Changes in Objectively-Determined Walkability and Physical Activity in Adults: A Quasi-Longitudinal Residential Relocation Study

Causal evidence for the built environment's role in supporting physical activity is needed to inform land use and transportation policies. This quasi-longitudinal residential relocation study compared within-person changes in self-reported transportation walking, transportation cycling, and overall physical activity during the past 12 months among adults who did and did not move to a different neighbourhood. In 2014, a random sample of adults from 12 neighbourhoods (Calgary, AB, Canada) with varying urban form and socioeconomic status provided complete self-administered questionnaire data (n = 915). Participants, some of whom moved neighbourhood during the past 12 months (n = 95), reported their perceived change in transportation walking and cycling, and overall physical activity during that period. The questionnaire also captured residential self-selection, and sociodemographic and health characteristics. Walk Scores® were linked to each participant's current and previous neighbourhood and three groups identified: walkability "improvers" (n = 48); "decliners" (n = 47), and; "maintainers" (n = 820). Perceived change in physical activity was compared between the three groups using propensity score covariate-adjusted Firth logistic regression (odds ratios: OR). Compared with walkability maintainers, walkability decliners (OR 4.37) and improvers (OR 4.14) were more likely (p < 0.05) to report an increase in their transportation walking since moving neighbourhood, while walkability decliners were also more likely (OR 3.17) to report decreasing their transportation walking since moving. Walkability improvers were more likely than maintainers to increase their transportation cycling since moving neighbourhood (OR 4.22). Temporal changes in neighbourhood walkability resulting from residential relocation appear to be associated with reported temporal changes in transportation walking and cycling in adults.