ACCIDENT DATA REQUIREMENTS FOR IMPROVING CYCLING SAFETY

Canadian cities have experienced a 50 to 60% increase in police-reported collisions between bicycles and motor vehicles over the past decade; along with this increase are growing problems in coping with bicycle traffic. In some locales, as many as 10% of police-reported injury-producing road accidents involve a cyclist. Even though cycling accidents are increasing, efforts to make urban areas more accommodating to cyclists are seldom based on accident experience. A thorough understanding of the nature and extent of cycling accidents is essential to develop meaningful countermeasures. The poor representation of cycling accidents in existing data sources inhibits the understanding of such accidents. Police data have traditionally been the primary information source on bicycle accidents and are generally limited to collisions involving motor vehicles. Emergency room surveys show that between 50 and 87% of these accidents do not involve motor vehicles. Thus police-reported data represent only a small fraction of cycling accidents. An analysis of police-reported data in Winnipeg, Canada, for 1990 reveals several problems: collision descriptions are inaccurate in 60% of the cases, inconsistencies in coding are common, and information on cyclist safety equipment and contributing factors is completely overlooked. Opportunities exist to improve data collection efforts and to extend data sources so that they cover all bicycling accidents. Hospital emergency room surveys can determine the nature and magnitude of cycling accidents and associated injuries. Travel surveys can be used to estimate distances cycled and to develop accident rates. Cyclist surveys can provide information on the nature of accidents, distances cycled, and accident rates.