Google Research model spots signs of high blood pressure from ordinary smartwatch sensors, without a cuff
In a study of 196 people published in Nature, the watch found 65.8% of people with hypertension. A standard office measurement found 55.3%.
Google Research has developed an AI system that detects signs of high blood pressure using only the pulse and motion sensors already built into modern smartwatches. It needs no blood pressure cuff for calibration. The study was published in Nature and described in a Google blog post. A neural network learned from nearly 500,000 hours of sensor data by comparing short segments and working out which ones came from the same person. That way it picked up typical pulse and heartbeat patterns without diagnoses to guide it. The system averages these patterns over several days and calculates a probability of high blood pressure. In a study of 196 people at five US research sites, the watch identified hypertension about as well as a traditional cuff. It found 65.8 percent of those with high blood pressure, compared with 55.3 percent for the standard in-office measurement. Google says the findings have been built into the blood pressure trends feature of the Google Health app. It stresses that the method is still a research project and does not replace a diagnosis. The goal is to encourage people to get checked: according to Google, more than half of people with high blood pressure do not know they have it.