Wearable AI: A clinician’s experience
Artificial intelligence (AI) tools are rapidly becoming indispensable for many industries, including healthcare. In fact, the FDA has already approved a number of AI algorithms for clinical use, as seen in this figure.
I will focus on atrial fibrillation detection, which is now a feature of the Apple Watch, and discuss in the context of a personal story.
What lead me to purchase the Apple Watch
I purchased the Apple Watch (Series 4) about a year ago, primarily to understand how it might be useful in healthcare. Early on I tested out the atrial fibrillation detection functionality by placing my fingertip on the watch crown. It always showed sinus rhythm (see screen shot), which is normal heart electrical activity.
I had minimally invasive mitral valve repair surgery in 2005 for a progressively leaky heart valve. Following this, I had a brief episode of atrial fibrillation after surgery, which is not unusual.
Atrial fibrillation is the most common heart rhythm disturbance in adults, often associated with high blood pressure and coronary artery disease. Untreated, it frequently results in a rapid heartbeat with palpitations and occasional shortness of breath, chest pain and fatigue. Stroke is the most severe complication, which can be prevented in many cases with blood thinning medication.
I had been doing well up until about a year and a half ago, when I experienced another brief episode of atrial fibrillation, which resolved spontaneously during a brief hospital stay. No serious cause was found.
How wearable AI helped detect my atrial fibrillation
About one month ago, while sitting on a couch at home watching a movie with my wife, I began experiencing chest pain and palpitations, similar to symptoms I had previously with atrial fibrillation. This time I was wearing my Apple Watch and placed my finger on the watch crown. This time it showed atrial fibrillation (see screen shot below).
As a physician, knowing my medical history and symptoms (along with knowledge of electrical heart rhythm patterns), I agreed with the interpretation. I immediately texted the watch recording to my cardiologist along with an explanation of my symptoms.
I went to the hospital, was admitted, and again my heart returned to normal sinus rhythm by the next morning with no intervention. I was prescribed an anti-arrhythmic (heart rhythm control) medication, called flecainide, which was to be taken as needed if there was recurrent atrial fibrillation. I also take a blood thinner to reduce my risk for stroke.
Over the course of the next week, I experienced recurrent atrial fibrillation daily, which I felt and then confirmed using the Apple Watch. This information was extremely helpful to objectively monitor my heart rhythm.
It provided invaluable information when I subsequently called my electrophysiologist (a cardiologist who specialized in heart rhythm disturbances) to get his advice, which eventually included taking flecainide daily to keep me in sinus rhythm.
Soon however, atrial fibrillation recurred and persisted despite the heart rhythm medication. I was switched to a different medication and am now in sinus rhythm again, feeling well and resuming my normal activity.
Given my age and potential for long-term serious side effects of the medication, the electrophysiologist recommended a procedure called a catheter ablation to definitively treat atrial fibrillation. I look forward to having it done in several weeks and anticipate eventually returning to sinus rhythm without medication and to my day-to-day routine.
Based on my experience, I see great potential for wearable AI. Having cared for patients whose first symptom of atrial fibrillation is a disabling or fatal stroke, earlier diagnosis using these types of tools may help others avoid devastating consequences of this heart rhythm disturbance.
I anticipate additional use cases will emerge with advances in AI with potential to markedly improve the health and well-being of people worldwide.
Click here to learn more about Allscripts collaboration with Apple.