Digital Health in the Clinical Practice: Benefits, Challenges, and Updates About the REACT-AF Trial

© 2023 HMP Global. All Rights Reserved.

Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of EP Lab Digest or HMP Global, their employees, and affiliates. 

EP LAB DIGEST. 2023;23(9):1,10-12.

In this episode, we feature a discussion with Rod S Passman, MD, and Hamid Ghanbari, MD, about how digital health is changing the cardiac electrophysiology (EP) practice, the benefits and pitfalls of current technology, challenges with data flow, and questions to be answered in the REACT-AF trial.

Hamid Ghanbari, MD: It is wonderful to be here. My name is Dr Hamid Ghanbari. I am a cardiac electrophysiologist at the University of Michigan and the vice chair of innovation at the Cardiovascular Center. I am excited to be talking with one of my colleagues and mentors, Dr Rod Passman.

Rod S. Passman, MD: Thank you. It is a pleasure to be here and share this with you. I am Rod Passman. I am a cardiac electrophysiologist at Northwestern University in Chicago and the director of the Center for Arrhythmia Research at Northwestern.

We are going to be talking about digital health, how we use this in our practices, and some of the challenges that we are facing. Then, we will hopefully have some time to talk about the REACT-AF trial, which brings in digital health not only to detect disease, but to manage it, which is important to all of us. Hamid, you have been a pioneer in the recognition of digital health as it pertains to the management of patients in the EP practice. How did you get interested in this and how do you use this technology on a day-to-day basis?

Ghanbari: I do not know if I would call myself a pioneer—that is a high bar! But I was interested in this technology from the beginning, mostly from a practical perspective, because my patients kept bringing these devices to me and asking me questions about which devices to use and what the information meant. So, it started out of the necessity in trying to understand how to use these devices in a clinical setting and answering patient questions. That evolved into my research interest, which was how to use this technology to have a better understanding of disease states, who was experiencing symptoms, and how the disease progresses outside of the clinical setting and visit. Over the past year or so, we have really been interested in building systems in our health care system to take in all that information, incorporate it into our clinical practice, and make it part of our medical records. So, that has been the evolution of my initial interest to what has become what we call the clinical digital health program at the University of Michigan.

It has been always focused on cardiology, and electrophysiology has been at the forefront, but it is slowly expanding to all the clinical entities at the University of Michigan. I would be happy to talk more about that.

Passman: Yes, because I think a lot of us use this technology as sort of a one-off, but very few of us have been able to build an integrated program that collects this information, gets it to the electronic health record (EHR), and uses it for clinical decision-making on a widespread level. So, I would love to hear how you did that.

Ghanbari: The clinical digital health program at the University of Michigan focuses on integrating the digital health program and aligning it with the overall mission of our institution, which is to lead in education, clinical care, and research. The research angle is something that you are very familiar with, in using these devices and developing algorithms to integrate these devices in our clinical practice. The clinical program leverages the research and tries to integrate it into clinical practice. So, the devices have really progressed at a much faster pace than I had anticipated, with having on-demand electrocardiogram (ECG) devices such as the Kardia (AliveCor), to now where we can detect atrial fibrillation (AF) with a high degree of accuracy and get an overall burden of AF over a monitoring period. Our program aims to use that information.

I will use the example of patients with AF, which is something both you and I, as well as our audience, are very familiar with. Patients are given devices to manage their AF. Through that device, they are then enrolled in our clinical digital health program. They send us information, which is typically their symptoms and recordings of their ECG. That is queued up in our EHRs and an assigned clinician reads them on a weekly basis. That aspect is similar to what we do with loop recorders, where patients are monitored over a 30-day process.

But what is really important to make these actionable is from a system perspective, one must show that this information can improve access. So, we could keep track of how many repeat visits are not needed. For example, after an AF ablation, patients are in sinus rhythm and do not need to come into our clinical office to be seen again, so those spaces open for new patients to be seen. Therefore, wait times have decreased and access time has significantly increased. Also, we have a dashboard that evaluates patient experience, so we can see how patients interact with the devices and what new insights they have, and then we use that information in real time to improve our care process.

That is a brief overview in AF, but the same kind of framework applies to heart failure and cardiac rehab, and to some extent, neurology.

Passman: You raise excellent points. I think this is what many of us want—for this information to come to us and be dealt with just like for remote monitoring of an implantable device. But the question is, have we shown that we save money to the system by giving patients an $80 device and spare them emergency department (ED) visits and doctor visits with the more medical-grade monitoring that we do? Because ultimately, we need to make this something that insurance companies recognize, that giving people these devices early in their AF journey helps them. Not waiting to post ablation, but titrating medications and deciding whether they need an ablation.

