Simultaneous evaluation of physical and social environmental correlates of physical activity in adults: A systematic review

Background Ecological models of physical activity posit that social and physical environmental features exert independent and interactive influences on physical activity, but previous research has focussed on independent influences. This systematic review aimed to synthesise the literature investigating how features of neighbourhood physical and social environments are associated with physical activity when both levels of influence are simultaneously considered, and to assess progress in the exploration of interactive effects of social and physical environmental correlates on physical activity. Methods A systematic literature search was conducted in February 2016. Articles were included if they used an adult (≥15 years) sample, simultaneously considered at least one physical and one social environmental characteristic in a single statistical model, used self-reported or objectively-measured physical activity as a primary outcome, reported findings from quantitative, observational analyses and were published in a peer-reviewed journal. Combined measures including social and physical environment items were excluded as they didn't permit investigation of independent and interactive social and physical effects. Forty-six studies were identified. Results An inconsistent evidence base for independent environmental correlates of physical activity was revealed, with some support for specific physical and social environment correlates. Most studies found significant associations between physical activity and both physical and social environmental variables. There was preliminary evidence that physical and social environmental variables had interactive effects on activity, although only 4 studies examined interactive effects. Conclusions Inconsistent evidence of independent associations between environmental variables and physical activity could be partly due to unmeasured effect modification (e.g. interactive effects) creating unaccounted variance in relationships between the environment and activity. Results supported multiple levels of environmental influence on physical activity. It is recommended that further research uses simultaneous or interaction analyses to gain insight into complex relationships between neighbourhood social and physical environments and physical activity, as there is currently limited research in this area.