Last updated on
Wearable devices (smartwatches and fitness trackers) can tell us a lot about ourselves, from how many steps we’ve taken in a day to how much water we’ve had and how many calories we’ve burned. But can they accurately flag heart abnormalities? We spoke with Dr. Daniel Yazdi from Brigham and Women’s Hospital to find out more.
Q: Let’s start with the concept of “precision health.” It’s a term we hear often. What exactly is it?
Precision health, in a very broad sense, refers to data and signals that we can use to provide personalized, targeted health care to an individual. There’s a wide spectrum of data that we can use, ranging from information we obtain from blood samples to data we can collect from wearables and other devices. Precision health allows us to reach patients at points that are convenient for them, so I think it allows for the decentralization of health care to some degree.
Q: How are these data being used or implemented?
Precision health has actually been around for some time. For example, we’ve been using genetic analysis for several years to optimize treatments in patients with cancer. But now, there are more consumer precision health tools on the market, such as “23andMe” and other genome sequencing services that can provide information about your susceptibility to certain diseases and things like that.
We’re using devices to target heart health, too. In patients with hypertension, for instance, we have blood pressure monitors that patients can use at home. We’re trying to extract more information from patients out of the clinic and in their everyday lives.
Q: In the vein of heart health devices, let’s talk about the preliminary results from the Apple Heart Study that were recently released. In this study, patients used the Apple Watch to identify irregular heart rhythms such as atrial fibrillation (Afib). Can you unpack some of the key findings from the study?
The first point to note is that study participants wore Apple Watch models 1, 2, and 3 to look at heart rate variability. The study did not include the most recent iteration of the watch, Apple Watch 4, which has a built-in electrocardiogram (ECG) monitor.
The study investigators wanted to see if a mobile app on the watch could detect episodes of Afib using data from a heart rate sensor. Because Afib is a disorder that comes and goes, the researchers sent some patients a heart rate patch to wear on their skin after their watches detected Afib. Roughly 30% of the patients who were wearing both the watch and the patch were found to have Afib.
The watch had a high positive predictive value, meaning that when it detected Afib, the patients actually had Afib 70-80% of the time—although again, because Afib is so transient, all episodes may not have been identified. A very small percentage of the patients received Afib notifications (0.5%), so there weren’t a lot of “false positives,” either.
Q: What do these findings tell us?
What they tell us on a basic level is that a wearable, like an Apple Watch, has the ability to detect Afib. It’s important to identify Afib because it can lead to stroke.
So, what do we do with this data? Well, we already know that patients with paroxysmal Afib (Afib that occurs off and on) benefit from blood thinners (anticoagulation). Generally, we say that the primary risk of anticoagulation is a major bleed while the major benefit is stroke prevention. We also say that a patient needs to fall roughly 200 times for the risk of a serious bleed to outweigh the benefit of preventing stroke. But, of course, we haven’t had this level of granularity before where patients can be monitored 24/7— that’s an exciting development. At the same time, we’re entering uncharted territory with this continuous monitoring and we’re going to have to reassess if risk/benefit analyses favor using anticoagulation in these patients. I’m sure researchers will be doing that in the future.
Q: Afib typically doesn’t affect younger people. What does it mean if a younger patient receives an Apple Watch alert?
I think that looking at probability is important when it comes to Afib. Younger patients (under 55 years old) are less likely to have Afib. Of course, risk factors like hypertension, diabetes, and other conditions can increase those odds, so an alert for a young patient should still be taken seriously.
If your watch tells you that you’re in Afib—even if you’re under the age of 55—it’s important to contact your primary care provider or cardiologist and let them know. They should have a discussion with you about the nature of the episode(s) and symptoms and should address any other medical conditions that may increase your risk for Afib. Some clinicians may want to further investigate with an event monitor. If Afib is detected, the patient and clinician will have a conversation about anticoagulation treatment.
Again, with the Apple Watch 4, you can download a PDF of a heart rhythm strip and send it to your healthcare provider, which can be helpful in guiding discussion and management.
Q: Are these devices actually reliable or is it too early to tell?
Newer devices are reliable in that they can actually detect Afib—the sensitivity and specificity are over 98%. I think there’s definitely a signal benefit. It’s too early to tell if clinicians are seeing lots of patients who are receiving Afib notifications, but again, only 0.5% of participants in the Apple Study were notified that they had Afib. I personally haven’t encountered a big uptick of patients who’ve received alerts.
Q: One of the drawbacks of these technologies is that they can’t really tell a patient’s full story. They don’t account for genetics, environmental factors, etc. What are your thoughts on this?
Right, these devices only offer one piece of evidence. Of course, clinicians should always take the whole patient story into consideration and patients should be forthcoming about their histories. If you have a history of heart disease, high blood pressure, or other risk factors for Afib, those things should be discussed.
Again, age is also a factor to consider with Afib. If you’re a healthy 20-year-old patient, I’d think twice about Afib and would need to inquire about your medical history and order additional studies if warranted. I definitely agree that devices only represent one piece of data.
Q: Is there anything else to be aware of when purchasing one of these devices?
We know these devices can be really helpful in chronic conditions like hypertension or diabetes. They can help patients collect organized metrics, which, in turn, can help clinicians optimize treatment and management. On the other hand, people should be aware that these devices can identify abnormal values that aren’t truly meaningful. Our bodily signals change a lot throughout the day with exercise and activities. These devices will make us more aware of how our bodies are operating, which is exciting but may also pose challenges.
The bottom line is that patients should listen to their bodies and talk to their providers if they have specific questions or concerns. Likewise, providers will have to work on filtering data from wearable devices and determining if it’s clinically meaningful or not. We need to be careful about overburdening the healthcare community with potentially spurious results. I think we’ll all continue to learn more as these devices become more widespread.
Dr. Yazdi is an internal medicine resident interested in the intersection of digital health and clinical medicine. His current research focuses on noninvasive devices and their role in promoting cardiac health. He currently serves as the Clinical Innovation Lead in the One Brave Idea program at Brigham and Women’s Hospital.
*Originally published in The Beat – June 2019. Read the full newsletter here.