So, I would like to see these things given for free in the ED, at post hospitalization, or in a clinic visit. Then, we need to prove not only that patients like it better, but that the system saves money and it is worth doing. What do you think?

Ghanbari: It is important to pay attention to that aspect to show that the unit economics of this works, meaning that it improves access so you are able to increase the number of patients that can be seen, you can capture some of the remote patient monitoring codes that are generated through the interaction with these devices, so you are revenue-positive in that aspect and can cover the cost of clinicians who evaluate them. Then, you must be able to show that to the health system so there is wider adoption. We know from a clinical standpoint, there is a lot of evidence that patients like it and there is better quality of life measures associated with use of these devices. But if you can show that the unit economics also work for the system, then you can expect that wide adoption would be downstream from that.

Passman: You mentioned something early on that these devices are highly accurate, and while most of us believe that they are accurate, we have all encountered situations where there are false-positives and false-negatives, which can create a lot of anxiety. Many of these devices are not FDA cleared, and we do not really know the sensitivity and specificity of these devices. It is very difficult to discern that, because often it is not published or the studies were done in highly controlled environments. How do you deal with all of this? I do not think I know the answer. Sometimes it can be difficult to sort out a single-lead ECG, whether it is sinus with PACs versus AF.

I think all of us believe that if a patient presents with a photoplethysmogram diagnosis of AF, and the irregular rhythm notification on an Apple Watch or Fitbit shows AF, we need to see the ECG before we treat. The question is, if you saw a single-lead ECG that was read as AF, I am assuming that we all would want to see that data ourselves, right? We would not trust that electronic overread. So, do you still do other forms of monitoring, and why? Or, in the right patient, do you treat based on that single 30-second ECG and then get more information? Or do you not believe that it is truly AF until you see it from your own testing?

Ghanbari: There are a lot of questions embedded in your question, and I would love to hear more about your thoughts on that. But I see these devices as being more complementary to our existing arsenal of devices, not necessarily replacing what we have, but an addition to the already existing arsenal that we have.

Obviously, we have devices that are very good. They have been used in clinical practice for a long time and can give us pretty accurate information over a 2-week period. We also have good devices that are impactful and that can monitor patients over a 3-year period.

What we do not have are non-invasive devices that can give us an opportunity to monitor patients beyond 2 weeks and up to a year. I think that information, particularly when obtaining information on burden of AF, can be very helpful. So, there is still going to be a need for confirmatory tests like the 2-lead patch ECG monitoring.

As far as accuracy of these devices, you are right that the published literature leaves a little to be desired. Part of it is that we as clinicians should demand that more be published in a variety of different populations so we can have a better idea of the actual performance of these devices for detection and estimation of AF, and have a better understanding of how to apply them in clinical practice.

For example, a lot of new information is coming out that in certain populations, particularly in darker skinned populations, the devices do not perform very well. We are also learning that when doing high levels of activity, these devices do not perform very well. We need to understand this information more so we can better use these devices. But every time a new diagnostic test becomes widely available, part of our understanding must evolve as well.

Passman: We are all taught not to take things at face value—we want to see the primary data ourselves. Particularly with the consumer grade device, if a patient has multiple risk factors, is symptomatic, has a 30-second ECG of AF, and 2 hours of palpitations, I will believe that. I may still want to get a monitor to see what their burden is, post-conversion pauses, and rates during sinus versus AF—that is critical data. But a 30-second ECG, whether it is from a patch or consumer electronic device, I take for real.

We have to believe that those 30 seconds are part of a much larger episode. We are not going to treat for 30 seconds. But it is a great opportunity to reframe how we think about AF and look more at things like AF burden and burden reduction, and then start to correlate burden with the outcomes that we care about, not just stroke, but heart failure and cognitive decline. Because we have done ourselves a disservice by looking at the success rate of an intervention as time to recurrent AF and defining that as 30 seconds.

The good thing is that these devices do not pick up very short episodes. Certainly, you need dozens of minutes or an hour or more depending on the device. So, we do not want devices that are going to pick up very short episodes, because they are going to be highly susceptible to PACs and PVCs that create false-positives. I would like to see it incorporated more into the research that we do, evaluations of new technologies, and more. We all know that snapshots of the rhythm, whether it is 1 week, 2 weeks, or even a 30-second ECG once a day is not an accurate assessment. Because of the density of these episodes, we need long-term monitoring, and ideally, long-term non-invasive monitoring, to better understand this disease and the impact of treatment.

