Understand this story
Google Research published in Nature the AI-PPG-ACC-HTN multimodal AI system, which can passively detect signs of hypertension using only the PPG and accelerometer signals of consumer-grade smartwatches, without cuff calibration. The system first performs self-supervised pretraining on unlabeled sensor data totaling 492,347 hours from 2,109 participants, then aggregates multi-day waveform embeddings into person-level features and, combined with height, uses logistic regression to output the probability of hypertension. In a prospective study of 196 participants across 5 U.S. sites, in the initial screening setting the sensitivity was 65.8%, specificity 90.0%, and PPV 80.6%; in the confirmation setting among 48 people with elevated office blood pressure, the sensitivity was 78.4%, specificity 90.9%, and PPV 96.7%. The research team states that this work is an early feasibility exploration, does not replace cuff-based diagnosis, and does not point to any existing or forthcoming commercial feature.
AI-generated from public reports
Latest development · Report
Google Research Uses Smartwatch Sensors to Passively Detect Signs of HypertensionDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 6
Google ResearchFirst-partySelected
Google Research Uses Smartwatch Sensors to Passively Detect Signs of Hypertension
Google Research published the AI-PPG-ACC-HTN system in Nature, which can passively detect signs of hypertension using only PPG and accelerometer signals from consumer-grade smartwatches, without the need for cuff calibration. The system first uses 492,347 hours of unlabeled sensor data from 2,109 participants for self-supervised pretraining, then aggregates multi-day waveform embeddings into person-level features, and combines them with height using logistic regression to output a hypertension probability. In a prospective study of 196 participants across 5 U.S. sites, the initial screening scenario achieved 65.8% sensitivity, 90.0% specificity, and 80.6% PPV; in the confirmation scenario involving 48 individuals with elevated office blood pressure, it achieved 78.4% sensitivity, 90.9% specificity, and 96.7% PPV. The research team states that this work is an early feasibility exploration, does not replace cuff-based diagnosis, and does not point to any existing or forthcoming commercial feature.