When does alcohol hurt? A driving simulator study.

World-wide, alcohol is still a major cause of traffic accidents. The dose-related accident risk function has been found in a large number of risk studies. A plethora of laboratory studies has examined the effect of alcohol with regard to different information processing capabilities of drivers. Summarizing the results, alcohol effects occur at lower blood alcohol concentrations (BAC) the more complex the tasks get. However, in contrast, typical alcohol-related crashes are frequently single vehicle crashes but not so often crashes in complex situations like at intersections. It may be that the subjective assessment of the traffic situation and the adaptation of behavior under the influence of alcohol plays a major role in accident causation. In order to examine this hypothesis, two driving simulator studies were conducted at a target BAC of 0.5g/l comparing two (alcohol vs. placebo; n=48, Experiment 1) and three (sober, placebo and alcohol; n=63, Experiment 2) groups of subjects in two critical scenarios. The first scenario was a seemingly easy traffic situation and was supposed to lead to a relaxed driving behavior under alcohol. The second scenario involved a complex intersection situation where especially drivers under the influence of alcohol should try to concentrate and compensate their experienced alcohol effects. In all scenarios, a critical object appeared suddenly and the driver had to react fast in order to prevent a (simulated) accident. Overall, the results support the hypothesis. Accidents were more frequent for alcohol drivers as compared to placebo/sober drivers in the easy scenario, but not the complex one. The initial speed of the driver when entering the scenario seems to play a major role in the accident causation. Drivers under the influence of alcohol seem to lower their speed in complex scenarios, possibly to thus counteract alcohol effects. In seemingly easy scenarios this does not seem necessary for them and the arousing effect of alcohol may contribute to driving faster. The results are summarized in a model of alcohol-related crashes. Further in-depth analyses of real crashes would be an interesting next step to further corroborate this model.