical function and physical activity among older adults, but few assessed those living with dementia. This study examined the bidirectional relationship between physical function and physical activity among older adults living with dementia during hospitalization and post discharge periods, and whether the relationship differs by severity of dementia. METHODS: This secondary analysis included 455 older adult patients aged 55 years and older living with dementia from a randomized clinical trial, assessed during admission, discharge, 1-, 6-, and 12-month post-discharge periods. Random intercept cross-lagged panel models (RI-CLPMs) were used to assess the bidirectional relations, controlling for age, comorbidities, admission location, length of stay, discharge location, and intervention status. RESULTS: Average age was 82.47 (SD = 8.49) and majority were female (62.6%) and White (65.3%). The average SLUMS score was 7.51 (SD = 5.90) with 77% (n = 351) having a severe level of cognitive impairment. Antecedent physical function at admission, discharge, and 1-month predicted physical activity at corresponding cross-lagged timepoint separately (range of unstandardized coefficients b's: 0.037-0.043, p's < .05); physical activity at discharge predicted physical function at 1-month (b = 0.708, p = .016). This bi-directional relationship varied by severity of dementia, appearing at the first 2 cross-lagged time points in patients with severe dementia (b's: 0.039-0.049 and 0.464-0.848, all p's < .05), but not in those with moderate dementia. CONCLUSIONS: Physical activity and physical function commonly co-occur among older adults with dementia. Intervention studies promoting both physical function and physical activity among older adults with dementia may achieve greater effectiveness when tailored to differences in dementia severity.