Accuracy of 11 Wearable, Nearable, and Airable Consumer Sleep Trackers: Prospective Multicenter Validation Study

Area of Science:

• Biomedical Engineering
• Sleep Science
• Wearable Technology

Background:

• Consumer sleep trackers (CSTs) are popular for self-monitoring sleep.
• Limited validation studies exist for widely used CSTs.
• Assessing CST performance against polysomnography is crucial.

Purpose of the Study:

• To validate the accuracy of various consumer sleep trackers (CSTs) using polysomnography.
• To provide comprehensive insights into CST performance in a hospital setting.
• To evaluate CST applicability for sleep monitoring.

Main Methods:

• Compared 11 commercial CSTs (wearables, nearables, airables) against polysomnography.
• Conducted a multicenter study with 75 participants and analyzed 3890 hours of sleep data.
• Evaluated epoch-by-epoch agreement in sleep stage classification.

Main Results:

• Significant performance variation observed among the 11 CSTs, with macro F1 scores ranging from 0.26 to 0.69.
• Specific CSTs excelled in certain sleep stages (e.g., SleepRoutine for wake/REM, wearables for deep sleep).
• Device type influenced accuracy: wearables showed bias in sleep efficiency, nearables in sleep latency.

Conclusions:

• Certain CSTs demonstrate substantial agreement with polysomnography, indicating potential for sleep monitoring.
• Other CSTs showed partial consistency, highlighting performance variability.
• Findings offer insights into CST strengths across different classes for wellness-focused sleep management.