How accurate is a smartwatch at flagging sleep apnea?

In a 2026 lab comparison, a Samsung Galaxy Watch flagged moderate to severe obstructive sleep apnea with 94.1% sensitivity and 66.7% specificity at its default setting. In practice it caught most true cases and also flagged about one in three people who did not meet the threshold. Participants were already at higher pretest risk, which limits how far the numbers travel.

What did the smartwatch study measure?

Alavi and colleagues enrolled 152 adults (147 completed both nights) who were aged 22 or older and either had a prior diagnosis of moderate to severe sleep apnea or scored 3 or more on the STOP-Bang questionnaire, a screening tool. Each person wore the watch during two nights of in-lab polysomnography, the reference test, plus at least three nights at home.

The target was moderate to severe apnea, defined as 15 or more breathing events per hour. At the watch’s default threshold, sensitivity was 94.1% (95% CI 84.1 to 98) and specificity 66.7% (95% CI 51 to 79.4), with an AUROC of 0.94. A threshold optimised on this same cohort raised specificity to 94.9% and lowered sensitivity to 82.4%, so that setting is specific to this sample.

How does that compare with other wearables, and what is missing?

Samsung funded the study, and the authors note that data and code cannot be shared for proprietary reasons. The sample was enriched for people likely to have apnea, so the results say less about the general population, milder disease or people without symptoms.

For context, a meta-analysis of 59 studies of wearable and contactless devices (Hasanain et al., Sleep and Breathing 2025) pooled wearable results from 2,743 participants: sensitivity 0.87 (95% CI 0.81 to 0.92) and specificity 0.86 (95% CI 0.78 to 0.91) for moderate to severe apnea. Its abstract does not say which device types counted as wearables, and the authors say more validation is needed across populations and care settings.

The detection question echoes the smartwatch atrial fibrillation trial: a wearable can flag a possibility, and a formal test settles it.

The bottom line

• In 152 higher-risk adults, a Samsung smartwatch had 94.1% sensitivity and 66.7% specificity for moderate to severe sleep apnea against overnight lab testing.
• A pooled analysis of wearable devices across 59 studies reported sensitivity 0.87 and specificity 0.86, though its abstract does not say which device types were counted.
• The watch study was company-funded and enriched for likely cases, so it describes screening in a higher-risk group, not the general public.