Smartwatch heart monitoring: what the evidence actually shows
Founder & CEO, Smartlet - CentraleSupelec engineer - Concours Lepine 2025, Awarded - CES 2026
Contents
Most heart rhythm problems do not announce themselves. Atrial fibrillation, the most common one, can come and go for months without a symptom, which is exactly what makes it dangerous and exactly why a device worn all day has something to offer. The claims around smartwatches and the heart run from careful to wildly overstated. This article stays with the numbers: how common the condition is, what the largest studies found when watches flagged an irregular rhythm, how accurate the detection is, and the plain limitation that a watch in a drawer detects nothing.
- 52.55 million people lived with atrial fibrillation or flutter worldwide in 2021, a 137% rise since 1990 (EP Europace, 2024).
- Alerts are rare and meaningful. In the Apple Heart Study, 0.52% of 419,297 participants received an irregular pulse notification.
- An alert is not a diagnosis. It is a reason to see a doctor, who confirms with a medical ECG.
- Detection is good, not perfect. Pooled studies put smartwatch AF detection at 95% sensitivity and 97% specificity.
- Wear time is the hidden variable. Intermittent AF can only be caught while the watch is on.
A common condition that often goes unnoticed
Atrial fibrillation is an irregular, often fast rhythm of the upper chambers of the heart. Its importance is not the rhythm itself but what follows it: AF raises the risk of stroke and heart failure, and it frequently has no symptoms until one of those happens.
The scale is large. Using Global Burden of Disease data, a 2024 analysis in EP Europace counted 52.55 million people with AF or atrial flutter in 2021, a 137% increase since 1990, with 4.48 million new cases that year and 0.34 million deaths (Cheng et al., EP Europace, 2024). A review in The Lancet Regional Health Europe puts the lifetime risk at about one in three to one in five people after the age of 45 (Linz et al., 2024).
When AF comes and goes, a single ECG at a clinic can miss it entirely. Detection then depends on how many hours of the day something is listening.
How a watch notices an irregular rhythm
Two different tools are involved, and they are often confused. The first is the optical sensor, which measures the pulse continuously in the background. When the interval between beats looks irregular across several readings, some watches send an irregular rhythm notification. The second is the ECG, a 30-second recording you start yourself by touching the crown or bezel (how to read those ECG results).
The optical layer is passive. It only works while the watch is worn, and it works best when the wrist is still, which is why so many notifications are generated at night or at rest (how the sensors work). The ECG is active: it gives a doctor something to look at, but only if you take it.
What the Apple Heart Study found
The largest study of its kind enrolled 419,297 people over eight months. Over a median of 117 days of monitoring, 2,161 participants, 0.52%, received an irregular pulse notification. Those who were notified were sent an ECG patch to wear, and among the 450 who returned usable patches, atrial fibrillation was present in 34% (Perez et al., New England Journal of Medicine, 2019).
The figure most often quoted is the positive predictive value: when a new notification arrived while a participant was wearing the patch, AF was confirmed on the patch 84% of the time. Of those who answered the 90-day survey, 57% had contacted a health care provider outside the study.
| Apple Heart Study measure | Result |
|---|---|
| Participants | 419,297 |
| Received a notification | 2,161 (0.52%) |
| AF on ECG patch, notified participants | 34% |
| Positive predictive value, concurrent notification | 0.84 |
| Contacted a provider (survey respondents) | 57% |
Read those numbers twice. Alerts are rare, which is good: a watch that cried wolf daily would be useless. And a 34% confirmation rate on a later patch is not a failure of the watch, it is the nature of intermittent AF, which may simply not recur during the patch week.
How accurate is detection
A 2025 meta-analysis in JACC Advances, pooling 26 studies and 17,349 patients, found smartwatch AF detection at 95% sensitivity and 97% specificity. By brand, Apple Watch reached 94% and 97%, Samsung 97% and 96%, and the Withings ScanWatch 89% and 95% (Barrera et al., JACC Advances, 2025). Good numbers. Not perfect ones.
A living review restricted to the Apple Watch was more conservative for AF, with a sensitivity of 0.79 and a specificity of 0.91 (Lambe et al., npj Digital Medicine, 2026). The difference is mostly about which studies are included and how unclassifiable readings are counted. The American College of Cardiology's 2024 expert analysis also reports a positive predictive value of 98.2% in the Fitbit Heart Study (Buda and Sengupta, ACC, 2024).
A notification is a reason to book an appointment, not a diagnosis. A doctor confirms AF with a medical ECG, and decides what, if anything, to do about it.
One documented case
Individual stories prove nothing statistically, but they show the mechanism. In February 2025, AppleInsider, citing FOX 7 Austin, reported that Mike Gomez, of Austin, Texas, received a high heart rate alert and then five notifications warning of atrial fibrillation, with no other symptoms. Doctors found he was in serious atrial fibrillation (AppleInsider, 2025). The value was not in the hardware. It was in the fact that the watch happened to be on his wrist that day, during a stretch of weeks when nothing else in his life suggested anything was wrong with his heart. Timing.
The gap is wear time
Every figure above assumes one thing: that the watch was worn. Paroxysmal AF is caught by whatever is listening at the time it happens, and a smartwatch spends a surprising number of hours off the wrist for reasons that have nothing to do with health. Charging is one. For people who love mechanical watches, the bigger one is choosing the other watch for the day.
This is the problem the Smartlet was made for. It is a metal link bracelet that holds a mechanical watch on top of the wrist and a smartwatch underneath, in the central part of the bracelet, against the skin. The mechanical watch takes watch adapters that seat in its lug holes in place of the spring bars, and the bracelet is sized to the wrist by its links. The smartwatch keeps running its background rhythm checks on the days it would otherwise have stayed home (a customer story on exactly this).
One honest limit. An ECG requested by a doctor should be taken exactly as the manufacturer describes, with the watch on its own band. The Smartlet adds hours of wear; it does not change the sensor or replace medical care. At night, the Clasp accessory closes the bracelet in place of the mechanical watch so the smartwatch stays on for sleep tracking.
Smartwatches with rhythm features
- Apple Watch: irregular rhythm notifications and a 30-second ECG from the Digital Crown. Both Apple Watch adapters are in the Smartlet box.
- Samsung Galaxy Watch: ECG and irregular rhythm features. The Galaxy adapter is provided when the Galaxy Watch 8, 8 Classic, 9, Ultra or Ultra 2 is selected at order.
- Withings ScanWatch: a single-lead ECG recorded in 30 seconds by resting a finger on the bezel, on demand (Withings).
The full list is in the smartwatch compatibility index, and mechanical watches are covered in watch brand compatibility. If you already wear an Apple Watch for its ECG, this piece on the trade-off with a mechanical watch goes further.
This article summarises published research on consumer devices. It is not medical advice. Smartwatch alerts are not a diagnosis. If you have symptoms such as palpitations, breathlessness or chest pain, seek medical care.
Smartlet, Bronze Medal at Concours Lepine 2025 and shown at CES 2026, keeps the smartwatch on the wrist on mechanical watch days.
