Modeling the probability of a vehicle-bicycle hit-and-run

Purpose:  Understanding bicycle-vehicle collisions that result in hit-and-run behavior is an important concern for both law enforcement and public health. If bicyclists are injured, this issue has implications for expedient access to medical care and for protection from the financial burden of associated medical costs. This study aimed to identify significant predictors of a hit-and-run, the results of which can potentially inform deterrence interventions for this type of crime. Method: Data were collected from Boston Police bicycle crash reports for 2009-2012. The data identified whether a crash was a hit-and-run and other predictor variables including road and bicyclist characteristics. The probability of a hit-and-run was fit to selected variables through logistic regression models. Best-fitting models were selected through likelihood ratio chi-squared tests comparing nested models.  Effects of the predictors were reported as odds ratios. Results: Of the1626 bike-vehicle collisions, 6% (n=93) resulted in a hit-and-run and 75% (n=70) of this subset involved an injury to the bicyclist.  Controlling for other variables, the odds of a hit-and-run were 0.38 (95% CI: 0.17, 0.87) times larger on weekends versus weekdays and 1.71 (95% CI: 1.11, 2.63) larger during daylight hours. The odds of a hit-and-run for appearing injured compared to not appearing injured among females were 3.73 times as large as the odds of a hit-and-run for appearing injured compared to not appearing injured among males. Conclusion: The probability of a hit-and-run depends on gender and perceived injury status. Findings may have implications for interventions aimed at preventing this type of crime.