There is a growing body of studies indicating that extended shift duration has an adverse effect on fatigue, consequently leading to reduced work performance and higher risk of accident. Following modern fatigue risk management systems (FRMS), acceptable performance could be maintained by the mobilization of appropriate mitigation strategies. However, the effective deployment of such strategies assume that workers are able to assess their own level of fatigue-related impairments. In this study, we sought to determine whether emergency physicians’ subjective feelings of sleepiness could provide accurate knowledge of actual fatigue-related impairments while working extended shifts. We conducted a prospective observational study with a within-subjects repeated measures component. We collected sleep logs, sleepiness ratings and reaction times on a Psychomotor Vigilance Task (PVT) at different time points during shifts. Our results show that the PVT is sensitive to sleep loss and fatigue, with a 10% increase in mean reaction time across the shift. Subjective sleepiness, however, showed no significant association with time since awakening and was not a significant predictor of PVT performance. Our results are consistent with experimental studies showing that individuals tend to underestimate fatigue-related impairments when sleep deprived or functioning under adverse circadian phase. The discrepancy between subjective sleepiness and actual fatigue-related impairments may give workers the illusion of being in control and hinder the deployment of mitigation strategies. Further research is needed to determine the relative weight of circadian phase shifting and cumulative sleep deprivation in the decline of self-knowledge in extended shifts.