A new model to estimate pedestrian deaths from speed-related interventions

Abstract

Objective The S-shaped function of probability of pedestrian fatality with respect to impact speed of vehicle is well known in road safety literature. However, the implication of this evidence for changes in speed for a population of drivers has not been explored. Method An integrated model has been developed to estimate pedestrian fatalities resulting from changes in speed of traffic. The model uses combined relative risks of injury crash rate of drivers and probability of pedestrian fatality resulting from an injury crash. Two approaches have been used—an individual approach using probability distribution of speed, and an aggregate approach, using only mean speed. The application of the model has been demonstrated using four case studies with speed of traffic before and after an intervention. Results It is found that even small reductions in mean speed translate to large reductions in pedestrian fatalities. The risk contributed by speeding vehicles is highly disproportionate to their share in vehicle population. In a case study from Delhi (India), in which at-grade junction was replaced with a grade-separated junction, 74% of risk of pedestrian fatalities is contributed by less than 5% speeding vehicles (>50km/h). Changes in mean speed is a poor indicator of changes in injury risk if the standard deviation of the distribution also changes significantly. A surprising finding was that large variation in the S-shape of pedestrian fatality risk function has relatively small effects on overall results. Conclusions A new model has been developed to estimate changes in pedestrian fatalities resulting from changes in traffic speed. The application of the model using four speed-related interventions emphasize that deterrence of a small proportion of drivers using speed enforcement can have large implications for pedestrian safety.