The Apple Watch ECG is highly accurate for the one thing it’s designed to do: distinguishing atrial fibrillation (AFib) from a normal heart rhythm. In Apple’s own clinical testing of 588 subjects, the algorithm showed greater than 98% sensitivity and 99% specificity for that task. Independent research confirms strong performance, but with important caveats about what the watch can and cannot detect.
What the ECG App Actually Measures
The Apple Watch records a single-channel electrocardiogram similar to what doctors call a Lead I ECG. It works by completing an electrical circuit between two points on your body: a sensor on the back of the watch touching your left wrist, and the digital crown, which you press with your right index finger. The electrical signals your heart generates travel through your chest and arms, and the watch picks up that tiny current to produce a 30-second waveform.
A clinical ECG uses 12 leads placed across your chest and limbs, capturing your heart’s electrical activity from multiple angles. The Apple Watch captures just one angle. That single view is enough to reliably detect the irregular rhythm pattern of AFib, but it misses a great deal of information that a full ECG would reveal.
Accuracy for Detecting AFib
The numbers vary depending on how you define “accuracy,” and the distinction matters. When a physician reviews the actual PDF waveform the Apple Watch generates, a study published in Circulation found 96% sensitivity and 100% specificity. That means the waveform correctly showed AFib in 96 out of 100 confirmed cases, and it never falsely labeled a normal rhythm as AFib.
The watch’s built-in algorithm, which classifies your rhythm automatically as “AFib,” “sinus rhythm,” or “inconclusive,” performs differently. In the same study, the automated notification caught AFib with only 41% sensitivity, though it maintained 100% specificity (zero false positives). The gap exists because the algorithm labels many recordings “inconclusive” rather than making a call. It classified 31% of readings from patients in normal rhythm as inconclusive, meaning it simply couldn’t determine the result.
The Stanford-led Apple Heart Study, which enrolled over 400,000 participants, found that when the watch flagged an irregular pulse and a simultaneous ECG patch was recording at the same time, the detection was confirmed as AFib 84% of the time. Only 0.52% of all participants ever received an irregular pulse notification, which suggests the watch is conservative by design. It would rather stay quiet than cry wolf.
How It Compares to a Clinical ECG
The WEAR-TECH study compared 400 simultaneous recordings from the Apple Watch Series 6 and a standard 12-lead ECG. For measuring specific heart timing intervals (the gaps between electrical events that doctors use to assess heart function), the watch agreed best with certain leads on the 12-lead machine but not all of them. The Apple Watch tended to slightly overestimate some intervals compared to the clinical ECG.
When researchers looked at how closely the measurements matched, about 71% of readings fell within 40 milliseconds of the 12-lead results, and only 21% fell within 10 milliseconds. For rhythm detection, that level of agreement is often adequate. For precise diagnostic measurements, it’s a meaningful limitation. Your doctor may find the waveform useful as a snapshot, but it won’t replace the detail of a full workup.
What the Apple Watch ECG Cannot Detect
The FDA clearance is narrow. The ECG app has been evaluated only for detecting AFib or confirming a normal sinus rhythm. It cannot detect heart attacks. It cannot identify other types of arrhythmias. If you’re experiencing chest pain, the watch is not a diagnostic tool for that situation.
Researchers have explored whether the Apple Watch could theoretically capture enough data to spot heart attack signs by holding the watch against multiple positions on the chest, mimicking additional ECG leads. But this remains a proof-of-concept experiment, not a real-world capability. Even when those extra readings were successfully captured, they still required interpretation by a physician.
The FDA also notes the app is not intended for people who already have a diagnosed arrhythmia other than AFib, and Apple states it does not guarantee you’re free of heart problems even when no irregular rhythm is detected.
Why You Might Get an Inconclusive Result
Inconclusive readings are common, and they don’t necessarily mean something is wrong. The app cannot assess for AFib when your heart rate is below 50 bpm or above 120 bpm, so the recording defaults to inconclusive. Heart rates between 100 and 120 bpm can also trigger inconclusive results because the algorithm has difficulty classifying rhythm in that range.
Physical factors matter too. A loose watch, wet skin, sweaty wrists, or movement during the recording can all introduce electrical noise that drowns out the signal. Apple’s instructions recommend resting your arms on a table, keeping your wrist and watch clean and dry, staying still for the full 30 seconds, and moving away from plugged-in electronics that could cause interference. People with certain neurological conditions like Parkinson’s disease may have difficulty getting a clear signal due to tremor.
Who the ECG App Is Designed For
The FDA cleared the ECG app for over-the-counter use in adults 22 and older. It was not tested in younger users. The intended use case is someone without a known arrhythmia who wants to check whether their heart rhythm looks normal or shows signs of AFib, a condition that affects millions of people and often goes undiagnosed because episodes can be brief and sporadic.
For that specific purpose, the Apple Watch performs well. It offers a convenient way to capture a rhythm snapshot during moments when you feel your heart racing or fluttering, giving your doctor real data from the moment symptoms occurred rather than relying on your description alone. Many cardiologists find these recordings genuinely useful, particularly when a patient’s symptoms are intermittent and hard to catch during a scheduled office visit.
Where the watch falls short is in the assumptions people make about it. A normal reading does not rule out heart disease, coronary artery blockages, or arrhythmias other than AFib. It is a screening tool for one specific condition, and within that scope, it performs remarkably well for a consumer device on your wrist.