Subjective sleepiness and accident risk avoiding the ecological fallacy

Area of Science:

• Occupational Health
• Human Factors Engineering
• Transportation Safety

Background:

• Shift work is associated with disrupted sleep patterns and increased daytime sleepiness.
• Sleepiness is a known contributing factor to motor vehicle accidents.
• Quantifying the precise relationship between subjective sleepiness levels and accident risk is crucial for safety interventions.

Purpose of the Study:

• To determine subject-level relative risks (RR) for accidents based on varying levels of subjective sleepiness.
• To model accident probability using a Generalized Linear Mixed Model (GLMM) approach for subject-specific risk estimation.
• To investigate the impact of sleepiness, measured by the Karolinska Sleepiness Scale (KSS), on driving safety in a simulated environment.

Main Methods:

• A high-fidelity driving simulator was used to assess 10 shift workers (5 male, 5 female, mean age 37) during a 2-hour drive.
• Subjective sleepiness was recorded every 5 minutes using the Karolinska Sleepiness Scale (KSS), ranging from 1 (very alert) to 9 (very sleepy).
• Incidents, accidents, and crashes were recorded, and accident probability was analyzed using a GLMM to estimate individual subject effects.

Main Results:

• A strong correlation was found between increased sleepiness and a higher risk of accidents.
• Compared to KSS level 5, an average subject faced an estimated 28.2 times increased accident risk at KSS 8 and 185 times at KSS 9.
• Significant individual variability in event propensity was observed, complicating precise individual risk prediction.

Conclusions:

• Subjective sleepiness is a critical predictor of accident risk in driving scenarios.
• The study provides robust, subject-specific estimations of accident risk associated with different sleepiness levels.
• While strong correlations exist, individual differences in susceptibility to sleepiness-related risks warrant further investigation.