Skip to content
ShiftDelete.Net
Search news/ TR Türkçe
News Feed AI Smartphones Games Cars Computers Social Media Space Reviews Latest news
  1. Home
  2. Wearables

Smartwatch Hypertension Detection Rivals Office Cuff in Study

Google Research says a regular smartwatch flagged high blood pressure without a cuff, catching 65.8% of cases versus 55.3% for an office reading in a Nature study.

Hakkı ALKANHakkı ALKAN

Published 3 min read

Smartwatch hypertension detection concept: a smartwatch on a wrist next to a blood pressure cuff

Google Research says an ordinary smartwatch can flag signs of high blood pressure without a cuff, and in a 196-person clinical study its smartwatch hypertension detection caught more cases than a standard office reading. The work was published in Nature on November 5, and Google says the findings already helped shape the blood pressure trends feature in its Google Health app.

How does the watch spot high blood pressure without a cuff?

The system, which Google calls AI-PPG-ACC-HTN, combines two sensors most watches already have. The optical heart sensor records photoplethysmography, the tiny changes in blood volume under the skin. The accelerometer picks up the faint wrist vibrations caused each time the heart pushes blood into the aorta.

An AI model turns those waveforms into embeddings, averages them over several days and adds the wearer's height to estimate whether that person likely has hypertension. It never needs calibration against a cuff. According to the Nature press summary, the model was trained on nearly 500,000 hours of watch data from 2,109 people.

How accurate was the smartwatch hypertension detection?

The team tested it at five US sites on 196 adults with no known heart disease who were not taking blood pressure medication. Each wore the watch for seven days of normal life, and results were checked against 24-hour ambulatory monitoring, the clinical gold standard. The figures below come from the Google Research blog post and the paper.

MethodSensitivitySpecificityPositive predictive value
Smartwatch, initial screening65.8%90.0%80.6%
Office cuff, initial screening55.3%90.0%77.8%
Smartwatch, confirmatory (48 people)78.4%90.9%96.7%
Repeat office cuff, confirmatory67.6%63.6%86.2%
Multi-day home cuff, confirmatory89.2%81.8%94.3%

The watch found slightly more true cases than the office cuff during first screening, though Nature notes the gap was not statistically significant. In the confirmation group it beat repeat office visits but still trailed several days of home cuff readings on sensitivity.

Will this replace a blood pressure cuff?

No, and Google is explicit about that. The researchers pitch it as passive, opportunistic screening that nudges people who may be at risk to see a doctor and confirm the result with a cuff. Google also calls the method early feasibility research rather than a promise of a specific product feature. It joins a wider push to put health tools on everyday devices, alongside efforts like ChatGPT Health on Android.

Our Take

A cuffless screen that needs nothing but a watch you already wear could matter a lot, because more than half of adults with hypertension do not know they have it. A 65.8% hit rate means plenty of cases will still be missed, so a quiet watch is no all-clear. With Google also readying Android XR glasses, its wearables are clearly meant to watch more than steps. The realistic win here is catching people who would never have booked a check-up, and that is worth having even at this accuracy.