Can we rely on wearable sleep-tracker devices for fatigue management?

Area of Science:

• Sleep science
• Human performance
• Biomedical engineering

Background:

• Wearable sleep trackers are widely used but often show discrepancies with polysomnography (PSG).
• The acceptable error margin for sleep trackers in specific applications remains unclear.
• Assessing alertness impairment requires accurate sleep parameter estimation.

Purpose of the Study:

• Develop a quantitative method to evaluate sleep tracker accuracy for alertness impairment assessment.
• Determine the operational acceptability of sleep tracker errors.
• Establish a criterion for acceptable sleep measurement errors.

Main Methods:

• Literature data on 18 sleep trackers' errors versus PSG were analyzed.
• Simulations compared alertness (Psychomotor Vigilance Test) with and without sleep measurement errors.
• Acceptable error was defined as <30 milliseconds difference in mean response time.

Main Results:

• On average, devices overestimated sleep by 19 minutes (SD=44).
• In nearly 80% of simulations, predicted alertness differences were below the 30-millisecond threshold.
• This held true across various sleep schedules over 30 days.

Conclusions:

• A method is provided to quantitatively assess sleep tracker operational acceptability.
• This method aids in fatigue management by validating sleep measurement tools.
• The findings support the use of certain sleep trackers for alertness monitoring.