Associations Between Built Environment Measures and Physical Activity among Insufficiently Active Adults: Preliminary Insights from a Multi-Level Trial

Background: Physical activity (PA) is critical to preventing chronic disease, yet most US adults do not meet the recommended weekly amount of aerobic PA. The built environment (BE) is an important factor influencing PA behavior, but limited knowledge exists on how it influences PA behavior among insufficiently active adults. To fill this gap, we conducted a secondary analysis using baseline data from the Steps for Change clustered randomized controlled trial to determine how various BE attributes were associated with multiple PA outcomes among such adults residing in or near ten affordable housing sites in the San Francisco Bay Area. We hypothesized (1) that site-level measures of PA resource features and amenities would be positively associated with PA behaviors while incivilities would be negatively associated and (2) that individual-level perceptions of walkability would be positively associated with PA behaviors.

Methods: Researchers measured BE attribute data at each housing site (n=10) using the Physical Activity Resource Assessment (PARA), a Likert-scale instrument that scores the quality of the features, amenities, and incivilities of each PA resource (e.g., parks, fitness clubs). Participants (n=255) responded to the abbreviated Neighborhood Environment Walkability Survey (NEWS-A), which measures and scores perceptions of the BE, which includes domains selected a priori (e.g., street connectivity, access to services). Multiple PA behaviors—walking, moderate-to-vigorous PA (MVPA), and total PA (min/week)—were estimated using the validated Community Healthy Activities Model Program for Seniors (CHAMPS) questionnaire. Site-level Spearman’s correlations were examined between PARA quality scores and mean PA outcomes. Linear regression was used to examine associations between domains of NEWS-A and individuals’ PA behaviors—with models including terms for housing site and, in the adjusted models, sociodemographic characteristics.

Results: In the site-level analysis, MVPA was positively correlated with the PARA features (ρ=0.59) and amenities (ρ=0.12) quality scores and negatively correlated with the incivilities score (ρ=-0.62), while walking was weakly negatively correlated with incivilities (ρ=-0.08). At the individual level, every one-point increase in the aesthetics score from NEWS-A was associated with a 57.37 min/week increase in walking (β=57.37 [95% CI: 20.84, 93.91], p=0.002) in the adjusted model for the entire sample. Moreover, a one-point increase in street connectivity was associated with a 24.31 min/week increase in MVPA (β=24.31 [95% CI: 3.22, 45.41], p=0.02).

Conclusions: Our findings indicate that site- and individual-level measures of the BE, such as the quality of neighborhood PA resources, aesthetics, and street connectivity, may play a role in promoting PA behavior. Future research with repeated measures should examine how such differences in the BE influence the effectiveness of PA interventions.