The wrist device buzzes: resting heart rate elevated, sleep fragmented, recovery score red. Maybe you drank wine. Maybe you’re getting sick. Maybe stress from a deadline. The algorithm lacks context but offers certainty in percentages.
Wearables crossed from step counters to provisional medical adjuncts — ECG alerts cleared by FDA on some models, irregular rhythm notifications, SpO2 during sleep, menstrual cycle prediction, glucose trends via patches in select ecosystems. Useful. Leaky. Barely regulated at the consumer boundary.
What wearables measure well
Trends over time — personal baseline comparison beats absolute thresholds. Resting heart rate drift upward may precede illness.
Activity and recovery — motivates movement; imperfect calories but directionally useful.
Sleep duration and stages — actigraphy approximates; not polysomnography but captures schedule chaos.
Heart rhythm flags — can prompt clinical follow-up; also false positives flooding cardiology (cost and anxiety).
What they measure poorly (so far)
Blood pressure — cuffless optical methods emerging; clinical validation uneven.
Non-diabetic glucose — lifestyle curiosity market; medical claims require restraint.
Stress — HRV proxies correlate; do not capture psychological complexity.
Single-night conclusions — noise, alcohol, altitude distort; users treat scores as verdicts anyway.
Overlap with blue light and sleep — devices both diagnose and contribute to problem via midnight notifications.
The privacy architecture
Data flows: sensor → phone app → cloud → sometimes employer wellness program, insurer pilot, or research partnership.
Terms of service — often permit anonymized aggregation; “anonymized” re-identification demonstrated repeatedly.
Employer programs — discounts for sharing steps; coercion dressed as wellness. HIPAA may not apply if employer uses vendor intermediary — legal gaps.
Insurance — usage-based life insurance experiments; adverse selection and discrimination risks by health status inference.
Law enforcement — court orders for cloud data; period tracking post-Dobbs raised reproductive privacy alarms.
Read our online privacy guide — wearables are continuous biometrics broadcast.
Medical boundary problems
Users treat Apple Watch ECG as diagnosis; it is screening. False reassurance dangerous. False alarm expensive.
FDA clearance specific to features, not whole device as doctor replacement. Marketing blurs line intentionally.
Integration with electronic health records promised; interoperability still fragmented. Doctors distrust consumer-generated noise.
Who benefits
Users — early nudges, accountability, data for clinician visits if summarized intelligently.
Platforms — lock-in; health ecosystem increases switching cost from iPhone to Android or vice versa.
Researchers — population-scale datasets if ethical collection; history includes exploitative paths.
Healthcare systems — remote patient monitoring for chronic disease reduces admissions when implemented with care, not as cost-cut only.
How to use wearables without surrender
- Minimize connected apps with broad permissions
- Disable features you do not use (always-on location, social sharing)
- Ask employer wellness opt-in consequences before enrolling
- Export personal copies; do not assume cloud perpetual access
- Bring trends to doctors, not screenshots as self-diagnosis gospel
Regulatory pressure building in EU (health data classification under GDPR+) and US state privacy laws — enforcement lags device sales.
Connection to broader tech ethics
Wearables preview BCI and synthetic companion debates — bodily data as product. Consent once at setup insufficient for decade-long sensor relationship.
Conclusion
Wearable health tech works best as mirror with memory — not oracle, not doctor, not gift to insurers. The biometric story on your wrist is intimate. Treat the export settings accordingly.
Your body generates the data. The business model decides who else keeps a copy.
Lumen is edited by Leo Hartmann. Related: Blue Light and Sleep · Online Privacy Guide