Cyclist-motor vehicle collisions before and after implementation of cycle tracks in Toronto, Canada

Cycling, as a mode of active transportation, has numerous health and societal benefits, but carries risks of injury when performed on-road with vehicles. Cycle tracks are dedicated lanes with a physical separation or barrier between bicycles and motor vehicles. Studies on the effectiveness of cycle tracks in urban areas in North America, as well as the area-wide effects of cycle tracks are limited.

Study objectives were to examine the effect of cycle track implementation on cyclist-motor vehicle collisions (CMVC) occurring: (1) on streets treated with new cycle tracks; (2) on streets surrounding new cycle tracks in Toronto, Canada.

Intervention and outcome data were obtained from the City of Toronto. All police-reported CMVC from 2000 to 2016 were mapped. Analyses were restricted to 2 years pre- and 2 years post-track implementation. Rates were calculated for CMVC on streets with cycle tracks (objective 1) and in five defined areas surrounding cycle tracks (objective 2). Zero-Inflated Poisson regression was used to compare changes to CMVC rates before and after cycle track implementation for both objectives. All models controlled for season of collision and cycle track.

The majority of CMVC on cycle tracks occurred at intersections (75%). The crude CMVC rate increased two-fold after cycle track implementation (IRR = 2.06, 95% CI: 1.51-2.81); however, after accounting for the increase in cycling volumes post-implementation, there was a 38% reduction in the CMVC rate per cyclist-month (IRR = 0.62, 95% CI: 0.44-0.89). On streets between 151 m - 550 m from cycle tracks, there was a significant 35% reduction in CMVC rates per km-month following track implementation (IRR = 0.65, 95% CI: 0.54-0.76).

Cycle track implementation was associated with increased safety for cyclists on cycle tracks, after adjusting for cycling volume. In addition, there was a significant reduction in CMVC on streets surrounding cycle tracks between 151 m - 550 m distance from the tracks (a 'safety halo' effect), suggesting an area-wide safety effect of cycle track implementation.