Risk factors for transportation injury and mortality in the Canadian Census Health and Enviroments Cohorts (CanCHEC): a descriptive analysis

Author(s): Branion-Calles, Rothman, Winters, Harris

Slidedeck Presentation:

Slide deck link



Canada has data gaps that limit the measurement population-based transportation injury risks, especially for active transport modes. Police data underreport bicycling and pedestrian injury and are disproportionately composed of incidents with motor vehicles. Additionally, these datasets may not have detailed sociodemographic information associated with injured persons and will not capture any measurement of exposure. To address these gaps, this study will leverage a powerful new secondary data linkage: the Canadian Census Health and Environment Cohort (CanCHEC). CanCHEC consists of a series of population-based cohorts that link long-form census data to hospitalization data from the Discharge Abstract Database and mortality data from Vital Statistics, enabling detailed stratification of transportation injuries by sociodemographic and travel behaviour information captured in the Census.


The goal of this research is to leverage a large population-based linked injury dataset to describe differences in 1) hospital admissions and 2) mortality risks between different population groups.


We combined 4 cohorts to analyze mortality (1996, 2001,2006, 2011) and two to analyze hospitalization (2006, 2011). We used cox proportional hazards regression modelling to quantify the relationship between three different transportation modes and the likelihood of (i) hospitalization and (ii) a fatal injury. More specifically, for each outcome we separately modelled injuries from (i) bicycling, (ii) walking, or (iii) being a motor vehicle occupant. Independent variables included main mode of commute at baseline, age, gender, low-income status, racialization, and recent immigration.


For both the hospitalization and mortality outcomes, there was a significantly higher risk of injury as an occupant of a motor vehicle for cohort members whose main mode of commute was by car, truck or van, were aged 15-24, a man, low-income and self-identified as Indigenous. For bicycling injuries, the highest risk groups included those who indicated their main mode of commute was by bicycle, walking or other modes (public transit, motorcycle, scooter or moped), men, low-income and those who self identified as a non-visible minority. Finally, the highest risk groups for pedestrian injuries included participants who indicated their main mode of commute was by bicycle, walking or other modes (public transit, motorcycle, scooter or moped), men, low-income and self identified Indigenous persons.


Our results are consistent with previous research in both Canada and around the world which has found that in industrialized nations, transportation injury occurs disproportionately within disadvantaged populations, including low-income populations as well as Indigenous people. We also find that the Census main mode of commute to consistent with expected trends (i.e commuting by bicycle at baseline correlated with high risk of bicycling injury etc.), suggesting that this information does capture differences in exposure. These data do not allow for capturing changes in commute behaviour over time.


This research found that there are significant differences in the likelihood of experiencing severe road traffic trauma across different modes of travel and between broad population groups in Canada. Further more detailed inquiries into inequities in road trauma at both a geographic and individual level should be conducted.