A systematic review and meta-analysis of 44 studies (42 articles) found no meaningful association between common wearable-derived sleep metrics and adult anxiety (e.g., sleep efficiency Fisher z=-0.07, 95% CI -0.14 to 0.002; prediction interval -0.19 to 0.05), while lower physical activity and higher heart rate were qualitatively linked to greater anxiety symptoms.

Why It Matters To Your Practice

• Patients increasingly bring smartwatch data to visits expecting it to explain anxiety symptoms, but pooled evidence doesn’t support using sleep metrics alone as an anxiety signal.

• Prediction intervals were wide enough to suggest real-world effects could vary substantially across settings and populations, limiting generalizability for point-of-care decisions.

• Wearable-only machine learning models showed mixed performance, improving when combined with other data sources (e.g., self-report, clinical data).

Clinical Implications

• Don’t over-interpret wearable sleep summaries (sleep efficiency, total sleep time, wake after sleep onset, sleep onset latency) as screening or monitoring tools for anxiety in adults.

• If you use wearables in care plans, prioritize trends that are more consistently directionally associated in the literature (lower activity, higher heart rate) and interpret them as nonspecific arousal/behavioral markers, not diagnostic signals.

• For AI-enabled “anxiety detection” tools, ask whether the model integrates multimodal inputs (wearable + patient-reported outcomes + context/clinical variables) and whether it was validated in a population like yours.

Insights

• The review included both inferential studies (n=36) and machine learning studies (n=8), with overall study quality mostly fair and sample sizes ranging from 17 to 170,320.

• Across four sleep metrics, meta-analyses were null (all P>.05), suggesting sleep features captured by wrist wearables may be too noisy, too indirect, or too confounded to track anxiety reliably.

• Machine learning performance varied when using wearable data alone, implying that “more data” isn’t the same as “right data” for mental health prediction.

The Bottom Line

• Wearable-derived sleep metrics, by themselves, shouldn’t be used to infer adult anxiety status.

• Wearables may add clinical value when combined with other signals—an important design constraint for AI tools aimed at anxiety screening or monitoring.