Watches now detect irregular heart rhythms, measure blood oxygen, track sleep stages, and flag falls. Several of these features have regulatory clearance as medical functions.
The genuine benefit
Some conditions are intermittent. Atrial fibrillation can come and go, and a clinic recording lasting minutes may miss what a device recording continuously catches.
Detecting it matters because untreated it substantially raises stroke risk. There are documented cases of devices identifying it in people with no symptoms.
The problem with screening everyone
Any test applied to a population where the condition is rare produces false positives that outnumber true ones, even when the test is good. That is arithmetic, not a criticism of the device.
The consequence is people arriving at clinics anxious about a notification, requiring investigation that carries its own cost and occasionally its own risk.
Health systems have absorbed a meaningful increase in this workload without any additional resource.
The overdiagnosis question
Continuous monitoring finds abnormalities that would never have caused a problem. Treating them is not necessarily beneficial and is not necessarily harmless.
Medicine has learned this repeatedly with screening programmes, and the lesson does not transfer easily to a consumer product sold on the promise of finding things.
What is genuinely useful
Monitoring a known condition, where a baseline exists and changes are meaningful. Tracking recovery. Detecting deterioration in someone already under care.
Screening a healthy person with no symptoms is the weakest use and the most heavily marketed.
Practical advice
Take a notification to a doctor rather than to a search engine. And treat trends over weeks as more meaningful than any single reading, because single readings from these devices are noisy.