Evaluating the effectiveness of the Safety Improvement Program in Saskatchewan using an observational before–after study with the full Bayes approach

Abstract

A key step in highway safety management is to determine whether the frequency and (or) severity of collisions has been reduced after implementing a specific improvement program. This research focused on evaluating the safety performance of 50 sites that have been improved under the Saskatchewan Ministry of Highways and Infrastructure’s Safety Improvement Program (SIP). SIP projects were designed to reduce the frequency and severity of collisions on provincial highways in rural areas through the implementation of different safety countermeasures. The methodology adopted was an observational before–after study using the full Bayes approach. The results showed that SIP projects reduced total collisions by 14.8% and severe (fatal-plus-injury) collisions by 25.4%. The reduction in property-damage-only collisions was not found to be statistically significant. Crash modification factors (CMFs) for the two most frequent SIP treatments, that is, right-turn lanes and delineation lighting at intersections, were estimated and compared with the results of the literature.