The research used a deep neural network that was based on photoplethysmographic (PPG) sensors that are frequently found in smart watches. When paired with an artificial intelligence-based algorithm, Apple Watch’s heart rate sensor can detect AF.
There were 6,158 participants in the University of California San Francisco Health eHeart Study who all used Cardiogram for the Apple Watch. The data gathered from the study was used to train a deep neural network to distinguish AF from normal heart rhythm automatically.
A group of 51 patients who were undergoing a cardioversion validated the deep neural network. They each wore an Apple Watch before and after their cardioversion for 20 minutes. The deep neural network was able to correctly detect AF with 97% accuracy, 98.04% sensitivity and 90.20% specificity.
“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 senior author on the study Gregory M. Marcus in a press release. “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 AF.”
AF is an irregular heartbeat that can eventually cause blood clots, stroke, heart failure and other heart-related complications. Approximately 2.7 million Americans are currently living with AF, according to the American Heart Association.
A recent report also suggested that the Apple Watch could include a glucose monitor in future versions. A source claimed that Apple is working on developing the ability to measure glucose through optical sensors on the Watch through light shining through the skin.