Most people think of fitness wearables as step counters, or maybe sleep trackers. But it turns out that strapping on a gadget that continuously monitors your vital functions can have some surprising, even life-saving results.
Dr. Greg Marcus, director of Clinical Research for Cardiology at the University of California San Francisco, recently presented a study that caught my eye at the Heart Rhythm Society conference. It was designed to see if an Apple (AAPL) Watch could detect atrial fibrillation.
Turns out it can—with 97% accuracy.
Now, that statement needs a lot of explaining. But it’s important.
First of all, what’s atrial fibrillation? It’s the world’s most common heart-rhythm abnormality. At least 2.7 million Americans live with a-fib, according to the American Heart Association. It’s when the top chambers of the heart (the atria) irregularly quiver instead of pumping.
As a result, the blood just sits there. And you know what blood does when it’s not moving, right? It clots. If pieces of that clot break off and find their way to your brain, you can have a stroke. In fact, a-fib increases stroke risk four to five times compared to people who do not have the condition, according to the CDC.
What’s not so widely known, though, is that those clots can also cause heart attacks, kidney disease, and even dementia.
Now, if you have a-fib, you can take medicine for it (usually blood thinners). But how do you know if you have a-fib? You can get tested at a doctor’s office. Trouble is, a-fib is often intermittent—the quivering happens only sometimes. You might go in for the test, and the doctor will find no problems at all, because it’s not happening at that moment.
“There are monitors that have gotten easier for patients to wear,” Marcus told me. “But generally, they don’t monitor for more than two to four weeks. So again, if they don’t have that atrial fibrillation in those four weeks, you’re not going to catch it.” And, of course, you wouldn’t even have one of those monitors unless you already suspected you have a-fib.
The idea behind Marcus’s study: See if something that you’re wearing all the time, like a smartwatch, can detect atrial fibrillation when it happens.
Collecting the data
Marcus teamed up with the creators of an Apple Watch app called Cardiogram, which analyzes the watch’s heart-rate data. The hard part, though, was teaching the app how to recognize atrial fibrillation—coming up with an algorithm that would distinguish it from regular heart pumping.
To do that, the UCSF team assembled two groups of test subjects.
The first was composed of people who already know they have a-fib. For this, UCSF’s team exploited its own database, from something called the Health eHeart Study. (“This is a study that any adult can enroll in. We have more than 100,000 signed up,” Marcus says. “We hope that it will be a really valuable resource to help understand heart disease, potentially prevent heart disease, identify new targets for therapies for heart disease.”)
In this database, 166 people have (a) atrial fibrillation, and (b) an Apple Watch. They were equipped with a portable EKG monitor called an AliveCor, which displays its heartbeat readouts on your phone. Researchers already know that AliveCor does a good job of detecting a-fib; they compared the Apple Watch’s data with the AliveCor’s data to teach the Cardiogram app what bouts of atrial fibrillation look like.
The second group of subjects was patients who were coming in to UCSF for a procedure called a cardioversion. “This is a very common procedure to shock people out of atrial fibrillation,” Marcus says. “It’s not like in, you know, TV shows, although we do yell ‘Clear!’ before we shock. But it’s a very effective way to get people out of atrial fibrillation.”
Marcus outfitted these patients with the smartwatch, too. That way, he could collect data on what the watch detected during a-fib (before the shock) and after, when these patients’ hearts were beating normally.
(Why two groups of people? “In the training set, we used AliveCor. This second group undergoing cardioversion was the validation set—the reference standard, measured with the conventional 12-lead [12-wire] EKG that we use in the office,” Marcus says.)
The bottom line: This study compared the app’s measurements against traditional EKG readouts from people who have atrial fibrillation.
And how did it do?
“The app was able to detect or discriminate a-fib from normal rhythm with 97% accuracy,” Marcus says.
“Now, critics will very fairly point out that not everybody has a smartwatch,” he adds. “And that’s a limitation that’s really important to mention. But I think that these watches and this technology are going to become more and more common.”
(I wondered why the researchers used the Apple Watch, which is very expensive and therefore fairly uncommon—and checks your pulse only once every 10 minutes. Why not a Fitbit, which is smaller, cheaper, and takes continuous heartbeat measurements?
Simple: You can’t add apps to the Fitbit.)
Unfortunately, you can’t start monitoring yourself for a-fib with a smartwatch—yet. “There are quite a few additional hurdles we need to jump before we successfully roll this out,” Marcus says. “First, we need to figure out how to do this in a way that’s not going to unnecessarily alarm people who don’t have the disease. That’s often an issue in medicine—identifying the people with the disease without too many false positives.
“Second, in this study, we were only studying people we know had atrial fibrillation. So what we’d like to do is prove that, yes indeed, we can identify someone who previously didn’t know they had it. So there’s quite a bit more to be done, but we’re working hard on it.”
That’s good—because the ultimate goal of this project is reducing reduce stroke and death.
The real genius of this study is how it collects data as you go about your life. “Even the highest caliber studies generally are conducted in some study facility, some hospital, a place outside that participant’s natural environment. It’s not their home, it’s not their work, it’s not while they’re asleep,” Marcus says.
“But this is an extraordinary opportunity to capture real-time, real-life data that is dense. It’s continuous. And it reflects what’s going on in natural environments.”
More from David Pogue:
David Pogue, tech columnist for Yahoo Finance, is the author of “macOS Sierra: The Missing Manual.” He welcomes nontoxic comments in the comments section below. On the web, he’s davidpogue.com. On Twitter, he’s @pogue. On email, he’s firstname.lastname@example.org. You can read all his articles here, or you can sign up to get his columns by email.