Apple Watch High Blood Pressure Detection: Technology and Validation
How the high blood pressure alert works
The analysis runs passively in the background over a longer period: the algorithm continuously evaluates data over a 30-day period. You only get a notification if signs of high blood pressure appear consistently over weeks. This longer observation period is meant to reduce false alarms from short-term fluctuations or measurement noise artifacts. Important: the watch does not measure a specific systolic or diastolic value in mmHg. It also does not make a diagnosis – instead it just notifies the user when the detected patterns suggest chronic high blood pressure. The alerts are intended to encourage users to have their blood pressure checked, since hypertension often runs silent and goes unnoticed.
Independent research also supports that this PPG-based approach can work. Studies show that AI algorithms can detect elevated blood pressure from PPG traces with useful accuracy. For example, a deep-learning model using a short PPG recording (10–15 seconds) on the UK Biobank dataset achieved a ROC-AUC of ~0.82 for detecting hypertension (and ~0.87 for more severe hypertension). In another study with long-term wrist PPG recordings a ResNet-based model also reached a very high specificity (>95%) when distinguishing normal blood pressure from hypertension.
Scientific validation and clinical studies
Apple extensively validated the high blood pressure alert before release. First, the algorithm was trained on very large datasets: according to Apple, PPG and health data from multiple studies totaling over 100,000 participants went into development. The finished algorithm was then tested in a prospective clinical study with more than 2,000 participants. In that study adult participants (without known hypertension) wore an Apple Watch for about a month for ~12 hours a day and simultaneously measured their blood pressure twice daily with a standard cuff. This allowed a direct comparison of whether the watch warned the people who also showed persistently elevated values on the standard blood pressure device. The result: Apple Watch notifications proved similarly reliable to the conventional cuff at detecting hypertension patterns. In other words: the watch could detect the same signs of chronic high blood pressure in participants who were identified by daily cuff measurements.
This evidence of accuracy was key to regulatory approval – in September 2025 Apple received FDA clearance in the US for the Hypertension Notifications. Apple and independent experts emphasize, however, that the watch does not replace seeing a doctor or using validated measurement devices. The new feature is meant primarily as a screening and early-warning tool. So if the watch issues a high blood pressure alert, Apple explicitly recommends that users re-measure their blood pressure carefully with a validated device within the next week (for example at home in the morning and evening) and discuss those readings with a doctor. This approach matches current American Heart Association guidelines, which call for repeated measurements over several days to diagnose hypertension. To record and analyze these measurements properly we recommend the BloodPressureDB app.
Cardiologists welcome the new Apple Watch feature as an opportunity to identify hidden high blood pressure earlier, but they also point out that it does not replace real-time blood pressure monitoring and will not catch every rise in blood pressure. Overall, the thorough validation – from big-data training to the controlled 2,000-patient study – shows that this technology is scientifically sound. It could help uncover many previously undetected cases of hypertension at scale (Apple estimates over 1 million in the first year) and thus help prevent cardiovascular disease.
Criticism
Apple has now published the validation study. We reviewed it and came away somewhat disappointed: the proportion of cases in which high blood pressure is correctly detected is smaller than expected.
Stage 1 hypertension is detected in less than 30% of cases, while stage 2 hypertension is detected in about half of cases.
There is also a noticeable uneven distribution: detection is worse in people under 60 and in those with a BMI under 30. That the feature reports high blood pressure more often in older people with higher BMI is not surprising – you don't need a sophisticated algorithm for that.
You have to see it this way: the feature is an additional capability of the Apple Watch, and because it's so widely used it will surely identify many previously undiagnosed hypertension cases. That's an impressive achievement by Apple. But: no one should buy an Apple Watch just for this. The feature is not a reliable safeguard.
Sources
Author Horst Klier has been dealing intensively with high blood pressure since 2002 – initially from personal experience and since 2009 as a developer of BloodPressureDB – and, thanks to his app and specialist platform used millions of times and numerous publications, he is now regarded as an established blood pressure expert. As the author of several health guides and professional articles, he explains complex topics clearly and in a practical way.
