A spatio-temporal analysis of the impact of congestion on traffic safety on major roads in the UK

A spatio-temporal analysis has been conducted aiming to explore the relationship between traffic congestion and road accidents based on the data on the M25 motorway and its surrounding major roads in England during the period 2003–2007. It was hypothesised that increased traffic congestion may be beneficial to road safety as the number of fatal/killed and serious injury (KSI) accidents would be less due to the low average speed when congestion is present. If this is confirmed, then it poses a potential dilemma for transport policy makers. A series of classical count outcome models (random-effects negative binomial models) and spatial models using a full Bayesian hierarchical approach have been developed in this study in order to examine whether congestion has any effect on the frequency of accidents. The results suggest that increased traffic congestion is associated with more KSI accidents and traffic congestion has little impact on slight injury accidents. This may be due to the higher speed variance among vehicles within and between lanes and erratic driving behaviour in the presence of congestion. In addition, traffic speeds even within congested situations are likely to be relatively high on major roads compared to other parts of the road network. Certain strategies are then proposed to optimise traffic flow which would be beneficial to both congestion and accident reduction.