Background Pedestrian incidents represent an increasing proportion of serious injuries resulting from motor vehicle collisions in Canada. However, few studies have examined the effect of pedestrian crossing location in urban areas on injury severity. The objective of this study was to investigate the relationship between pedestrian–motor vehicle collision injury severity and crossing location.

Methods This study was a population-based analysis of police-reported pedestrian collision data. The study group was pedestrian collisions from 1 January 2000 to 31 December 2009 in Toronto. Main outcome measures were a binary indicator of severe injury, and a four-level categorical variable of injury severity. The exposure variable was crossing at mid-block with no traffic control compared to signalised intersection. Analysis was via binary and multinomial logistic regression models to estimate ORs of injury severity with 95% CIs.

Results The analysis included 9575 pedestrian–motor vehicle collisions, of which 7325 occurred at signalised intersections when crossing and 2230 occurred at uncontrolled mid-block locations when crossing without right of way. Uncontrolled mid-block collisions resulted in greater injury severity when controlling for road type. The odds of severe injury were 1.75 (95% CI 1.07 to 2.86) for children, 2.55 (95% CI 2.13 to 3.05) for adults and 1.68 (95% CI 1.23 to 2.28) for older adults. The odds of death at uncontrolled mid-block crossings were 4.97 (95% CI 3.11 to 7.94) in adults and 3.49 (95% CI 2.07 to 5.89) in older adults.

Conclusions Crossing at uncontrolled mid-block locations resulted in greater injury severity compared with crossing at signalised intersections. This has important implications for pedestrian behaviour and traffic environment design and emphasises the need for safe pedestrian crossings on urban roads.