Is an ECG from your smartwatch actually useful?
For detecting atrial fibrillation, smartwatch ECGs show positive results across multiple studies, including one meta-analysis. The strongest evidence applies to the Apple Watch and Samsung Galaxy Watch, which achieved sensitivities of 85% in clinical research. Beyond AF, such as for cardiac ischaemia or heart muscle disorders, the evidence is too thin to claim practical value. A smartwatch signals and alerts, but does not replace a doctor.
Smartwatch ECGs have been studied most thoroughly for detecting atrial fibrillation, a common heart rhythm disorder that can cause strokes. A meta-analysis of multiple studies found a sensitivity of 93% and a specificity of 94%, with an AUC of 0.94. That is impressive accuracy for a wristwatch, comparable to that of smartphone-based ECG apps1.
In practice, performance is more variable. A clinical study involving 201 patients (the Basel Wearable Study) tested five popular brands: Apple Watch 6, Samsung Galaxy Watch 3, Withings Scanwatch, Fitbit Sense, and AliveCor KardiaMobile. Sensitivity ranged from 58% (Withings Scanwatch) to 85% (Apple Watch 6 and Samsung), with specificity between 69 and 79%. In addition, the algorithm was unable to reach a verdict for 17 to 26% of measurements. When a physician manually reviewed those inconclusive cases, 99% were still classified correctly. This means that a smartwatch ECG without medical follow-up is a limited tool2.
In older patients, such as those who have had a stroke (mean age 79 years), the Fitbit Charge 5 behaves differently. It measures heart rate reliably during a normal rhythm, but during atrial fibrillation the heart rate measurement is unreliable and the automatic AF alert misses as many as 66% of cases. The flip side is that an alert that is given is almost always correct: specificity 100%. People in this group may therefore take a green notification seriously, but should not interpret the absence of a notification as reassurance3.
A major practical advantage of a smartwatch over a one-off ECG at the GP's office is continuous measurement. Atrial fibrillation sometimes occurs only in brief episodes that are missed during a short reading in the consulting room. Large studies, including the Apple Heart Study and the Fitbit Heart Study, show that smartwatches detect these kinds of episodes more often4. That advantage does not extend to other conditions: for detecting cardiac ischaemia (oxygen deficiency in the heart muscle) or cardiomyopathies, the smartwatch ECG is not yet clinically usable, due to technical limitations in how signals are measured and displayed5,6.
Anyone who wants to track their heart rate variability (HRV, a measure of recovery capacity and stress level) via a Samsung smartwatch is best advised to do so during sleep. During the day, movement artefacts cause large measurement errors in most HRV parameters, whereas nighttime measurements of heart rate and most variability values do agree well with a gold-standard measurement7. Finally, a smartwatch is a tool, not a treatment. Review authors point out that devices without healthy lifestyle habits make no noticeable difference to cardiovascular health8.
Based on one meta-analysis (PMID 32921618), one prospective clinical study with n=201 (PMID 36858690), one validation study in stroke patients (PMID 37430546), two technical reviews (PMID 37562224, 33441002), one validation study with n=28 (PMID 36480505), and two large observational registry studies (Apple Heart Study, Fitbit Heart Study, summarised in PMID 39363440) plus a review comment (PMID 35624593). The evidence for AF detection is strongest; for other applications it is thinner.