An empirical analysis of how asleep/fatigued driving-injury severities have changed over time

Abstract Asleep/fatigued driving has proven to be a serious and persistent highway-safety problem. This study investigates aspects of this problem by studying the temporal changes in driver-injury severities in single-vehicle crashes that involve asleep/fatigued driving. To do this, random parameters logit models with unobserved heterogeneity in means and variances were estimated to compare injury-severities in asleep/fatigued crashes in Florida in 2014 and 2019. The estimated models are based on available police-reported crash data that include a wide variety of factors related to the spatial, temporal, and weather characteristics as well as vehicle characteristics, traffic information, harmful events, roadway attributes, and driver characteristics. The model estimates show that there were many statistically significant factors determining driver-injury severities resulting from asleep/fatigued driving, and that the effect of these factors on driver-injury severities has changed significantly over time, with many explanatory variables producing temporally shifting marginal effects. While asleep/fatigued driving crashes remain a serious safety concern, the empirical findings indicate (using model prediction simulations) that the resulting injury severities in crashes involving asleep/fatigued driving have declined between 2014 and 2019, likely reflecting the effectiveness of safety campaigns and ongoing improvements in vehicle safety technologies and highway safety features.