Exploring regional differences in cyclist safety at roundabouts: A comparative study between the UK (based on Northumbria data) and Belgium.

The level of safety for cyclists at roundabouts may vary according to national differences not only in the design itself but also sociodemographic, cyclist and driver behaviour as well as environmental factors. This paper investigates the national influence on cyclist casualty severity at roundabouts by comparing the United Kingdom (using Northumbria as a representative sample) and Belgium. The data included speed limits, socio-demographic characteristics, environmental conditions and driver/cyclist behaviour-related contributory factors. First, a logistic regression analysis for the UK data, including 864 cyclist casualties, was carried out. Increasing the speed limit by ten units (for example 30mph to 40mph) increased the probability of a cyclist being killed or seriously injured by 10%. A cyclist casualty was more than three times more likely to be killed or seriously injured (the odds ratio is 3.02) where sudden braking was recorded as a contributory factor. Second, a separate logistic regression analysis for Belgium was conducted. Cyclists ignoring the priority at roundabouts increased the probability of a fatal or seriously injured collisions (the odds ratio is 2.71). Comparing the individual analysis for both countries, the influence of cyclist age was consistent. Each one-year increase in cyclist age increases chance of being killed or seriously injured as opposed to not being killed or seriously injured by 2 % (odds ratio is 1.02) in both UK and Belgium. A final comparative analysis was applied considering proxy variables for both countries. Three-way chi-square tests of independence showed that all non-behavioural variables (i.e. sociodemographic characteristics, speed limit, and environmental conditions) were found to be statistically different between UK and Belgium for both slight and killed and seriously injured casualties. This suggests that driver/cyclist interaction and behaviour in the two countries is generally similar whilst speed limits, the sociodemographic characteristics of cyclists and environmental conditions are specific for each country. The third part of the logistic regression analysis suggested that the country residual was highly statistically significant. This indicates that there are some statistically significant differences with respect to the characteristics of the two regional datasets used in the analysis.