In brief: Apple Watch hypertension notifications do not measure blood pressure. They analyse the optical heart sensor over 30-day windows and flag patterns associated with hypertension. In Apple’s validation study the feature alerted 41% of people with hypertension and wrongly alerted about 8% of people without it. An alert is a strong reason to check your blood pressure with a validated upper-arm cuff; the absence of an alert is not evidence that your blood pressure is normal.
High blood pressure is common, usually silent and often undiagnosed for years. That makes it an obvious target for passive screening: if a device already worn every day could flag likely cases, more people might get measured, diagnosed and treated earlier.
Since September 2025, Apple Watch has offered exactly that kind of feature. It has been cleared by the US Food and Drug Administration and rolled out in many other countries, including the European Union. Millions of people now have it switched on. The question we hear most is not how it works, but what a notification — or the lack of one — should actually change.
How the feature works
What it measures
The optical heart sensor (photoplethysmography, PPG) records 60-second segments roughly every two hours during waking time, when the accelerometer shows you are sitting still. A machine-learning model scores each segment for patterns linked to hypertension.
The 30-day window
Scores are averaged over non-overlapping 30-day periods. If the average crosses a threshold, you receive a notification suggesting you check your blood pressure. Nothing happens on a single day.
Who it is for
Adults aged 22 or older without a previous hypertension diagnosis, outside pregnancy. It runs on Apple Watch Series 9 or later and Ultra 2 or later (not SE), and availability varies by country.
What it does not do
It never shows a blood pressure number, it does not monitor people already diagnosed, and it is not a diagnosis. Apple states that it will not detect every case of hypertension.
What the validation study actually showed
Apple validated the feature in a separate study of 2,229 adults without a known diagnosis; 1,863 had enough data for the primary analysis. The reference was the average of home cuff readings taken twice a day for 30 days, classified with home thresholds (stage 1: 130–134 / 80–84 mm Hg; stage 2: 135 / 85 mm Hg or higher).
| Group | Result | 95% CI |
|---|---|---|
| Sensitivity — stage 1 hypertension | 29.6% | 24.5–35.1% |
| Sensitivity — stage 2 hypertension | 53.7% | 47.7–59.7% |
| Sensitivity — overall | 41.2% | 37.2–45.3% |
| Specificity — normal blood pressure | 95.3% | 93.7–96.5% |
| Specificity — elevated blood pressure | 82.8% | 78.1–86.8% |
| Specificity — overall | 92.3% | 90.6–93.7% |
The design choice is deliberate. Apple chose an operating point that favours specificity: few false alarms, at the cost of missing most milder cases. Sensitivity was also lower in people under 60 and in people with a BMI under 30, while sex, race and skin tone showed no clinically meaningful difference after adjustment.
That trade-off is reasonable for a consumer feature switched on by millions of people. It also means the feature behaves very differently from a test that “catches” hypertension: it is a net with large holes, designed so that whatever it does catch is usually real.
Why silence is not a clean bill of health
In February 2026, researchers from the University of Utah, Columbia and other centres applied Apple’s sensitivity and specificity to nationally representative US survey data (NHANES 2017–2020) to estimate what an alert, or its absence, means in practice. Their key point: the meaning depends on how common undiagnosed hypertension already is in your group.
| Group (eligible US adults) | Before the watch | After an alert | After no alert |
|---|---|---|---|
| Under 30 | 14% | 47% | 10% |
| 60 and older | 45% | 81% | 34% |
| Non-Hispanic Black adults | 36% | 75% | 26% |
| Hispanic adults | 24% | 63% | 17% |
Read the last column carefully. In a person over 60 who never receives a notification, the estimated probability of undiagnosed hypertension falls only from 45% to 34% — roughly one in three. The higher your baseline risk, the less reassuring silence becomes. The authors conclude that the feature should supplement cuff-based screening, not replace it, and warn that false reassurance could lead people to skip measurements they need.
If you receive a notification
- Do not treat it as an emergency. The alert summarises a 30-day pattern, not today’s reading. It is a prompt to measure, not a diagnosis.
