Detecting signs of hypertension using wearable sensors

Judgment

Google's Health Guardian product team now decides whether a no-calibration, multi-day hypertension signal ships as a passive screening prompt or stays a trend display; the cost is that a state aggregated over several days cannot answer a single-morning reading, and screening framing pulls in a clinical confirmation path.

Basis

Staff Research Scientist Paolo Di Achille and Senior Staff Research Scientist Ming-Zher Poh write that the system, AI-PPG-ACC-HTN, detects signs of hypertension from smartwatch PPG and accelerometry without any cuff calibration, and was validated against 24-hour ambulatory blood pressure monitoring (ABPM), reaching accuracy comparable to traditional cuff-based diagnostic workflows for both initial and confirmatory screening. The mechanism is a deliberate reframing: rather than estimating point-in-time systolic and diastolic values, which they describe as highly susceptible to transient environmental noise, the system aggregates passive physiological signatures over multiple days to identify an individual's underlying, persistent hypertensive state. They argue existing commercial cuffless devices need regular arm-cuff calibration, so they track only relative change from a baseline rather than identifying absolute hypertensive states, which limits their use as independent screening tools. Context given: hypertension affects over one billion people globally, more than half of adults with it are unaware, and the awareness gap is most severe among adults aged 18 to 44. The findings helped inform the blood pressure trends Health Guardian feature in the Google Health app.