Accuracy of Three Commercial Wearable Devices for Sleep Tracking in Healthy Adults.
Level 3 - non-randomized controlled study
Diagnostic validation study comparing index tests against a reference standard in a single cohort
PubMed 39460013 · doi:10.3390/s24206532
What was done
Thirty-five healthy adults aged 20 to 50 years without sleep disorders participated in a single-night inpatient validation study. Each participant wore three commercial sleep trackers simultaneously (Oura Ring Gen3, Fitbit Sense 2, and Apple Watch Series 8) alongside gold-standard laboratory polysomnography (PSG) to evaluate accuracy for sleep/wake detection and sleep staging.
What was found
Sensitivity for distinguishing sleep from wake was ≥95% across all three devices. For discriminating sleep stages: - Oura Ring had 76.0–79.5% sensitivity and 77.0–79.5% precision, with no significant differences from PSG for wake, light, deep, or REM sleep duration. - Fitbit Sense 2 had 61.7–78.0% sensitivity and 72.8–73.2% precision, overestimating light sleep by 18 minutes (p < 0.001) and underestimating deep sleep by 15 minutes (p < 0.001). - Apple Watch Series 8 had 50.5–86.1% sensitivity and 72.7–87.8% precision, underestimating wake by 7 minutes (p < 0.01) and deep sleep by 43 minutes (p < 0.001), while overestimating light sleep by 45 minutes (p < 0.001).
Why it matters
This study shows that while modern commercial wearables reliably measure total sleep time, sleep architecture breakdowns—particularly deep and light sleep—remain variable and device-dependent.
Limits
The study was limited to 35 healthy adults monitored during a single inpatient night. It did not evaluate real-world multi-night wear, older adults, or individuals with sleep disorders.
Cited by
- supports Consumer wearable sleep staging algorithms have approximately 60% to 80% accuracy relative to polysomnography.