Factors Predicting Local Effectiveness of Impaired Driving Laws, British Columbia, Canada.

Abstract

In 2010, British Columbia introduced laws allowing police to suspend licenses and impound vehicles of drinking drivers based on roadside breath alcohol analysis results. This study examines regional variation in the effectiveness of the new laws and identifies local factors that may have modified the laws' effectiveness. First, we used an interrupted time series design with a negative binomial mixed model to study the change in rate of single-vehicle nighttime crashes (SVNCs; a proxy for alcohol-related crashes) in each of British Columbia's 118 police patrols after the laws were introduced. Next, the role of explanatory factors was studied using patrol-level estimated percent reduction in SVNCs as the response variable in a linear regression model with spatially autocorrelated errors that included explanatory factors as covariates. The average rate of SVNCs across police patrols was 12% lower (95% CI [9.1%, 14.8%]) following the new laws, and all police patrols included in our model had estimated reductions in SVNCs. Reductions in SVNCs were positively associated with the number of alcohol-serving establishments per capita and with socioeconomic status and were negatively associated with baseline rates of traffic citations. The laws were associated with improved road safety across the province but were least effective in more socioeconomically disadvantaged regions. Increased effectiveness in regions with more alcohol-serving establishments may have indicated a specific deterrent effect on drinking drivers who frequent these establishments. The laws were also more effective in regions with lower baseline traffic law enforcement (i.e., lower citation rates), perhaps because media reports about the immediate roadside penalties regime created unease in drinking drivers who previously knew that their risk of punishment for traffic law violations was low.