Google's AI screens for high blood pressure using ordinary smartwatches

Google Research has developed an artificial intelligence system that looks for signs of high blood pressure using only the pulse and motion sensors found in modern smartwatches. The approach does not require a blood-pressure cuff for calibration. The study was published in Nature.

A neural network trained on nearly 500,000 hours of sensor recordings learned typical pulse and heartbeat patterns by comparing short data segments associated with the same person, without using diagnostic labels. It combines measurements collected over several days to estimate the probability of hypertension. Google’s SensorFM health model also learns from smartwatch data.

In a study involving 196 people at five research sites in the United States, the watch identified 65.8 percent of participants with high blood pressure. The standard in-office measurement identified 55.3 percent. Google said the findings have been incorporated into the blood pressure trends feature of the Google Health app, but the method remains a research project and is not a substitute for diagnosis. Its intended role is to encourage people to seek medical checks.