Abstract
Bends in roads are known to cause road-traffic crashes, but do areas with many road bends have more collisions than areas with straighter roads? A geographical information system was used to generate indicators of average road curvature from a road-network dataset of England and Wales at the local-authority district level. The indicators were the number of bends per kilometre, the ratio of road distance to straight distance, the proportion of road lengths that were straight, the cumulative angle turned per kilometre and the mean angle of each bend. Generally the five measures were associated. Road curvature was highest on minor roads and least on major roads, and metropolitan districts had straighter road networks than nonmetropolitan districts. Counts of the number of road-traffic crashes resulting in fatalities, serious injuries, and slight injuries in each district were obtained from police ‘Stats 19’ records. The association between each of the curvature measures and the number of fatal, serious, and slight collisions in each district was determined by negative binomial regression analysis. Collision numbers were negatively related to road curvature after adjusting for other risk factors, so districts with straighter roads had more crashes. The cumulative angle was the curvature measure most strongly related to fatal road crashes. An increase of 1° per km was associated with approximately a 0.5% reduction in crashes, enough to explain more than a two fold difference in collision rates over the range of the data. Separate analysis of crashes on major roads, ‘B’ class roads, and minor roads confirmed the conclusion. Although individual road bends may be hazardous, these results suggest that road curvature at the district scale is protective.
Get full access to this article
View all access options for this article.