Background: Data showing long term efficacy and sustained active lifestyle change are scant. Identifying social determinants for maintaining an active lifestyle overtime may help pinpoint barriers/facilitators for sustaining an active lifestyle change.
Methods: We evaluated the associations between social determinants of health (SDH) and changes in physical activity (PA). The analysis dataset included 3,741 participants in the Jackson Heart Study (JHS) who completed the PA questionnaire at exams 1 and 3, approximately 8 years apart. Participants were classified into 3 PA groups-poor (0 minutes/week moderate/vigorous activity), intermediate (some moderate/vigorous activity), or ideal (≥75 minutes/week vigorous or ≥150 minutes/week moderate or ≥150 minutes/week combined moderate/vigorous activity). SDH evaluated were individual level socioeconomic status (SES), social support, perceived stress and neighborhood characteristics. Other covariates included age, sex and body mass index (BMI). Unadjusted analyses followed by multivariable logistic regression analyses with backward selection were performed to explore the most significant SDH associated with changes in PA.
Results: Twenty-eight percent (28%) participants had poor PA at both exams; 35.5% had at least intermediate PA at both exams; 19.2% changed from poor to intermediate/ideal PA and 17.8% changed from intermediate/ideal to poor PA overtime. In unadjusted analyses, among participants with intermediate/ideal PA at exam 1, younger age, higher levels of SES and greater social support; as well as higher neighborhood SES, greater social cohesion, fewer neighborhood problems and violence were positively associated with maintaining intermediate/ideal PA at exam 3. Among those with poor PA at exam 1, the same factors other than social cohesion were associated with improved PA to moderate/ideal level at exam 3. Sex, perceived stress, number of PA facilities in the neighborhood or being overweight/obese at either exam was not associated with changes in physical activity levels. Exploratory multivariable logistic regression with backward selection suggested that all three individual SES indicators were important factors for maintaining moderate/ideal PA levels at exam 3; and younger age, higher education, and higher neighborhood SES were important factors for improving PA levels from poor to intermediate/ideal.
Conclusions: Individual SES indicators were significantly associated with maintaining an intermediate/ideal PA overtime; whereas younger age, higher educational attainment and higher neighborhood SES were significantly associated with changes in PA from poor to intermediate/ideal. Targeting individual education and environmental factors (e.g., neighborhood problems, violence) associated with neighborhood SES may provide most benefit both in initiating and maintaining an active lifestyle.