The Pulse
Google Research Tests Watch-Based Hypertension Screening
Google Research evaluated a smartwatch system that it says uses optical and motion sensors to detect signs of hypertension without cuff calibration. In a clinical study, its screening results were compared with 24-hour ambulatory blood pres
AI.info Team · Written with AI · fact-checked ·
Google Research’s smartwatch system recorded 65.8% sensitivity and 90.0% specificity in an initial hypertension-screening scenario, with a positive predictive value of 80.6%. The researchers benchmarked the system against 24-hour ambulatory blood pressure monitoring, or ABPM, which the post describes as the clinical gold standard for hypertension diagnosis. Google Research published its account on November 5, 2026.
Office-based initial screening in the same comparison had 55.3% sensitivity and a 77.8% positive predictive value. The system is designed to detect signs of a persistent hypertensive state from wearable data, not to estimate exact blood-pressure readings at a particular moment. Google says hypertension affects over one billion people globally, and more than half of adults with the condition are unaware they have it.
A score built from several days of wrist data
The system, named AI-PPG-ACC-HTN, uses photoplethysmography (PPG) and accelerometry (ACC), sensors Google says are already active on modern smartwatches. PPG measures changes in blood volume at the wrist through reflected light, while the accelerometer identifies still periods and captures subtle wrist vibrations associated with blood being ejected from the heart. The model combines those sensor signals with the user’s height to predict a hypertension probability score.
For its first training stage, Google Research used self-supervised contrastive learning with a one-dimensional deep convolutional neural network. The training data comprised 492,347 hours of sensor recordings from 2,109 consenting participants, with the model learning from overlapping 15-second PPG and ACC waveforms. In the second stage, the system mean-aggregates waveform features across multiple days of wear and feeds the resulting person-level representation, together with height, into a regularized logistic regression classifier.
Clinical testing included two screening situations
The prospective, multicenter study was conducted across five US sites and involved 196 consenting participants. Participants were free from known cardiovascular disease and were not taking blood-pressure medication. Researchers assessed the system for initial screening and for confirmatory testing after an elevated office reading.
For confirmatory testing, the researchers examined 48 participants whose initial office blood pressure was elevated. The smartwatch system recorded 78.4% sensitivity, 90.9% specificity and a 96.7% positive predictive value in that group. Return office measurements recorded 67.6% sensitivity, 63.6% specificity and an 86.2% positive predictive value.
Multi-day home monitoring in the confirmatory-testing comparison recorded 89.2% sensitivity, 81.8% specificity and a 94.3% positive predictive value. Google Research says the smartwatch system’s accuracy was comparable to traditional cuffed BP devices in the existing clinical framework.
Google describes the work as early feasibility research
The researchers examined average pulse waveforms from people with the highest and lowest predicted hypertension probabilities. Higher predicted risk was associated with slightly narrower pulse width, a slower systolic upstroke and a lower d/a ratio. The accelerometer also captured ballistocardiogram signals, with deeper K waves and wider IJK complexes observed in participants with higher predicted risk.
Google says smartwatch-based screening is not intended to replace standard clinical arm cuffs or provide a definitive standalone diagnosis. The post describes the work as early feasibility research and says its methods are exploratory, not indicative of an existing or upcoming commercial product feature or device capability. Google says the research findings helped inform the blood pressure trends Health Guardian feature in the Google Health app.