Identifying Motorist Risk Factors in Youth Bicycle-Motor Vehicle Collisions
Author(s): Pitt, Graff, Howard, Nettel-Aguirre, McCormack, Owens, Anderson, Rowe, Hagel
Slidedeck Presentation Only (no paper submitted):
Abstract:
Bicycle-related injuries are among the most common injuries during recreational activities for children in Canada. Serious and fatal injuries most commonly result from bicycle-motor vehicle (BMV) collisions. Factors associated with BMV collisions need to be identified to develop effective primary prevention strategies. To date, studies have typically focused on youth (<18 years of age) risk factors rather than driver or environment BMV collision risk factors. To undertake a case-control study design that will determine the driver characteristics contributing to youth BMV collisions. We used Edmonton and Calgary police collision report data from the years 2010-2014. From these data, we identified 423 youth BMV collisions and 423 drivers (i.e., cases) involved in those collisions. The controls were drivers who, over the same period, were involved in collisions but deemed not-at-fault using an automated culpability analysis tool (n=244,442). This control selection uses the quasi-induced exposure method, which indicates that drivers who are not-at-fault in collisions are representative of the typical driver (source population). Descriptive statistics, including proportions, median and interquartile range (as appropriate) were used to describe the characteristics of the two groups involved. Descriptive results indicate that 59.2% of motorists in the control group and 55.9% of motorists involved in BMV collisions were male. Control drivers had a median age of 40 years (IQR=13), and cases had a median age of 42 years (IQR=13). There appears to be a difference in the proportions of cases and controls who drove passenger vehicles (53.1% vs. 49.2%), as well as the proportion of cases and controls who drove pick-up trucks (10.8% vs. 15 .8%). Lower proportion of control motorists were impaired by alcohol (0.06% vs. 0.28%) than of cases. Most BMV collisions occurred between 18:00hrs and 24:00hrs (59.2%), followed by events occurring between 06:00hrs and 12:00hrs (19.7%). More controls were driving between 18:00hrs and 24:00hrs (50.7%) and between 06:00hrs and 12:00hrs (33.1%). Some driver characteristics appear to be different between cases and controls; however, we will continue our analysis using unconditional logistic regression to assess the potential effects of the considered factors. A strength of our study is the large sample size of objectively and consistently recorded data that was facilitated by the use of an automated culpability analysis tool. As these are administrative data, we will assess proportion of missing data and the way by which it is missing. If appropriate, we may use multiple imputation analysis to better understand the effect of missing data in traffic collision reports. Although the drivers appear similar, in terms of demographics, there are some differences in vehicle, time of day and exposure to alcohol. This study helps to highlight often ignored motorist characteristics in youth BMV collisions. In doing so, we hope to inform primary prevention strategies for the motorists and environment. Culpability analysis tools and quasi-induced exposure techniques are typically applied to motorists to identify transient exposures; however, we have shown that these techniques are possible in vulnerable population collisions.