FDA approves AI-powered monitoring device

A medical device for heart failure patients which uses artificial intelligence (AI) and machine learning was approved October 3 by the FDA, according to a news release from the manufacturer.

Biovitals Analytics Engine, manufactured by Biofourmis, received 510(k) clearance as part of the FDA’s continued recognition of machine learning and AI in the Software as Medical Device category. The company also received approval in May for Biovitals RhythmAnalytics, which they are calling the “Biovitals ecosystem,” the release said.

"The Biovitals Analytics Engine helps to fill a critical unmet medical need in heart failure care," says Raj Khandwalla, MD, MA, assistant professor at the Cedars-Sinai Smidt Heart Institute, director of Cardiovascular Education at the Cedars-Sinai Care Foundation, and Biofourmis Clinical Advisory Board member. "Numerous studies consistently demonstrate that guideline-directed medical therapy is underutilized in patients with heart failure, which is likely the reason that historic decreases in morbidity and mortality have not only plateaued, but have unfortunately begun to increase. The Biovitals Analytics Engine within the context of the BiovitalsHF platform will likely lead to improved clinical outcomes."

Biovitals Analytics Engine receives physiological data using FDA-approved sensors and uses AI and machine learning to identify correlations between vital signs and a heart-failure patient’s daily activities. It uses this information to alert physicians to changes in the patient’s vital signals from their baseline, the release said.

The company says the device may be able reduce hospital readmissions and decrease the need for emergency department visits.

Biofourmis’ application for clearance on the new software was based on clinical studies which monitored patients with complex chronic conditions like heart failure, COPD, and atrial fibrillation at home, the release said.

One such study showed an earlier clinical intervention in a 73-year-old man by predicting heart failure decompensation 12 days before a heart failure-related hospitalization who suffered from multiple comorbidities like hypertension, COPD, and diabetes. It achieved this by combining multiple physiology signals allowing physicians to step in.