Voice Analysis Moves Toward Predicting Heart Failure

Cordio Medical, providers of the first cardiac technology company to use voice and speech analysis for heart failure monitoring, announced that in an independent test of its HearO system, DETECT-HF found that HearO predicted more than 70 percent of heart failure events at an average of 24.6 days before they occurred, with high patient compliance and no device-related adverse events.

Patients with heart problems might experience gradual accumulation of fluid in the lungs that can eventually require hospitalization. HearO can detect changes associated with fluid accumulation in the lungs through short daily voice recordings on standard smartphones. HearO combines speech signal processing and artificial intelligence to build patient-specific voice models from short daily recordings. The system uses this model to detect changes associated with fluid accumulation in the lungs and notifies the care team through a secure web-based portal. The patented technology is language-independent.

"As clinicians, we have long been able to hear the change in a patient's voice when they arrive at the hospital decompensated, but at that point it's already too late," said Dr. William Abraham, College of Medicine distinguished professor at The Ohio State University Wexner Medical Center, heart failure specialist, and lead author of the study, in a statement. "DETECT-HF validated that an AI-assisted voice analysis biomarker can identify deterioration weeks before a heart failure event, with sensitivity well above that reported for daily weight monitoring and comparable to invasive monitoring technologies, without requiring an implanted device or additional hardware. This opens the possibility of bringing effective remote monitoring to a much broader heart failure population, giving healthcare practitioners an opportunity to intervene before patients require hospitalization."

In DETECT-HF testing, patients completied 85 percent of their scheduled daily recordings and ranked the system with mean usability ratings of 4.7 to 4.9 out of 5. Additionally, the rate of unexplained alerts was 2.65 per patient per year, below the study's predefined goal of fewer than three.

"Nearly a decade ago, Cordio set out to pioneer the use of speech as a biomarker, and DETECT-HF marks the first blinded, prospective validation of AI-assisted voice analysis for heart failure monitoring," said Tamir Tal, CEO of Cordio Medical, in a statement. "The high adherence rate we saw in DETECT-HF reinforces our vision of empowering patients to make heart failure monitoring a manageable part of their daily routines. HearO's compatibility with easily accessible standard smartphones, combined with its ease of use and deployment, creates the potential to bring effective monitoring to large and diverse patient populations, with the ultimate goal of reducing hospitalizations and improving quality of life at scale."

HearO has received Breakthrough Device designation from the U.S. Food and Drug Administration, is CE marked in Europe, and approved for use in Israel. The results from DETECT-HF will support Cordio's pursuit of FDA marketing authorization through the De Novo pathway ahead of a planned U.S. launch in 2027.