A car road deaths model to explain the annual road death peak near 1970 in high income countries, using driver experience and travel

Stipdonk, H.

The World Health Organisation estimates that 1,350,000 road deaths occur worldwide. Adequate models relating road death data to distance travelled or risk factors do not exist. Remarkable road safety events such as a common peak in many High Income Countries around 1970 in the annual number of road deaths lack understanding.

Sensible models require stratification by crash type (e.g. pedestrian-car), as different crash types show different trends and require different measures. Models often define annual numbers of road deaths equalling distance travelled times a risk component containing relevant risk factors. Such factors could be safety measures such as safety belt use for car occupants, helmet use for two wheeler riders, gradual behavioural changes triggered by the introduction of mobile phones, or general factors such as the weather. In this paper, we introduce average driver experience as an essential factor in car crash trends.

There are no available data of road crashes or distance travelled, stratified by driver experience. Therefore, we have to introduce some assumptions regarding this factor. We used total car distance travelled data (1950–2015) and car fleet data (1946–2016) to estimate the car distance travelled stratified by driver experience. We assumed that annual car fleet increase corresponded to the introduction of new cohorts of car drivers.

To model the experience factor in risk we assumed this factor to be a simple parametrized function, decreasing with driver experience and dependent on the age drivers start their driving career. For car-car crashes we incorporated the effect of experience factor and distance travelled for two car drivers simultaneously. For each crash group we further assume a parametrized negative exponential risk decrease to model the influence of all other risk factors.

We describe how changes in the annual number of three groups of car related road deaths coincide with changes in the proportion of inexperienced drivers, with considerable resemblance between model values and road death data. For Car-Car crashes and Pedestrian-Car crashes we show that average driver experience improved road safety until the nineties, while there is still some effect of this factor for Single Car crashes.

Published in
Safety Science
129 (104635)

SWOV publication

This is a publication by SWOV, or that SWOV has contributed to.