Factors associated with change in objectively measured physical activity in older people – data from the physical activity cohort Scotland study

Clare L. Clarke, Falko F. Sniehotta, Thenmalar Vadiveloo, Ishbel S. Argo, Peter T. Donnan, Marion E. T. McMurdo, Miles D. Witham

Cross-sectional relationships between physical activity and health have been explored extensively, but less is known about how physical activity changes with time in older people. The aim of this study was to assess baseline predictors of how objectively measured physical activity changes with time in older people. Longitudinal cohort study using data from the Physical Activity Cohort Scotland. A sample of community-dwelling older people aged 65 and over were recruited in 2009-2011, then followed up 2-3 years later. Physical activity was measured using Stayhealthy RT3 accelerometers over 7 days. Other data collected included baseline comorbidity, health-related quality of life (SF-36), extended Theory of Planned Behaviour Questionnaire and Social Capital Module of the General Household Survey. Associations between follow-up accelerometer counts and baseline predictors were analysed using a series of linear regression models, adjusting for baseline activity levels and follow-up time. Follow up data were available for 339 of the original 584 participants. The mean age was 77 years, 185 (55%) were female and mean follow up time was 26 months. Mean activity counts fell by between 2% per year (age < =80, deprivation decile 5-10) and 12% per year (age > 80, deprivation decile 5-10) from baseline values. In univariate analysis age, sex, deprivation decile, most SF-36 domains, most measures of social connectedness, most measures from the extended Theory of Planned Behaviour, hypertension, diabetes mellitus, chronic pain and depression score were significantly associated with adjusted activity counts at follow-up. In multivariate regression age, satisfactory friend network, SF-36 physical function score, and the presence of diabetes mellitus were independent predictors of activity counts at follow up after adjustment for baseline count and duration of follow up. Health status and social connectedness, but not extended Theory of Planned Behaviour measures, independently predicted changes in physical activity in community dwelling older people.