Should we treat fatal and injury crashes differently for road safety treatment selection? The evidence says sometimes yes sometimes no

Two opposing views exist regarding use of fatal versus injury crashes to guide selection of road safety treatment: (1) fatal and injury crashes are equally important in guiding road engineering treatments: (2) fatal crashes should be assigned more weight than injury crashes in determining treatment priorities. Each view is adopted into policy in various countries and programs. To date the debate on which view is correct has been theoretical with proponents arguing that fatal and injury crashes do or do not differ systematically in ways relevant to road design and engineering. For example, it is argued that other factors such as age and frailty of the victims, or safety afforded by their vehicle may determine whether the crash is fatal or injury, not the specific details of the crash or the road. The critical issue is whether fatal crashes predict that other crashes (of the same type at the same location) are more severe than if no fatality has occurred (at that location and crash type). This paper empirically informs this critical debate by testing this the predictive power of fatal versus injury crashes at crash cluster locations in South Australia. Extraordinarily, each view is correct in certain crash circumstances: fatal crashes predict more severe injuries in some circumstances (e.g. sideswipe crashes in <50kmh zone) but not others (e.g. rear-end crashes in <50km/h zone, or right-angle crashes in 70-80kmh zone). These results can be applied to improve use of fatal crashes in selecting and prioritising treatments to improve safety benefits.