HRS2017: AI-enabled Apple Watch detects AFib in UCSF study
Apple devices just keep getting smarter. New research presented at this year’s Heart Rhythm Society (HRS) conference in Chicago showed that when paired with an artificial intelligence-based algorithm, the watch can detect atrial fibrillation (AFib).
With an increase in people wearing fitness trackers, like Apple Watches or Fitbits, researchers saw an opportunity to address diagnosing and treating AFib in a convenient way, according to an HRS press release.
The study, which included more than 6,000 users of Cardiogram—a heart rate monitoring app used with the Apple Watch—was conducted at the University of California, San Francisco (UCSF). Data from the study included 139 million heart rate measurements and more than 6,300 mobile ECGs. It was used to train a neural network that could automatically distinguish AFib from normal heart rates.
Once the neural network was created, researchers tested it on 51 patients who were set to undergo cardioversion. Each patient wore an Apple Watch for 20 minutes before and after the procedure. Results showed that the watch could detect AFib with an accuracy of 97 percent, with a sensitivity rate of about 98 percent and with a specificity rate of 90 percent, which were all higher than other algorithms used to detect the condition.
“Our 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,” said Gregory M. Marcus, MD, lead author on the study and an endowed professor of atrial fibrillation research and director of clinical research for the Division of Cardiology at UCSF, in a statement. “While mobile technology screening won’t replace more conventional monitoring methods, it has the potential to successfully screen those at an increased risk and lower the number of undiagnosed cases of AFib.”