Assessing School Neighbourhood Environments for Active Transport to Secondary Schools in Dunedin, New Zealand

Background: Adolescent active transport to school (ATS) is influenced by demographic, social, environmental and policy factors. Yet, the relationship between school neighbourhood built environment and adolescents’ ATS remains largely unexplored. This observational study extended the Built Environment and Active Transport to School (BEATS) Study to examine associations between observed, objectively-measured and perceived school neighbourhood built environments and rates of ATS across all twelve Dunedin (New Zealand) secondary schools. Methods: Microscale Audit of Pedestrian Streetscapes Global (MAPS Global) audits were physically conducted within a 0.5 km street-network buffer-zone of all schools. Sub-scale and overall summary scores were calculated. Intersection density, residential density, land use mix and walkability of school neighbourhood built environments were computed using Geographic Information Systems (GIS) spatial analysis over several street-network buffer zones (0.5 km, 1.0 km, 1.5 km and 2.25 km). School-specific ATS rates and perceptions of the school route from adolescents living ≤2.25 km from school (BEATS Study data: n=473; age: 15.2±1.4 years; 56.2% female) complemented analyses. Data were analysed using Pearson’s Product Moment correlations and Spearman’s Rank correlation coefficients. Results: School-level ATS rates were not significantly correlated with MAPS Global, GIS measures (0.5 or 2.25 km buffer zones) or adolescents’ perceptions of the school route. MAPS Global pedestrian design sub-scale (0.5 km: r=0.62), GIS-determined intersection density (0.5/2.25 km: r=0.77/r=0.64), land use mix (2.25 km: r=0.64) and walkability index (0.5/2.25 km: r=0.86/r=0.73) positively correlated with adolescents’ perception of dangerous crossings. MAPS Global pedestrian infrastructure sub-scale (0.5 km: r=-0.66) negatively correlated with walking safety concerns. Intersection density positively correlated with cycling safety concerns (0.5 km: r=0.68) and perceived cycle lane absence (0.5 km: r=0.65), while land use mix positively correlated with adolescents’ perception of high traffic volume (2.25 km: r=0.72). GIS-determined walkability index positively correlated with walking/cycling safety concerns (0.5 km: cycling: r=0.72; 2.25 km: r=0.68/r=0.66), high traffic volume (0.5 km: r=0.73), and absence of footpaths (0.5 km: r=0.62) and cycle lanes (2.25 km: r=0.71) (all p