On Nov 5, 2026, researchers from Google Research, Beth Israel Deaconess Medical Center and Harvard Medical School published a study in Nature. They trained an AI model to read two sensors an ordinary smartwatch already has and look for signs of hypertension, the medical term for high blood pressure, without any cuff calibration. The authors call it early feasibility research; Alphabet, Google's parent company, funded the study, and the authors are Alphabet employees.
The number
65.8%. That is the share of people with hypertension that the watch model flagged in a test of 196 adults (its sensitivity, in medical terms). An office check with a cuff at one clinic visit (both arms, then two more readings on the higher arm, threshold 140/90) flagged 55.3% of the same people. Both methods wrongly flagged about 1 in 10 people with normal blood pressure (specificity 90.0%). The authors call the results comparable: the ranges of statistical uncertainty overlap (54.0 to 76.3% for the watch, 43.4 to 66.7% for the office check). The reference for who really had hypertension was a 24-hour cuff monitor.
What was found
- The model reads an optical sensor that tracks the pulse wave at the wrist and a motion sensor that picks up tiny vibrations from each heartbeat.
- 196 adults at five US sites wore the watch for about seven days of ordinary life.
- The reference was 24-hour ambulatory blood pressure monitoring, the standard method for diagnosing hypertension.
- According to the authors, hypertension rarely gives obvious symptoms, and more than half of adults who have it do not know.
What it does not mean
- It is a screening signal, not a diagnosis. The authors say the method is not meant to replace arm cuffs or give a standalone diagnosis, only to prompt at-risk people to get a formal check.
- A quiet watch is not a clean bill of health: the model missed about a third of the cases in the test.
- 196 people is a small sample, and the results say nothing about people with known heart disease or on blood pressure medication, who were excluded.
- It is not a medical device. Google says the findings helped inform the blood pressure trends feature of Health Guardian in the Google Health app, rolling out free with Pixel Watch 3, 4 and 5 (blog.google, Sep 24, 2026). Google describes that feature as intended for general wellness, not a replacement for a blood pressure monitor and not a prescreener for hypertension. The paper itself says its methods are exploratory, not a product.
- The Nature article is an early version, subject to further edits by the journal.
Why it matters
Many people wear a watch every day and rarely meet a blood pressure cuff. If a watch can point some of them toward a check they would otherwise skip, screening could reach people before symptoms do. Whether that holds outside a study is a question for larger trials.
Sources: Di Achille et al., "Opportunistically Detecting Signs of Hypertension on a Consumer Smartwatch", Nature, Nov 5, 2026, doi.org/10.1038/s41586-026-11336-3; Google Research blog, "Detecting signs of hypertension using wearable sensors", Nov 5, 2026, research.google/blog/detecting-signs-of-hypertension-using-wearable-sensors; preprint of the same study, medRxiv, Dec 15, 2025, doi.org/10.64898/2025.12.10.25341972; Google Keyword blog, "Health Guardian features go live in Google Health", Sep 24, 2026, blog.google/products-and-platforms/products/google-health/health-guardian-features-live. Checked Nov 9, 2026.
We are not affiliated with the institutions mentioned.