A 25-Year Retrospective on Bridge-Related Crashes in North Carolina: Frequencies, User Costs, and Crash Forecasting Models

Bridge management systems (BMS) rely on accurate estimates of both user and agency costs for multi-objective optimization of investments in bridge improvement projects. Costs due to vehicular crashes are the most significant component of user costs and are particularly sensitive to the number of vehicular crashes and rate of higher severity crashes, which must be accurately forecasted to facilitate reliable decision making. This paper revisits a seminal study on bridge-related crash rates and associated costs, performed using data for North Carolina bridges. This approach facilitates a unique 25-year retrospective of bridge-related crashes in a specific location, providing insights into changes in 1) bridge-related crash frequencies, 2) predicted user costs associated with bridge-related crashes, and 3) factors influencing bridge-related crashes. The average number of bridge-related crashes occurring in counties included in the study remained relatively constant, and the frequency of fatal and high-severity injury bridge-related crashes decreased significantly. Paired with locally-sourced per-crash cost data, use of updated, lower crash frequencies results in lower, more accurate, user cost predictions in the state BMS. An updated bridge-related crash prediction equation developed by statistical regression provides plausible results linking crash rates to bridge characteristics recorded in the BMS. Average daily traffic (ADT) and structure length remain influential in predicting bridge-related crashes, similar to previous findings. Average Index (BMS), a composite bridge condition rating, was found to be a significant predictor, indicating that regular and preventative maintenance should reduce bridge-related crashes. Interstate bridges are associated with lower incidences of bridge-related crashes, relative to bridges on other roadway types.