Risk assessment of rural intersections based on predictive modelling

Rural intersections present unique road safety issue: they are numerous and typically have low crash counts. Hence, crash history can be a poor predictor of future safety performance. Given that 10% of all Victorian severe crash injuries occur at rural intersections, their effective prioritisation for improvement is a priority. This paper explores prioritisation of rural intersections based on severe injury estimates using three alternative approaches. The findings show that predictive models based on traffic flow, road hierarchy, intersection type, and maximum speed limit optimise both prioritisation logic and ease of applicability. These findings may be useful in development of future safety infrastructure investment programs.