Hypertension is nicknamed the silent killer for a reason. High blood pressure usually produces no symptoms at all. You don't feel it. There's no ache to notice, no twinge to mention to a doctor. It simply grinds away at your arteries, heart, kidneys and brain for years, and the first sign is often the event you were trying to avoid.
Which is why the Apple Watch hypertension notification feature got so much attention when it arrived with watchOS 26. A device already strapped to millions of wrists, using a sensor that's been there for years, quietly raising a hand when it thinks your blood vessels are under chronic pressure.
It's genuinely useful. It's also widely misunderstood, including in some of the coverage that followed the launch. So let's do this properly: what the feature actually reads, what the notification means, what it definitely doesn't mean, and how accurate Apple's own validation data says it is.
TL;DR: Apple Watch can flag possible hypertension using the optical heart sensor and a machine learning algorithm, not by measuring blood pressure. It works passively in the background during waking hours, analysing roughly 30-day windows of pulse data, and it does not give you a blood pressure number in mmHg. It's intended for people aged 22 and over who have never been diagnosed with hypertension, and it shouldn't be used during pregnancy. Only Apple Watch Series 9 or later and Apple Watch Ultra 2 or later support it — not Apple Watch SE. In Apple's pivotal study of more than 2,000 people, the feature caught roughly 41% of people who had hypertension, with about 92% specificity, meaning most positives turn out to be real flags but a large share of people with hypertension never get an alert. A notification is not a diagnosis. It's a prompt to go measure your blood pressure properly with a cuff for seven days and talk to a clinician.
YouTube: Related video
What the feature does — and the one thing it doesn't
Start with the thing it doesn't do, because this is where most of the confusion lives.
Apple Watch cannot measure your blood pressure. It doesn't produce a systolic/diastolic reading. It won't show you 128/82. There's no on-demand measurement, no cuff calibration, nothing you can trigger and watch fill in.
What it does instead is classify a risk pattern. Apple's own documentation is explicit: the feature "does not measure blood pressure directly, and does not surface a blood pressure reading to users." It looks at the shape of your pulse signal over a long period and decides whether that pattern is more consistent with someone who has hypertension than someone who doesn't.
That's a meaningful distinction. A blood pressure reading is a measurement. A hypertension notification is a suspicion, arrived at statistically, that you should go get the measurement.
The reason Apple took this route is practical. Cuffless blood pressure estimation from a wrist-worn optical sensor is genuinely hard, and the industry has a track record of producing numbers that look precise but aren't. Rather than ship an estimate that invites false confidence, Apple shipped a yes/no nudge toward a real cuff. That's a more honest product, even if it's a less satisfying one.
How Apple Watch detects possible hypertension
Every Apple Watch with a heart-rate sensor already shines green and infrared light into your wrist and watches how the blood volume underneath changes. That's photoplethysmography, or PPG. It's the same signal used for heart rate, HRV and irregular rhythm notifications.
What's new is what Apple does with it.
The hypertension feature works passively, in the background, during your waking hours. It doesn't require you to sit still for a test or start a session. Roughly every two hours, the watch captures a 60-second segment of PPG data, and it uses the accelerometer to filter for moments when you're actually sitting still, because a moving wrist produces a useless signal.
Those segments then run through a three-stage machine learning pipeline that Apple describes in its validation paper:
- A PPG feature extractor, built with self-supervised deep learning on large-scale unlabelled data from the Apple Heart and Movement Study. This part was trained without any blood pressure data at all — it just learns what pulse waveforms look like.
- A hypertension risk score per segment, produced by linear models trained on PPG data paired with reference readings from at-home blood pressure cuffs.
- A notification decision, where segments are filtered to remove anything collected during sleep, then averaged across a 30-day window and compared against a threshold.
Apple says the algorithm was developed using data from more than 100,000 study participants. The evaluation happens over non-overlapping 30-day windows, which is why nothing can happen in a day or a week. If you enable it today, the first window closes in a month.
One practical consequence: because samples from sleep are excluded and the model needs you seated and still, the feature leans on daytime, stationary moments. Someone who is constantly moving during the day, or who takes the watch off for long stretches, will generate a thinner data set.
