Multidisciplinary Investigation Of Rural Intersection-Related Crashes And Crash Injuries: Characteristics, Risk Factors And Prevention

Abstract

Introduction: Rural intersections disproportionally contribute to crashes in rural United States (U.S.). Few available studies have examined potential causal associations between important intersection-relevant exposures and crash outcomes. Additionally, the efficacy of an advanced Rural Intersection Conflict Warning System (RICWSs), implemented at high-risk rural two-way stop-controlled intersections, has been largely unknown. Methods: This multidisciplinary study included: 1) an ecological epidemiologic analysis to identify the magnitude of and risk factors for rural intersection-related crash rates (per 100 million Annual Average Daily Traffic — AADT) by the level of injury severity, among various intersection characteristics; 2) a quasi-experimental epidemiologic investigation to evaluate the efficacy of the RICWSs for crash reductions, among 56 intersection risk-sets, each containing one RICWS treatment and three relevant control intersections; and 3) a driving simulator study to evaluate how different RICWS designs impacted drivers’ behaviors among 120 participants. The study population involved a representative sample of rural intersections in Minnesota (n = 4,172). Adjusted crash rate-ratios were obtained from the Generalized Estimating Equations (GEE), while adjusting for relevant confounders. Potential causal associations between RICWSs and various crash rate outcomes (per intersection per year) were determined, using Difference-in-Difference (DID) and Difference-in-Difference-in-Difference (Triple-D) analytical methods. Various regression models, applying the GEE, enabled efficacy determination of each RICWS design and an aggregated RICWS treatment effect, averaged across all simulated RICWS designs, among different levels of drivers' ages, mainstream traffic volume, and intersection visibility. Results: Rural intersection-related crash rates were significantly associated with various important intersection relevant risk factors by the level of injury severity. Minnesota RICWSs demonstrated a potential significant protective effect against targeted right-angle crashes, continuously, over time, post-treatment (p < 0.05). The original RICWS design potentially improved gap acceptance performance among drivers; however, it also increased the risk of stop-sign violations (Risk Ratio = 2.2, 95% Confidence Interval (CI) = 1.03 to 4.64). Conclusions: This study serves as a basis to provide insight into prioritizing the prevention of rural intersection-related crashes and associated injuries. It has significant research relevance to mitigate the overall mortality and morbidity of motorists, including the working population, on the U.S. rural roads.