Every second of every day, a staggering amount of health data is streaming from devices on and inside of people’s bodies. About one in three American adults, for example, wear a smart watch or band to track their health and fitness; one in 10 Americans are implanted at some point in their lives with a medical device, many of which transmit wireless data. But what happens to this rich trove of health information once it reaches the remote servers we’ve come to call “the cloud?”
Artificial intelligence (AI) is giving the healthcare industry the ability to turn data into valuable insights. “AI technology is helping us move from the old model of reactive medicine to proactive and even predictive medicine,” says Kenneth Stein, M.D., Boston Scientific senior vice president and global chief medical officer. “It’s already enhancing doctors’ ability to provide quality care,” including by enabling earlier diagnoses, simplifying treatment decisions and accurately predicting medical emergencies, allowing physicians time to intervene.
At Boston Scientific, we’ve been applying AI to medtech for nearly a decade. Here are two examples of how we’re currently using AI to help people with cardiovascular disease.
Predicting heart failure weeks in advance
Heart failure is a long-term condition in which the heart gradually becomes unable to pump enough blood to meet the body’s needs. In the developed world, heart failure causes more hospitalizations than any other medical condition – more than a million per year in the U.S. alone. It’s also hard to identify early. Symptoms (including shortness of breath, persistent coughing and fatigue) can be subtle enough that some people don’t recognize their symptoms until too late, requiring hospitalization. In addition, patients typically go through long periods of relative stability punctuated by quick declines, making the signs of heart failure easy to miss.
AI is helping physicians identify patients at risk for worsening heart failure quite early – even before they have symptoms – with the HeartLogic™ Heart Failure Diagnostic, which became the first predictive analytic alert in an implantable device approved by the U.S. Food and Drug Administration (FDA). Heart patients who have an implanted cardioverter defibrillator (ICD) generate a huge volume of around-the-clock physiological data, including:
- heart rate
- heart sounds
- thoractic impedance (fluid levels in the chest region)
- breathing rate
- activity levels
The algorithm analyzes all of these biomarkers to track a patient’s physiological trends over time, searching for patterns that predict heart failure events. HeartLogic has been shown to predict up to 70% of cases of worsening heart failure about 34 days before an event occurs and alerts doctors, giving them time to intervene.
Monitoring 3 billion heartbeats daily for a single errant beat
An irregular heartbeat (arrhythmia) is another potentially dangerous condition that can be difficult to detect, this time thanks to the challenge of picking out a single dangerous heartbeat from a person’s roughly 100,000 beats per day. This is another area where AI is helping physicians analyze a tremendous amount of data to produce actionable, sometimes lifesaving, information.
The BeatLogic™ deep learning algorithm allows physicians to analyze data from cardiac monitoring devices worn outside the body. The device “listens” to every heartbeat in near-real time to catch any variations that could signal a dangerous arrhythmia. With cardiac data streaming in from patients all over the United States, every day BeatLogic processes over 50 million minutes’ worth of heartbeats – more than three billion heartbeats daily – and alerts doctors to the individual cases that require further attention.
Based on positive results with BeatLogic, we are beginning to test its AI power on other cardiovascular conditions. The ultimate hope is that AI can someday help predict health events across an array of medical specialties and in doing so, perhaps prevent those emergencies from ever happening.
Learn more about how we are working to address heart failure and arrhythmias.