The Slow Merger of Fitness and Medicine
Wrist-worn sensors started as step counters and ended up in clinical trials. The regulatory and medical systems are still deciding what to do with the data people already wear.
The distance between a fitness tracker and a medical device was once clear: one motivated, the other diagnosed. That line has been blurring for years, feature by feature — heart rate, then heart rhythm, then blood oxygen, then sleep staging, then early-warning flags for conditions a wearer did not know to look for. The devices on millions of wrists now produce data of a kind that medicine spent decades learning to value, and the two systems are meeting each other with mutual interest and mutual suspicion.
The clinical evidence has matured faster than the clinical workflow. Large studies have shown that wearable rhythm notifications can flag atrial fibrillation early enough to matter, and that continuous activity and sleep data carry genuine prognostic signal. What studies have not yet resolved is what a health system should do with a population-scale stream of mostly-normal readings punctuated by mostly-false alarms. Every alert generates a clinical encounter; at scale, sensitivity costs money and attention, and the false-positive burden is borne by systems that were not designed for it.
The regulatory response has been to build a middle category. Features that make diagnostic claims require clearance as medical devices; features framed as wellness do not. The boundary is policed by wording — a watch may notify you of an irregular rhythm suggestive of a condition, but may not say you have it. This produces the odd spectacle of companies engineering sentences as carefully as sensors. It also produces a genuinely useful discipline: the claims that survive clearance are the ones with evidence behind them.
The deeper change is in where care happens. Continuous monitoring moves measurement out of the clinic, where readings are rare and artificial, and into daily life, where they are constant and contextual. Blood pressure measured quarterly in a surgery is a different instrument from blood pressure tracked across months of ordinary Tuesdays. Remote monitoring programmes for chronic conditions are already showing reductions in admissions in structured trials, and hospital-at-home models lean on the same plumbing. The constraint is no longer the sensor; it is the reimbursement code, the integration with records, and the question of who is watching the dashboard.
The limitations here deserve emphasis. Wrist devices remain consumer instruments with real accuracy limits, especially on darker skin tones, during motion, and at the edges of their algorithms. Validation studies are frequently industry-funded. And the equity question is unresolved: the benefits of continuous monitoring currently flow to those who buy the devices, while the systems that could distribute them move slowly. The merger is real, but it is early, and the joining tissue — workflow, payment, liability — is still being built.
Every claim above is sourced to a document, a named person, or a record we hold. Where we could not verify a claim, we say so. Read our standards and corrections policy →