Tests done on the Apple Watch indicate the device can detect atrial fibrillation early, potentially reducing the risk of stroke.

Atrial fibrillation, a type of irregular heartbeat, is a silent killer. Many people don’t know they have the condition until they have a stroke or other related cardiovascular problems.

But detection of the disease could soon be as easy as a glance at your wrist.

Heart rate data gathered by an Apple Watch can be used to detect atrial fibrillation (AFib), according to researchers who used an artificial intelligence-based algorithm to analyze information from the watch.

In a study group of 51 patients receiving treatment for irregular heartbeat, AFib was detected with a 97 percent success rate by the watch.

When applied to a group of 1,617 people enrolled in the University of California at San Francisco (UCSF) Health eHeart Study, the program successfully predicted AFib in 72 percent of those who had previously self-identified as having an irregular heartbeat.

AFib affects 34 million people worldwide and is a leading cause of stroke.

The condition can be successfully treated with medication or electrical stimulation of the heart, called cardioversion.

“The challenge is to find people with AFib earlier, since the first hint is often stroke or death,” Dr. Gregory M. Marcus, lead study author and director of clinical research for the UCSF division of cardiology, told Healthline. “This study can be seen as the first proof of concept that a watch can passively detect atrial fibrillation.”

Apple itself launched the Apple Heart Study app last fall in collaboration with Stanford Medicine.

The app uses the Apple Watch’s heart rate sensor to detect irregular heartbeats and then alerts users if AFib is suspected.

Apple Heart Study participants suspected of having AFib also get a free consultation with a physician and an electrocardiogram (ECG) patch for additional monitoring.

In the UCSF study, the watch was found to be somewhat less accurate than an ECG in detecting AFib.

“The technology is already there to screen for AFib,” said Marcus, “but the final determination still needs to be done using an ECG by a treating physician.”

Results from the Apple and Stanford study haven’t yet been made public.

Marcus and colleagues didn’t have access to the raw sensor data collected by Apple. Instead their research is based on consumer-accessible information on heart rate and step count.

The algorithm developed by the UCSF group looked for irregular patterns in the data, which itself varied from sampling heart rate every five seconds while the watch was in workout mode to every few minutes during normal activity.

Among the smaller study group, for example, only workout-mode data was used, while both types of data were studied for the larger group.

The UCSF study was published in the journal JAMA Cardiology.

The Apple Watch uses flashing LED lights and light-sensitive photosensors to detect the flow of blood through the wrist and determine heart rate and rhythm.

The watch’s software allows heart rhythms to be isolated from other noise, according to Apple.

About 9 percent of American adults currently own a smartwatch. That’s expected to grow to 15 percent by 2019, according to the 2017 NPD Group Connected Intelligence Wearables Survey.

Marcus said that smartwatch ownership is also growing among older adults, who are most at risk of AFib.

“There is tremendous potential in wearable technologies in detecting disease on a population scale,” Dr. Michael Levy, a cardiologist at Lahey Hospital and Medical Center in Massachusetts, told Healthline. “Additionally, if the technology becomes more reliable, a smartwatch may one day be used in place of conventional monitors that are difficult for some patients to wear.”

Nora Zetsche, co-founder of Veta Health, which integrates data gathered from wearable devices, clinicians, and manual user input, said that wearables like the Apple Watch “could really result in better relationships between caregivers and patients.”

It can do this, she said, by generating reminders and alerts that “can help patients understand their care better and also help them self-manage their condition.”