Using AI and Biosensors to Snuff Out Heart Failure in Veterans

The week after most of the health-tech world was talking theory and sales at the gigantic HIMSS conference in Las Vegas, Nevada, a concrete example of technology’s promise was presented at a medical specialty conference in Orlando, Florida.

At the American College of Cardiology’s 67th Annual Scientific Session and Expo (ACC18), researchers from the US Department of Veterans Affairs (VA) and analytics firm PhysIQ showed how wearable biosensors and a smart monitoring platform could be used to predict hospitalizations from acute heart failure, allowing earlier interventions and potentially preventing complications or death.

Exactly 100 VA patients were given a 90-day supply of VitalConnect biosensors, which are disposable adhesive monitoring patches that are FDA-approved to be worn for 120 hours during a hospital stay or post-discharge. Participants were also equipped with a corresponding smartphone that continuously transmitted data from the devices to PhysIQ’s cloud analytics platform.

PhysIQ’s artificial intelligence (AI)-based analytics, which are also FDA-approved, were able to predict patient readmission with impressive accuracy: The predictive area under the curve for the intervention was 0.88, compared to 0.58 for a random decision generator. Specificity (85.9%) and sensitivity (84.2%) were also promising.

“The results of this study suggest a highly favorable relationship between sensitivity and specificity of event detection, as well as a sufficient warning lead time for clinicians tasked with managing patients at risk for admission for heart failure exacerbation,” lead author Josef Stehlik, MD, said in a statement. He added that study participants were highly compliant in using the patches, which are roughly 4.5 inches long and rise less than half an inch above the skin.

Stephen Ondra, MD, chief medical officer of physIQ, called the findings a potential victory for value-based care. The combination of wearable biosensors and AI shows “tremendous promise in an approach that supports a proactive care delivery model—one where clinicians can see personalized changes early enough in the deterioration process to take action that can potentially head off an acute or even catastrophic clinical event.”

PhysIQ is currently testing its tech in multiple scenarios and states of care. For the VA population, however, an intervention like this could be particularly valuable. The health system is currently facing widespread staff shortages, and heart disease is the leading cause of hospitalizations in VA facilities. Complicating comorbidities like stress and substance use are also abundant in the veteran population.

The agency has begun to embrace smart solutions like AI in recent months. In mid-February, it announced an arrangement with Google-owned DeepMind to track patient deterioration to prevent rehospitizations and fatalities. DeepMind will be given access to over 700,000 de-identified historical health records that it can use to build predictive algorithms for health deterioration.

Related

VA Plans to Use AI to Track Deteriorating Health in Veterans

Predictive Patient Surveillance System Receives FDA Clearance

VitalPatch Biosensor Receives Clearance for Longer Wear Duration