D. Eustace, M. Almothaffar, P.W. Hovey, F. Samandi
Pages: 141-152
Abstract
The Ohio Department of Transportation (ODOT) began installing median cable barriers in 2003 along interstate medians narrower than 59 ft. The central goal of this effort was to reduce the occurrence of cross-median crashes (CMCs), which were concerning due to their frequency and the severe injuries they caused. A CMC occurs when a vehicle leaves its travel lane, enters or crosses the dividing median, and collides with vehicles moving in the opposite direction. This study analyzed data from 41 locations, totaling 201 miles of installed median cable barriers from 2009 to 2014. These locations experienced 2,498 median-related crashes before and after installation. The study involved reviewing police reports to identify target crashes and how vehicles interacted with the barriers. The study found that median cable barriers were effective in stopping vehicles from breaching the barrier; 95.4% of all median cable barrier crashes did not penetrate the barrier, meaning the vehicles were stopped or bounced back by the cables. The findings of overall statewide CMC reduction after the installation of median cable barriers, compared to the period before installation, are based on safety effectiveness computed by the Empirical Bayes (EB) before-after study method. The safety effectiveness of Ohio’s statewide cable barriers on CMCs was found to be 73.9% for total crashes, 80.4% for fatal and injury crashes, and 80.1% for fatal, incapacitating, and non--incapacitating injury crashes combined. Based on the Highway Safety Manual’s procedure, Crash Modification Factors (CMFs) were quantified as a result of a safety effectiveness evaluation. Therefore, the estimated CMFs for median cable barriers installed in Ohio’s Interstate system in reducing CMCs are 0.261 for total crashes, 0.196 for fatal and injury (FI) crashes, and 0.199 for fatal, incapacitating, and non-incapacitating injury (KAB) crashes.
Keywords: median cable barriers; cross median crashes; Safety Performance Functions; Crash Modification Factors; Empirical Bayes; before-after evaluation