"Don't wait for accidents to happen"

October 2005. Helsinki

Traffic safety effects of roundabouts – a review with emphasis on bicyclist’s safety

DANIELS, Stijn; WETS, Geert
The paper reviews research results about the design of roundabouts and their effects on traffic safety. Attention is given to specific accommodation for bicyclists and its effects on traffic safety.
Firstly the safety of roundabouts is regarded from a theoretical point of view. A roundabout can be an appropriate type of intersection design because of several reasons. Roundabouts reduce the number of conflicts between road users (e.g. from 32 to 8 at four-legged crossroads). They also reduce vehicle speeds because of the circulatory movement around the traffic island.
The most useful technique to evaluate the safety effects of roundabouts is the before- and after study. Few studies seem to account for effects of general trends in traffic safety and regression-to-the-mean. Other research techniques have been applied, but do have their limitations.
Several studies delivered evidence about the safety performance of roundabouts. Safety effects of roundabouts are high, when injury accidents are considered. For example, in Flanders-Belgium an average decrease of 34% of the number of injury accidents was noticed after construction of a roundabout. Effects are higher on the number of accidents with seriously injured in comparison with accidents causing slight injuries. Effects on accidents with property damage only are highly uncertain and might even be negative.
But the effects are not equal for all road users. Particularly for bicyclists, safety effects seem to be smaller and are not really proven. Depending on the design of the bicyclist’s facilities there seem to exist differences in safety level for bicyclists. Separated bicycle lanes (the bicyclist drives around the roundabout on a separated bicycle lane) are the safest. Designs with bicycle lanes adjacent to the carriageway or ‘mixed traffic’ designs (cyclists and motorised traffic both conducted on the carriageway) perform less. Special attention is needed for conflicts between bicyclists and motorised road users when cars/trucks are leaving the roundabout and come into conflict with bicyclists who are driving around the roundabout.
It is suggested to apply methodologically reliable techniques to evaluate safety effects for bicyclists too.