The app was able to predict hospitalization needs around three weeks ahead of a heart failure event in a trial including over 400 adult patients with heart failure in Israel.
More than 75% of hospitalizations were predicted by a smartphone app that used artificial intelligence to identify changes in a heart failure patient’s voice, roughly three weeks in advance, according to late-breaking research presented today at the American Heart Association’s Scientific Sessions 2023. The conference, which takes place in Philadelphia from November 11–13, is a prominent worldwide forum for exchanging the most recent findings, research, and updates on evidence-based clinical practice in cardiovascular science.
Lead study author William T. Abraham, M.D., FAHA, a professor of medicine, physiology, and cell biology as well as a College of Medicine Distinguished Professor in the division of cardiovascular medicine at The Ohio State University Wexner Medical Center in Columbus, stated that speech analysis is a 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. “By implementing proactive outpatient care in response to voice changes, this technology has the potential to improve patient outcomes, keeping patients well and out of the hospital.”
When the heart muscle is unable to pump enough blood to meet the body’s requirements for oxygen and blood, heart failure results. This may cause exhaustion, edema, dyspnea, and in rare cases, persistent coughing.
This study assessed how well a mobile app powered by artificial intelligence might forecast deteriorating heart failure in patients with heart failure before they require hospitalization or intravenous therapy. The smartphone app was created to track alterations in patients’ speech metrics over time. The changes in speech could be a symptom of early lung fluid accumulation, a warning of worsening heart failure.
Between March 2018 and April 2023, 416 Israeli people with a diagnosis of heart failure were included in the study. Every day, study participants entered five sentences onto the phone app in either Hebrew, Russian, Arabic, or English, depending on their home tongue. During the study’s training phase, the AI system was developed using unique voice measures from 263 participants. The tool’s efficacy was then confirmed by applying the algorithm to the remaining 153 subjects.
During the trial’s training phase, the app correctly identified 76% of heart failure events that worsened (44 out of 58 heart failure events), typically 24 days prior to hospitalization or the requirement for intravenous fluids. Every patient received three pointless alerts from the app on average year.
During the validation stage, 10 out of 14 heart failure episodes were detected by the app with a 71% accuracy rate, around three weeks ahead of time. In this group as well, there were roughly three unexplained warnings per patient annually.
The system accurately predicts future bouts of worsening heart failure with a low rate of needless notifications, according to research findings. The AI tool is validated as a potentially useful means of reducing hospitalization and improving patient outcomes due to its high accuracy rate and early notification of increasing heart failure.
Read More: https://eurohealthleaders.com/
