| Summary: | Abstract Background Pedestrian related deaths have recently been on the rise in Canada. The effect of changing posted speeds on the frequency and severity of pedestrian motor vehicle collisions (PMVC) is not well studied using controlled quasi-experimental designs. The objective of this study was to examine the effect of lowering speed limits from 40 km/h to 30 km/h on PMVC on local roads in Toronto, Canada. Methods A 30 km/h speed limit on local roads in Toronto was implemented between January 2015 and December 2016. Streets that remained at a 40 km/h speed limit throughout the study period were selected as comparators. A quasi-experimental, pre-post study with a comparator group was used to evaluate the effect of the intervention on PMVC rates before and after the speed limit change using repeated measures Poisson regression. PMVC data were obtained from police reports for a minimum of two years pre- and post-intervention (2013 to 2018). Results Speed limit reductions from 40 km/h to 30 km/h were associated with a 28% decrease in the PMVC incidence rate in the City of Toronto (IRR = 0.72, 95% CI: 0.58–0.89). A non-significant 7% decrease in PMVC incidence rates were observed on comparator streets that remained at 40 km/h speed limits (IRR = 0.93, 95% CI: 0.70–1.25). Speed limit reduction also influenced injury severity, with a significant 67% decrease in major and fatal injuries in the post intervention period on streets with speed limit reductions (IRR = 0.33, 95% CI: 0.13–0.85) compared with a 31% not statistically significant decrease in major and fatal injuries on comparator streets (IRR = 0.69, 95% CI: 0.37–1.31). The interaction term for group and pre-post comparisons was not statistically significant (p = 0.14) indicating that there was no evidence to suggest a pre-post difference in IRRs between the intervention and comparator streets. Conclusions Declines in the rate of PMVC were observed on roads with posted speed limit reductions from 40 km/h to 30 km/h, although this effect was not statistically greater than reductions on comparator streets.
|
|---|
