The ability to self-monitor performance during a week of simulated night shifts.

STUDY OBJECTIVES Research has indicated that individuals are able to accurately monitor the performance decrements they experience during unitary periods of acute sleep deprivation. The aim of the current study was to investigate the ability to self-monitor performance during a week of simulated night shifts. DESIGN Subjects completed 7 consecutive 8-hour night shifts (11 pm-7 am). SETTING University sleep laboratory. SUBJECTS Fifteen young (7 men, 8 women, 19-25 years) healthy volunteers. INTERVENTIONS During the night shifts, performance was measured hourly on 4 performance parameters: psychomotor vigilance test (PVT), tracking, and grammatical reasoning (GRG) accuracy and response latency. Before and after each test, subjects completed visual analogue scales, which required them to rate their alertness and their performance speed and/or accuracy. RESULTS Analysis indicated that GRG response latency and tracking were significantly impaired (P<0.05) during the first 2 shifts only. The PVT performance displayed consistent impairment, with significant (P<0.05) declines during all but the final shift. The pattern of deterioration in subjective ratings of alertness was similar to that of the PVT data. Correlations between subjective alertness and self-ratings of performance were significant (P<0.01) for all parameters (r=0.39-0.69). Significant (P<0.05) correlations were found across the week between pretest performance ratings and actual performance for all parameters except GRG accuracy (r=0.29-0.58) and between posttest ratings and actual performance for all parameters (r=0.52-0.75). Correlations between pretest ratings and actual performance were also conducted separately for each shift. Highest correlations were found during the first shift, with r-values that were low for GRG accuracy (r=0.32) and GRG response latency (r=0.20), moderate for tracking (r=0.41), and high for PVT (r=0.82). In general, lower correlations were found later in the week. CONCLUSIONS Overall, results indicate that individuals have only a moderate ability to predict performance impairment during a week of night shifts. It is likely that performance ratings are based, at least to a certain extent, on subjective alertness levels. Furthermore, it seems that rating accuracy is improved on tasks providing performance feedback, such as the PVT. Finally, it appears that after testing, individuals have a more accurate perception of their performance.