AI turns wearable device data into actionable insights for heart failure patients
New early clinical data from a study using artificial intelligence (AI) to monitor widely available consumer wearables may help transform how heart failure patients are monitored and managed outside the hospital.
At the THT 2026 meeting in Boston, Afnan Tariq, MD, JD, an interventional cardiologist and assistant clinical professor of medicine at the University of California, Irvine, presented first-in-man results from a passive, device-agnostic AI platform designed to turn data from consumer wearables into actionable clinical insights. The technology was able to lower the number of hospitalizations required over time thanks to earlier interventions.
Tariq spoke with Cardiovascular Business about the study as well as other issues in heart failure management AI may be able to solve.
“The idea of disseminating technology to people who are so busy with day-to-day work is an important task,” Tariq said, describing the motivation behind the platform. “Technology has abounded in consumer's hands and it is everywhere. We as cardiologists—as physicians, frankly—struggled to keep up with that demand or really understand it.”
It can be difficult and time consuming to integrate wearable data into patient records or comb through that data to find actionable information. However, AI offers the ability to sort through the data and save valuable time.
Bridging the data gap between patients and clinicians
Consumer devices such as smartwatches and fitness trackers have long been capable of collecting continuous physiologic data, including heart rate, activity levels and oxygen metrics. However, integrating that data into clinical workflows and making it actionable has remained a major challenge. Tariq said both patients and clinicians often lack the tools to interpret this information effectively.
“A lot of this stems from patients wanting to advocate for themselves. There's a fundamental agency problem in medicine. Patients are at home, they don't know what to do with the data,” he explained. “When it comes to physicians, they don't know what to do with the data.”
The platform his team developed aims to solve that problem by aggregating and analyzing data from multiple devices, then presenting insights in a format that can be used in routine care.
“What we built is a platform that can help bridge that gap and help translate consumer data into actionable workflows,” Tariq explained. “It's not a medical device, but we leverage FDA-cleared devices to deliver and help understand that data to really know what's going on with patients between visits.”
Early feasibility study shows encouraging results for AI home monitoring
The initial study was a single-center, real-world observational pilot involving 71 heart failure patients, representing more than 108 patient-years of data. Participants used their own consumer devices with no mandated wear time.
The platform collected approximately 75 million physiologic measurements, with a median signal availability of 98%, demonstrating strong real-world usability.
“When you have high-frequency data, you're able to see things earlier,” Tariq said. “Even if you were not to use AI just to be able to visualize the data, it's important. But then when you can use AI to understand the data, to build the supportive evidence base that allows you to act with confidence, it helps clarify what is happening.”
A retrospective analysis showed all-cause hospitalization rates of 0.11 per patient year. This is substantially lower than historical Medicare heart failure rates, which range from rates of 0.87 to 2 per patient year.
“The math speaks very much for itself,” he said. “When clinicians are empowered with insights and able to act earlier, you're able to have a durable impact.”
Moving beyond single-device monitoring for a more holistic view of heart failure
Unlike traditional remote monitoring systems that rely on a single implantable or wearable device, the AI platform aggregates data across multiple sources, including smartphones, smartwatches and even implantable devices when available.
“We have always been trying to create one single device that can create one point of measure that is actionable,” Tariq said. “What we've shown is that gives us some signal, but there's limited utility. When you can create software that transcends that or that is able to collate and collect all of the clinically actionable data and then use AI to support clinicians when appropriate to take action, I think that's really the exciting portion.”
The system is fully device-agnostic, integrating with commonly used consumer products without requiring patients to adopt new hardware. Patients in the study used an array of devices they already had and used on a regular basis, including technology from Apple, Fitbit and Samsung. No new devices were issued.
Real-world impact of identifying deterioration earlier
Tariq shared a case example illustrating how the platform can detect clinical deterioration earlier than traditional care pathways. In one instance, an 88-year-old patient flagged worsening symptoms and physiologic changes through wearable data.
“In four minutes, I was actually able to see retrospectively 440,000 historical data points,” he said. “It doesn't take AI to see things. It takes translating things with technology.”
After reviewing the data and alerting the care team, it was determined the patient had pulmonary hypertension as a result of heart failure and they received timely treatment adjustments, avoiding further hospitalizations.
“It's just the ability to listen and make the consumer device data actionable,” Tariq said. “And then all of those supportive workflows, that's where AI can really help.”
Implications for heart failure prevention and value-based care
Beyond managing existing heart failure, the technology could also support earlier detection and prevention strategies—an area of growing focus in cardiology. He noted that heart failure costs about $35,000 per patient annually, with roughly 75% of those costs tied to hospitalizations. It is a major driver for Medicare expenses each year, costing the U.S. health system about $179.5 billion annually.
“I think it's a tool that's incredibly helpful for that,” Tariq said, referring to the earlier identification of at-risk patients. “The AI surfaces the patients that have rising risk in a way that you can understand and validate, then that's something you can act on.”
As healthcare systems increasingly shift toward remote monitoring and value-based care models, platforms that can convert the growing flood of patient-generated data into actionable insights may play a key role in reshaping chronic disease management. Tariq said AI may be able to play a role in the new Ambulatory Specialty Model (ASM) payment model for heart failure that begins on Jan. 1, 2027.
The new program aims to improve prevention and upstream management of high-cost chronic diseases with an initial push in heart failure and lower back pain. The payment model is an attempt to reduce avoidable hospitalizations and unnecessary procedures. Participation in the ASM will be mandatory for certain specialists who commonly treat these conditions in Medicare patients in an outpatient setting.
Dave Fornell has covered healthcare for more than 17 years, with a focus in cardiology and radiology. Fornell is a 5-time winner of a Jesse H. Neal Award, the most prestigious editorial honors in the field of specialized journalism. The wins included best technical content, best use of social media and best COVID-19 coverage. Fornell was also a three-time Neal finalist for best range of work by a single author. He produces more than 100 editorial videos each year, most of them interviews with key opinion leaders in medicine. He also writes technical articles, covers key trends, conducts video hospital site visits, and is very involved with social media. E-mail: [email protected]