Prediction of performance during sleep deprivation and alcohol intoxication using a quantitative model of work-related fatigue

Abstract

Shift work and particularly night work can cause fatigue with subsequent negative impacts on health, sleep, and alertness. To facilitate better management of work-related fatigue, we have developed, optimized and validated a computerized model that can predict changes in performance, vigilance, sleepiness, and tiredness. The present study is a laboratory-based validation that demonstrates the further utility of the model in predicting performance impairment resulting from sleep deprivation and alcohol intoxication. Twenty-two healthy volunteers (mean age=22.0 years) each completed three counter-balanced laboratory conditions: sleep deprivation, alcohol intoxication, and a placebo control condition. In each condition, subjects were woken at 0700 h and performance on a variety of tests was measured hourly from 0800 h. The tests at 0800 h were then used as a relative baseline to which all other performance data were expressed. The six measures of performance assessed were grammatical reasoning (response latency and accuracy), unpredictable tracking score, vigilance (response latency and accuracy), and simple sensory comparison score. Regression analyses indicated that the fatigue model predicted between 47 and 98% of the variance in actual performance measures. Thus, there were moderate to very strong significant relationships between work-related fatigue model predictions and neurobehavioral performance as measured under laboratory conditions.