Google researchers and medical collaborators developed an artificial intelligence system capable of identifying possible hypertension from pulse and motion signals in a smartwatch without prior calibration using a blood-pressure cuff. The prospective, multicenter study published in Nature on November 5 followed 196 participants for seven days and compared the results with 24-hour ambulatory blood-pressure monitoring. The system detected 65.8% of hypertension cases with 90% specificity but missed about 33% of affected participants. Its positive predictive value was 80.6% in this risk-enriched cohort, whose characteristics limit wider generalisation of the findings. Clinical confirmation with a cuff remains necessary, particularly because the watch does not provide an exact systolic or diastolic blood-pressure reading.
Source: Read the original article
