Landmark Apple Watch Heart Study Is "Not a Gold Rush," Says Cardiologist

"We need to be careful of interpreting this trial or thinking that it’s ready for prime time, because it’s not."

In November, Ed Dentel was diagnosed with atrial fibrillation by his Apple Watch. The Virginia man wasn’t shy about saying that his wearable saved his life. But was Dentel’s case was a fluke, or a sign that the Apple Watch really can be trusted to identify a lurking heart condition? Last week, the results of the Apple Heart Study were released, hinting at the answer.

In a presentation at the American College of Cardiology’s 68th Scientific Session, a team of scientists from Stanford University presented the results from the Apple Heart Study. This 419,093-person clinical trial was intended to evaluate how well the Apple Watch detects signs of atrial fibrillation — a heart condition that can cause stroke, if left unchecked. The work was presented by Dr. Mintu Turakhia, an associate professor of cardiovascular medicine at Stanford.

A screenshot from the Apple Heart Rate Study 

American Heart Journal 

Turakhia and his team showed that the Apple Watch successfully detected AFib in their patient sample. But Dr. Kim Eagle, the Director of the Frankel Cardiovascular Center and the editor the American College of Cardiology’s website, tells Inverse it falls far short of the standards that are required for it to be a real, dependable medical tool.

"You have to start asking yourself — how good is the technology?"

“The ability of wearables to inform health and disease is obviously an exciting future,” Eagle says. “At the same time, when you start thinking about the wearables either identifying or failing to identify dangerous problems like atrial fibrillation, you have to start asking yourself — how good is the technology? I think my sense is that it’s a start, but it’s nowhere near where it needs to be.”

What the Apple Heart Rate Study Found

Out of the people enrolled in the study, 2,161 people got pulse notifications from their watches indicating irregular heart behavior suggesting AFib. Of those, 450 ended up receiving and returning ECG patches sent to them by the researchers, which they wore for up to a week after the first notification. Those ECG patches, a traditional medical tool used to detect atrial fibrillation, were used to double-check the Apple Watch’s diagnoses. Only 34 percent of them had AFib diagnoses confirmed by the patch.

Overall, Turakhia and his co-authors actually wanted to report on how accurate the watch was when it gave a positive diagnosis — which they evaluated in a group of 86 patients. In that group of 86, the ECG confirmed AFib diagnoses in 84 percent of them. That might feel high, but Eagle worries that it’s still not high enough for a medical device.

"This is not a gold rush.

“The accuracy of our existing technologies is extremely high — high nineties,” he says. “We can’t assume that the technologies that are coming into our lives have the same diagnostic accuracy that we’re used to.”

By the same analysis, 16 percent of people whose watches told them they had AFib didn’t actually have it according to the ECG — indicating a false positive. And even more importantly, Eagle notes that these statistics don’t actually touch on the rate of false negatives — people who never even got any notifications about their heartbeat, but may still have the condition.

“Imagine if you’re a patient wearing an Apple Watch and it tells you you’re in AFib, and you go on the internet and it tells you could have a stroke,” he says. “Equally important side is that the watch also misses patients with AFib. Imagine a patient who has complications and the watch misses them.”

Inverse reached out out the authors of the Apple Heart Study, who were not available for interviews. We will update this article when we hear back.

What Does This Mean for Apple Watch Users?

Eagle explains that interpreting the results of the Apple Heart Study is all about perspective. Simply put, the Apple Watch algorithm used in this trial isn’t perfect, or even close to perfect as a medical device. But maybe it’s wrong to even think of them in that way.

The Apple watch's Afib algorithm will notify the wearer if they have five episodes of unusual heart activity out of every six readings. 

Apple Heart Study 

Instead, he adds that he would treat the Apple Watch’s AFib algorithm as an “intermediate technology.” Receiving no notifications isn’t confirmation that you don’t have AFib, nor is getting pinged by the watch definitive proof that you have the condition. It should be a sign to get more information from a doctor who runs a more rigorous test. In the paper they note that 57 percent of people who were pinged reached out to a doctor as a result.

“If my patient says, ‘It found a funky rhythm, I should let my doctor know so I can get monitored by a more reliable device,’ that would be one thing, Eagle adds. “But if we’re saying out there that it means you have AFib, that’s scary for people.”

But that’s not to say that one day devices like this might not actually hit the high standards we have for medical devices. The authors of the Apple Watch study have already begun another clinical trial, and the amount of wearable and digital health studies have jumped 29 percent per year between 2011 and 2017. Eagle says one day we might have devices as convenient as a watch that can perform like a medical device, but the Apple Heart Study just drives home that that day isn’t today.

“This is not a gold rush,” he says. “We need to be careful of interpreting this trial or thinking that it’s ready for prime time, because it’s not.”

Related Tags