News

Article

AI App Predicts Heart Failure Early by Listening to People Talk

Author(s):

An artificial intelligence can detect increased fluid in lungs by simply hearing voice changes over time—and the increased fluid in lungs can be a sign of heart failure.

AI App Predicts Heart Failure Early by Listening to People Talk

Credit: Pexels

In a recent study, a smartphone app using artificial intelligence (AI) predicted heart failure 3 weeks before hospitalization—all because of a person’s voice. The late-breaking science was presented Monday, November 13, 2023, at the American Heart Association’s (AHA) Scientific Sessions 2023 in Philadelphia, Pennsylvania.1

AI is growing—and this AI heart-failure-predicting app is just another growing piece. AI is becoming more prevalent in the medical field with aiding diagnostics. According to the Southern Medical Association, approximately 5% of outpatients in the U.S. in 2021 receive an incorrect diagnosis, so AI helps to assist diagnoses. Three years ago, scientists at Babylon developed new AI symptom checkers to help cut diagnosis mistakes in primary care.2 So, while using AI for diagnoses is not brand-new, the new app in the study can improve the treatment for heart failure specifically.

The app uses an AI technology called the Cardio HearO® system to detect changes in the voice over time. Speech measures include pitch, volume, dynamics, and other characteristics. Voice changes may indicate increases of lung fluid, a sign of progressing heart failure. In the study, the app predicted more than 75% of hospitalizations about 3 weeks before it happened.1

“Speech analysis is novel technology that may be a useful tool in remote monitoring of heart failure patients, providing early warning of worsening heart failure that frequently results in hospitalization,” the lead investigator William T. Abraham, MD, FAHA from The Ohio State University Wexner Medical Center in Columbus told the American Heart Association in a press release. “This technology has the potential to improve patient outcomes, keeping patients well and out of the hospital, through the implementation of proactive, outpatient care in response to voice changes.”

The investigators conducted the study from March 2018 through April 2023. The study included 416 adults in Israel diagnosed with heart failure. Most of the participants were male (75%), and the average age was 68 years old. Every participant recorded 5 sentences in their native language—Hebrew, Russian, Arabic, or English—into the app daily. The training phrase of the study included 263 participants to develop the AI algorithm, but then only 153 participants were used to test the tool’s effectiveness.

The app predicted 76% of worsening heart failure about 24 days before hospitalizations in the training phase of the study. The app only had 3 unnecessary alerts per patient, per year.

As for the validation phase, the app was 71% accurate in detecting heart failure in roughly 3 weeks. Like during the training phase, there were also 3 unexplained alerts per patient per year.

While the findings revealed the AI app can predict heart failure early, one limitation was that the participant sample had been small. Though, an ongoing U.S.-based study continues to train and validate the Cardio HearO® technology.

References

  1. AI-Phone App Detected Worsening Heart Failure Based on Changes In Patients’ Voices. American Heart Association. November 13, 2023. https://newsroom.heart.org/news/ai-phone-app-detected-worsening-heart-failure-based-on-changes-in-patients-voices?preview=c9dd. Accessed November 13, 2023.
  2. Glick, R. Artificial Intelligence in Medical Diagnosis. Southern Medical Association. October 7, 2021. https://sma.org/ai-in-medical-diagnosis/. Accessed November 13, 2023.

Related Videos
The APAC Recap: Peripheral Artery Disease at CAPP Live 2024 with Bob Ross, PA-C | Image Credit: APAC
John Stone, MD, MPH: Continuing Progress With IgG4-Related Disease Research
AMG0001 Advances Healing in CLTI with David G. Armstrong, DPM, PhD, and Michael S. Conte, MD | Image Credit: Canva
Malin Fromme, MD | Credit: RWTH Aachen
Pavel Strnad, MD | Credit: AASLD
Philip Conaghan, MBBS, PhD: Investigating NT3 Inhibition for Improving Osteoarthritis
Gideon Hirschfield, FRCP, PhD | Credit: UHN Foundation
Rheumatologists Recognize the Need to Create Pediatric Enthesitis Scoring Tool
Alexei Grom, MD: Exploring Safer Treatment Options for Refractory Macrophage Activation Syndrome
Jack Arnold, MBBS, clinical research fellow, University of Leeds, Leeds Institute of Rheumatic and Musculoskeletal Medicine
© 2024 MJH Life Sciences

All rights reserved.