Child pedestrian risk and social equity: spatial distribution of roadway safety features in Toronto, Canada
Author(s): Rothman, Cloutier, Manaugh, Howard, Macpherson, Macarthur
Slidedeck Presentation Only (no paper submitted):
Abstract:
Transport planning must take into account the needs of pedestrians in urban areas. Walking must be a safe and convenient mode of transport. Investments in traffic calming infrastructure and other road design features are generally made to improve pedestrian safety as well as to enhance neighbourhood walkability. Pedestrian motor vehicle collisions (PMVC); however, remain common, and occur at even higher frequency in lower income neighborhoods. Spatial distribution of the risk of PMVC is an important transport equity issue. While the literature has identified specific road environment risk and protective features, there has been little work examining the social inequities in the distribution of road environment features. The aims of this research were to 1) compare child PMVC rates in low versus high income census tract (CT) clusters and 2) determine the distribution of roadway environment features related to child pedestrian safety in low versus high income clusters, in Toronto, Canada. Spatial cluster detection by census tract identified low and high income clusters using data from the 2006 Canadian census. Police reported data of 2185 PMVCs involving children between the ages of 5-14 over the years 2001-2010 were mapped along with the locations and densities of speed humps, crossing guards, missing sidewalks, one-way streets and local roads. Multiple logistic regression was used to examine the relationship between roadway environment features and low versus high income cluster location. There were 524 census tracts, with fifty eight (11%) in high and 44 (8%) in low income clusters. Collision rates were almost 6 times higher in low income clusters. In the multivariate analysis, for every km/10 km road increase of speed humps there was 65 % lower odds of being in a low income cluster (Table 2). For every km/10 km road increase local roads there was a 38% lower odds of being in a low income cluster. For every single increase in crossing guards/10 km road there was a 43% greater odds of being in a lower income cluster. Inequities in child PMVC risk and road environment features that are related to child pedestrian safety were demonstrated in this study. Fewer lower speed local roads and speed humps in lower income areas may put children at increased risk. The inequity in spatial distribution of speed humps may due to process of request which is initiated by the community. A request-based may not be the most equitable method of identifying need for a proven traffic safety feature, which may favour higher income communities that tend to have greater political engagement. Greater numbers of school crossing guards may reflect more children walking to school in lower income areas, as well as attempts to ameliorate more dangerous road environments, as indicated by higher child PMVC rates in these areas. More objective methods of determining need for traffic safety features which are not request-based are warranted. These results have policy implications related to the safe design of neighbourhoods and the equitable distribution of roadway environment features to provide safe pedestrian environments.