Sensitivity of established driving simulator metrics to varying doses of alcohol during a brief motorway task.

Driving simulators are valuable for investigating driving impairment since ethical and safety considerations limit the ability to conduct such research on roads. Although simulators have generally demonstrated relative validity, alcohol is often referred to as a reference for impairment due to well-established legal limits and its known effects on driving ability. However, there is a relative lack of large-sample studies that systematically assess the sensitivity of established driving performance metrics to alcohol-induced impairment, alongside the application of rigorous methodological approaches designed to minimise bias in estimating alcohol-related effects on driving performance. Therefore, this study employed a between-subjects alcohol dose-response experimental design using a driving simulator to assess the effects of alcohol on driving performance. Participants (N = 165, 77 females, mean age = 34.62 years) were randomly allocated to one of three single-blind conditions (range BAC: 0.020% - 0.085) and completed a short motorway driving task while wearing an eye tracker. A causal framework was employed to identify and adjust for potential confounders, and Bayesian modelling was used to estimate alcohol-related effects on driving performance and quantify uncertainty. The study showed first that simulator sickness is unlikely to bias driving performance estimates in driving simulator studies using alcohol. Then, the study assessed the sensitivity of established driving performance metrics as a result of alcohol. The largest effects were found on speed selection and lateral control while driving on a motorway, followed by reaction time during an emergency braking event and longitudinal control when following a vehicle. Limited changes in gaze patterns were found on this simple motorway drive. The findings of this study provide the basis for the development of models of impairment that are anchored on BAC levels, which is currently the sole impairment with widely accepted levels in the road safety domain.