- Measure with a validated upper-arm monitor. Wrist cuffs and cuffless devices are less reliable. Current US guidelines recommend validated upper-arm oscillometric devices and do not recommend smartwatches or rings for clinical blood pressure assessment.
- Follow a standard protocol. Avoid caffeine, exercise and smoking for 30 minutes beforehand. Sit for five minutes with your back supported, feet flat and the cuff on a bare upper arm at heart level. Take two readings a minute apart, morning and evening, for about a week.
- Keep the record. Log the values (the Health app has a blood pressure log) and bring the averages, not a single number, to a clinician. Diagnosis usually combines office and out-of-office measurements.
- Know the urgent exceptions. Readings of 180/120 mm Hg or higher, or high readings with chest pain, shortness of breath, weakness, numbness, or changes in vision or speech, need prompt medical attention — regardless of what the watch has said.
If you have never received one
Keep measuring your blood pressure as you would without the watch. Age, family history, excess weight, diabetes, kidney disease and pregnancy-related hypertension all raise baseline risk, and the higher the risk, the less a silent watch tells you. A periodic check with a validated cuff, at home or in a pharmacy or clinic, remains the reference.
It also helps to remember what the watch does measure well. Resting heart rate, heart rate variability and cardio fitness are useful longitudinal signals, but none of them is a blood pressure measurement. For how these values are recorded and how they relate to cardiovascular health, see this guide to blood pressure data in Apple Health.
Where the feature is genuinely useful
Hypertension is frequently missed because nobody measures. A feature that works in the background, costs the user nothing and produces few false alarms can move some people from “never checked” to “measured and diagnosed”. Apple has estimated that it could notify more than one million people with undiagnosed hypertension in its first year. At population scale, that is a real public-health contribution.
The risk is on the other side of the ledger: people who feel reassured by a silent watch. The responsible message is therefore asymmetric. Take an alert seriously and confirm it with a cuff. Do not take the absence of an alert as confirmation of anything.
A note on data and design
Passive cardiovascular screening means continuous collection of physiological data. Users should be able to see what is collected, where it is processed, and how to switch the feature off. The same principles we apply to our own products hold here: use the minimum data needed, explain how results are produced, and keep the human decision — measuring, consulting, treating — clearly in the loop.
Frequently asked questions
- Does the Apple Watch measure blood pressure?
- No. It estimates the likelihood of hypertension from optical heart-sensor patterns over 30 days. It never displays a systolic or diastolic value.
- How accurate are Apple Watch hypertension alerts?
- In Apple’s validation study, 41.2% of people with hypertension received an alert (53.7% with stage 2), and 7.7% of people without hypertension received one wrongly.
- I never got an alert. Is my blood pressure normal?
- Not necessarily. Most people with milder hypertension receive no alert. Modelling published in JAMA estimates that about one in three adults over 60 without an alert may still have undiagnosed hypertension.
- Can I use it if I already have hypertension?
- No. The feature is designed for people without a previous diagnosis. If you are diagnosed, use a validated cuff and follow your clinician’s plan.
Related: walking mobility metrics — from wearable data to better decisions
Sources and further reading
- Apple: Hypertension Notification Feature on Apple Watch (validation paper). September 2025.
- US FDA 510(k) summary K250507: Hypertension Notification Feature. 2025.
- Cohen JB et al. Impact of a smartwatch hypertension notification feature for population screening. JAMA. 2026;335(11):1001-1003.
- Jones DW et al. 2025 AHA/ACC guideline for the prevention, detection, evaluation and management of high blood pressure in adults. Hypertension. 2025;82(10):e212-e316.
- Shimbo D et al. Self-measured blood pressure monitoring at home: a joint policy statement from the AHA and AMA. Circulation. 2020;142(4):e42-e63.
- Apple: watchOS feature availability by country.
About the author
Albert Arnó writes about digital health, data interpretation and privacy-first technology. This article is educational and does not replace individual medical advice.