Which watches get it
Hypertension notifications are available on:
- Apple Watch Series 9 or later
- Apple Watch Ultra 2 or later
- Not Apple Watch SE — any generation
No new hardware was required, which is why the feature could arrive as a software update to watches already on people's wrists rather than being locked to a new model. You need the latest watchOS and a paired iPhone running a current iOS, since setup happens in the Health app on the phone, not on the watch itself.
Availability also varies by country and region, because health features that make claims about a medical condition need regulatory clearance market by market. The feature was cleared by the US FDA via the 510(k) pathway, but that doesn't automatically make it available everywhere. If you can't find it in your Health app, region is the most likely reason.
What happens when you get a notification
This is the part people skip, and it's the part that actually matters.
If the algorithm detects a pattern consistent with hypertension, you get an alert on your watch and your iPhone. The notification does not tell you your blood pressure. Instead, it walks you into a structured follow-up:
- A Blood Pressure Log is set up in the Health app.
- You're prompted to measure with a third-party blood pressure cuff — a real one — twice a day for seven days, morning and evening.
- You log systolic and diastolic values manually into the app, with reminders scheduled for after you wake and before bed.
- The log classifies your readings according to established thresholds from the American Heart Association and the European Society of Cardiology.
- The result is a tidy record you can hand to a clinician, rather than a vague memory of "my watch said something."
That seven-day, twice-daily protocol isn't arbitrary. Hypertension diagnosis has always required repeated measurements across days, because a single reading in a clinic is noisy and often inflated by the setting itself. Apple is essentially nudging you toward doing the measurement the way a doctor would want it done.
The log can also run for a longer four-week period if you're monitoring known hypertension rather than screening for it.
How accurate is it, really
Now for the numbers, because they're more nuanced than either the hype or the backlash suggested.
In Apple's pivotal decentralised study, more than 2,000 adults without a prior hypertension diagnosis wore the watch for at least 12 hours a day and took home blood pressure readings twice daily for 30 days using a validated upper-arm cuff. The algorithm's output was then compared against those readings.
The headline result: roughly 41% sensitivity and 92% specificity. In plainer language:
- Specificity of about 92% means that when the watch does fire, it's usually pointing at something real. Not always, but usually. Apple reported specificity of 95.3% in the "normal" blood pressure category.
- Sensitivity of about 41% means the feature misses most people who have hypertension. Apple's own framing is that it notifies "nearly half" of users who have hypertension, and more in the higher-risk Stage 2 category.
Put together: a notification is worth taking seriously, but silence is not reassurance. If you have hypertension and your watch never pings, that is entirely expected, not a clean bill of health.
A 2025 commentary in the Journal of the American Heart Association made the same point more forcefully, arguing that the low sensitivity means the absence of a notification could produce false reassurance and delayed diagnosis, and that FDA clearance should not be read as proof of accuracy or clinical readiness. That's a fair critique and worth sitting with. This is a nudge feature, not a screening programme.
Why Apple didn't make it more sensitive
There's a design tension here, and Apple has been fairly open about it.
Every threshold on a risk algorithm trades one kind of error for another. Push sensitivity up and you catch more real cases — but you also generate more false positives, and every false positive is a person sent to a doctor, worried, for nothing. On a device that already sends irregular rhythm alerts and high heart rate alerts, a feature that cries wolf gets muted. And a muted feature helps nobody.
Apple's engineers have described deliberately weighting the algorithm toward protecting the trustworthiness of the notification, accepting that some cases will be missed as the cost of keeping the positives meaningful. You can argue that trade-off either way. What you can't do is treat the absence of an alert as evidence that everything is fine.
Who should turn it on — and who shouldn't
Apple is unusually specific about the intended population, and it's worth respecting those boundaries rather than enabling the feature by default.
It's intended for you if: you're 22 or older and you have never been diagnosed with hypertension.
It's not intended for you if:
- You have already been diagnosed with hypertension. The feature is a screening tool for undiagnosed people, not a monitoring tool.
- You are under 22.
- You are pregnant. Apple explicitly states the feature should not be used during pregnancy.
