Identification of hazardous road locations using ambulance data

Background: Police underreporting of traffic crashes is a substantial challenge in Denmark, as in many other countries. Alternative data sources are highly warranted for site-specific road traffic safety work. This study aimed to test the potential benefits and drawbacks of a merger of ambulance data, including crash coordinates, and hospital patient data. Methods: We tested a definition of a traffic crash based on a merger of hospital patient data and ambulance data, including prehospital medical records information and ambulance coordinates in Aalborg Municipality in the North Denmark Region of Denmark from 2016 to 2019 and presented a method of identification of hazardous intersections, road sections and local traffic areas. We compared selected police data characteristics for the same area and period. Results: Compared to police data in the same study period, we found 3.2 times as many unique crash sites through the use of combined hospital patient and ambulance data. We showed that as many as 47% of the crash sites found in the ambulance records were located outside the network of major roads, whereas 16% were located at intersections of major roads and 38% on sections of major roads. We found a certain imprecision in the identification of crash locations because some crash coordinates were physical addresses of buildings only close to the crash sites. Furthermore, we noted that the data privacy of patient data is an issue. Conclusions: Ambulance data have clear-cut potential in site-specific road traffic safety work. However, data quality assurance is necessary prior to the application of the data for black spot identification. As such, more precise coordinates are needed, as well as further methodological development and an anonymisation procedure legally approved by the authorities to make data available for traffic safety work operations.