Relationships between neighbourhood physical environmental attributes and older adults’ leisure-time physical activity : a systematic review and meta-analysis

Background: Activity-friendly neighbourhood physical environments with access to recreational facilities are hypothesised to facilitate leisure-time physical activity (LTPA) among older adults (>= 65 years old). The aim of the current study was to systematically review and quantitatively summarise study findings on the relationships between physical environmental attributes and LTPA among older adults. Methods: An extensive search of literature, including grey literature, yielded 72 articles eligible for inclusion. The reported associations between seven categories of environmental attributes and six LTPA outcomes were extracted, weighted by sample size and study quality, and quantitatively summarised. Reported moderating effects of individual and environmental characteristics and neighbourhood definition were also examined. Results: We observed positive associations for walkability (p = 0.01), land-use mix-access (p = 0.02) and aesthetically pleasing scenery (p < 0.001) with leisure-time walking. For leisure-time walking within the neighbourhood, evidence was found for positive associations with land-use mix-access (p = 0.03) and access to public transit (p = 0.05), and a negative association with barriers to walking/cycling (p = 0.03). Evidence for positive relationships between overall LTPA and access to recreational facilities (p = 0.01) and parks/open space (p = 0.04) was found. Several environmental attribute-LTPA outcome combinations were insufficiently studied to draw conclusions. No consistent moderating effects were observed for individual and environmental characteristics and neighbourhood definition. Conclusions: The observed significant relationships can be used to inform policy makers and planners on how to (re) design neighbourhoods that promote LTPA among older adults. Many environmental attribute-LTPA outcome relationships have been studied insufficiently and several methodological issues remain to be addressed