- You have an Apple Watch SE, or a Series 8 / Ultra 1 or older.
One more important limitation: the watch cannot detect a heart attack, and Apple says so directly in the setup flow. If you have symptoms that worry you — chest pain, breathlessness, sudden severe headache, vision changes — you call emergency services. You do not wait for a wearable to weigh in.
If you get a notification: what to actually do
The path is short and concrete.
- Don't panic, and don't self-diagnose. You have a statistical flag, not a diagnosis.
- Get a validated upper-arm cuff. Not a wrist cuff, not a smartwatch — a proper upper-arm monitor with a validated model. This is the single most important step, because everything downstream depends on the quality of those readings.
- Measure twice a day for seven days. Morning after waking, and before bed. Sit quietly for five minutes first, back supported, feet flat, arm at heart level, no caffeine or exercise in the preceding 30 minutes.
- Log every reading in the Health app. The Blood Pressure Log will remind you and will classify your numbers against AHA and ESC thresholds.
- Take the log to a clinician. Bring the actual numbers. A seven-day record of 14 readings is far more useful than a single clinic measurement, and it's exactly the kind of evidence that supports a proper diagnosis and a decision about treatment.
If you don't get a notification: still know your numbers
This is the part that deserves the most emphasis, because it's where the feature's limits bite hardest.
With sensitivity around 41%, most people with hypertension will never see an alert. So if you want to know your blood pressure status, you cannot outsource it to your wrist. A yearly check at the doctor, or a home cuff if your blood pressure has ever been borderline, is still the tool that actually answers the question.
The CDC's numbers show why it matters. Nearly half of US adults — 47.7% in the most recent national survey — have hypertension by the AHA/ACC definition, and about 59% of them know it. That leaves a substantial group walking around with elevated pressure and no idea. In 2024, high blood pressure was a primary or contributing cause of over 680,000 deaths in the United States.
Knowing your numbers is cheap. Finding out the hard way is not.
How it fits with the rest of your watch data
The hypertension feature is one more signal on a device that's already collecting a lot of cardiovascular data, and the signals are more useful together than apart.
Your resting heart rate and HRV shift in response to the same underlying stressors that influence blood pressure — poor sleep, alcohol, illness, chronic stress. Tracking those over weeks gives you a picture of your baseline and how it moves. If you want to understand what those numbers mean before adding another one to the pile, our HRV meaning guide is the place to start.
Sleep is a bigger input than most people realise. Poor sleep, fragmented sleep, and especially untreated sleep apnea all push blood pressure up, which is why the link between sleep and blood pressure is worth understanding alongside hypertension notifications.
And if you're wondering whether a consumer wearable can measure blood pressure directly, the short answer is not reliably — we covered exactly where that stands in our piece on whether your smartwatch can measure blood pressure.
A tool like Century AI helps here by pulling HRV, resting heart rate, sleep and recovery into one daily view, so a change in any of them shows up against your own baseline instead of a population average. That's the difference between noticing a trend and reading a single number on a single morning. It doesn't replace a blood pressure cuff, and it isn't meant to — but it does tell you when your body's signals are drifting.
Summary
- Apple Watch hypertension notifications flag possible high blood pressure using the optical heart sensor and a machine learning algorithm. They do not measure blood pressure and never show a reading in mmHg.
- The feature works passively during waking hours, analysing roughly 30-day windows of seated, quiet PPG samples. Nothing happens in a day or a week.
- Available on Apple Watch Series 9 or later and Ultra 2 or later — not Apple Watch SE — with regional availability that varies by country.
- Intended for people 22 and over who have never been diagnosed with hypertension, and not for use during pregnancy.
- In Apple's study of 2,000+ participants, sensitivity was around 41% and specificity around 92%: most alerts are meaningful, but most people with hypertension won't be alerted.
- If you get a notification, the right response is a validated upper-arm cuff, twice-daily readings for seven days, logged in the Health app, and a conversation with a clinician.
- If you don't get a notification, that is not an all-clear. Know your numbers the old-fashioned way: a cuff and a check-up.
Century AI helps you understand your body with a daily health score, recovery score, and sleep insights — using the watch you already wear.