Driver drowsiness contributes to 10–20% of motor vehicle crashes. To reduce crash risk, ideally drivers would be aware of the drowsy state and cease driving. The extent to which drivers can accurately identify sleepiness remains under much debate. We systematically examined whether individuals are aware of sleepiness while driving, and whether this accurately reflects driving impairment, using meta-analyses and narrative review. Within this scope, there is high variability in measures of subjective sleepiness, driving performance and physiologically-derived drowsiness, and statistical analyses. Thirty-four simulated/naturalistic driving studies were reviewed. To summarise, drivers were aware of sleepiness, and this was associated to physiological drowsiness and driving impairment, such that high levels of sleepiness significantly predicted crash events and lane deviations. Subjective sleepiness was more strongly correlated (i) with physiological drowsiness compared to driving outcomes; (ii) under simulated driving conditions compared to naturalistic drives; and (iii) when examined using the Karolinska sleepiness scale (KSS) compared to other measures. Gaps remain in relation to how age, sex, and varying degrees of sleep loss may influence this association. This review provides evidence that drivers are aware of drowsiness while driving, and stopping driving when feeling ‘sleepy’ may significantly reduce crash risk.