Ghanbari: That is true. The most important thing I wanted to highlight from your comments was that our traditional way of thinking about anticoagulation has to evolve. We must innovate and think beyond the CHA₂DS₂-VASc score. In that 30-second ECG, we have to put it in the context of the patient. Maybe we do not need to anticoagulate people all the time during the whole process. Risk is more dynamic than we used to believe. I think you are testing that in a large trial coming very soon.

Passman: Yes, so that is a great introduction. We are about to start enrollment in a clinical trial called REACT-AF. This is an NIH-funded trial that will compare the current standard of care of chronic oral anticoagulation, whether a novel or direct oral anticoagulant, versus an experimental approach, which is a pill-in-pocket approach using a wearable, AF-sensing smartwatch. The theory is that if you anticoagulate patients for only a fixed period and only in response to a prolonged episode of AF, that you can protect them against stroke during the high-risk period while minimizing their exposure to oral anticoagulation, and in doing so, reduce bleeds, reduce costs, and improve quality of life.

This came out of the clinical experience of over-anticoagulating people. I have seen adverse events happen to patients who, from my perspective, were in sinus rhythm all the time because of a drug or an ablation, and yet, because of their CHA₂DS₂-VASc score, the guidelines said they should be on chronic oral anticoagulation. I was quite motivated by a patient who had a life-threatening intracranial hemorrhage as a result of the current standard of care strategy.

We have all faced the issue where we have a patient on a drug or we do a cardioversion or ablation, and the patient comes back to us 6 months, 12 months, or years later, still on their oral anticoagulants with no evidence of recurrence. Since patients can now monitor themselves for recurrence, it is always a difficult discussion to have. Why does that patient need to be on oral anticoagulation for the rest of their lives?

My hope is that we can leverage these advances in digital health and pharmacology, and give a more targeted approach to stroke prevention, not the “one size fits all” based on CHA₂DS₂-VASc score alone that has no regard for AF burden. We need to rethink the role of AF in stroke and recognize that all of what we know is based on the clinical diagnosis of AF. These patients came into their doctor’s office with essentially enough AF to be picked up on a physical exam or an ECG, but that does not take into account people who have a “successful” rhythm control strategy or those that have very low AF burdens to begin with. Is that the same AF as the AF that was used to define the CHADS₂ or CHA₂DS₂-VASc scores that were originally seen in the Framingham study showing the association from stroke and AF?

We do not understand this disease well enough, and yet, we are applying recommendations based on a disease that has evolved as our ability to monitor patients has evolved.

Ghanbari: Definitely. Once the results are available, I think this trial will not only enhance our understanding but will change the way we will think about this disease altogether. I do not believe there is any other trial right now going on in AF that has a bigger potential impact on our understanding and on the treatment of this disease. So, I am excited and anxiously awaiting the trial results.

Passman: Well, we will have to talk again in about 5 years, because I think that is how long it will take. There are some trial results that are quite interesting that are not published yet but have been announced. We know that the NOAH trial, which evaluated device-detected AF of less than 24 hours randomized to anticoagulation, was terminated for futility.

The ARCADIA trial was also terminated. This was a study of patients who had suffered an ischemic stroke who had evidence of a left atrial myopathy either based on left atrial size, BNP, or terminal P-wave forces in V1. So, these were non-invasive markers of a clinical “myopathy.” The theory was that the myopathy causes the stroke and the AF is simply a marker. That trial was terminated for futility, suggesting once again that the only patients who seem to benefit from anticoagulation when it comes to the heart and stroke are patients with AF. As we know, those markers for left atrial myopathy are also markers for AF.

So, to me, it bolsters the argument that it is the AF that causes the stroke. Yes, it is more likely to cause stroke in patients with vascular risk factors who also have a myopathy. But the question is, if we eliminate or control the AF, can we reduce or eliminate the need for chronic anticoagulation? I think either way, we will learn a lot from the results of this trial.

This is just an incredibly fascinating time for our field, not just in the therapeutics that we have been able to offer and will be able to offer, but as you mentioned, in the way we define and manage this disease. Just think of how far we have come in the last 25 years. The future is really exciting. Do you feel the same way?

Ghanbari: Yes. I am more excited about this field now than I ever was in the beginning. There is a recognition about this important part of innovation that is happening in our field currently with the new therapeutics and modalities being available. As we improve on our algorithms for detection of these arrhythmias, I think the future is bright. 

The transcripts have been edited for clarity and length.