AppleInsider reports on a new study conducted by the University of California, San Francisco, revealing the Apple Watch heart rate sensor to be highly accurate in detecting atrial fibrillation, a common heart condition that can lead to stroke. Conducted in partnership with the Apple Watch app Cardiogram, the study paired a deep neural network with the heart rate sensor on the Apple Watch and then fed it a pool of 139 million heart rate measurements and 6.338 mobile ECG’s collected from 6,158 Cardiogram users who opted to participate in UCSF’s Health eHeart Study.
Presenting the findings at the Hearth Rhythm Society’s Heart Rhythm 2017 conference on Thursday, the report’s senior author Gregory M. Marcus, MD, MAS Endowed Professor of Atrial Fibrillation Research, and Director of Clinical Research for the Division of Cardiology at UCSF, stated that the “results show that common wearable trackers like smartwatches present a novel opportunity to monitor, capture and prompt medical therapy for atrial fibrillation without any active effort from patients.” Dr. Marcus noted that mobile technology isn’t ready to replace conventional methods of monitoring heart conditions, it can help with screening to reduce the number of undiagnosed cases of atrial fibrillation.
Brandon Ballinger, co-founder of the Cardiogram app used in the study, explained the process to AppleInsider, saying that about 200 participants took part in the study who had already been diagnosed with paroxysmal atrial fibrillation. These participants were asked to take at least one reading per day, using a mobile electrocardiogram, as well as taking readings when they felt the onset of specific symptoms such as lightheadedness or heart pains. The collected data was then used to train the deep neural network used in the study, which was then paired with heart rate data collected from the Apple Watch to identity atrial fibrillation. The deep neural network was later validated against a test sample of 51 patients who were scheduled to undergo treatment, and when compared against a 12-lead electrocardiogram, the Apple Watch and deep neutral network solution was found to be capable of identifying atrial fibrillation with an accuracy of 97 percent, sensitivity of 98 percent and specificity of 90.2 percent, surpassing prior detection algorithms.
