FORMULATING ALCOHOL-INFLUENCED DRIVER'S INJURY SEVERITIES IN INTERSECTION-RELATED CRASHES

Approximately one third of all traffic fatal crashes are alcohol-related in the US according to the National Highway Traffic Safety Administration (NHTSA), alcohol-related crashes cost more than $37 billion annually. Considerable research efforts are needed to understand better significant causal factors for alcohol-related crash risks and driver's injury severities in order to develop effective countermeasures and proper policies for system-wide traffic safety performance improvements. Furthermore, since two thirds of urban Vehicle Miles Traveled (VMT) is on signal-controlled roadways, it is of practical importance to investigate injury severities of all drivers who are involved in intersection-related crashes and their corresponding significant causal factors due to control and geometric impacts on flow progression interruptions. This study aims to identify and quantify the impacts of alcohol/non-alcohol-influenced driver's behavior and demographic features as well as geometric and environmental characteristics on driver's injury severities around intersections in New Mexico. The econometric models, multinomial Logit models, were developed to analyze injury severities for regular sober drivers and alcohol-influenced drivers, respectively, using the crash data collected in New Mexico from 2010 to 2011. Elasticity analyzes were conducted in order to understand better the quantitative impacts of these contributing factors on driver's injury outcomes. The research findings provide a better understanding of contributing factors and their impacts on driver injury severities in crashes around intersections. For example, the probability of having severe injuries is higher for non-alcohol-influenced drivers when the drivers are 65 years old or older. Drivers' left-turning action will increase non-alcohol-influenced driver injury severities in crash occurring around intersections. However, different characteristics are captured for alcohol-influenced drivers involved in intersection-related crashes. For example, more severe injuries of alcohol-influenced drivers can be observed around intersections with three or more lanes on each approach. The model specifications and estimation results are also helpful for transportation agencies and decision makers to develop cost-effective solutions to reduce alcohol-involved crash severities and improve traffic system safety performance.