Advertisement
Columbus, Ohio – Could the sound of your voice reveal hidden heart problems? A new speech analysis app predicts worsening heart failure weeks before symptoms appear. The app, used by heart failure patients at home, detects fluid in the lungs an average of 21 days before an unplanned hospitalization or escalation in outpatient drug treatment.
“The current standard of care is simply not good enough to keep heart failure patients healthy and out of the hospital,” study co-author William Abraham of Ohio State University said in a statement. “The system tested in this study was able to predict 80 percent in advance of worsening heart failure, compared to a 10 to 20 percent success rate for daily weight monitoring shown in previous studies. In the future, speech analysis could be used along with other clinical information to modify treatments before a patient’s condition worsens, thereby avoiding hospitalization.”
Abraham explains that in people with heart failure, the heart doesn’t pump blood around the body as well as it should. Fluid is not properly eliminated by the kidneys, and the excess fluid accumulates in the lungs or legs. Lung congestion is a common cause of hospitalization and can be life-threatening.
Currently, pulmonary congestion is monitored by asking patients to weigh themselves each day and to report any significant increase. Patients are also asked to report worsening symptoms such as shortness of breath, needing to raise their heads at night to breathe comfortably and sleep, and swelling in their feet or ankles.
The study included 180 heart failure patients who were taking medication recommended by the guidelines. At the beginning of the study, the participants recorded five sentences on a standard smartphone using the speech analysis app. During the study period, patients recorded the same five sentences with the app each morning before breakfast.
The app compared each day’s recordings to the original versions and alerted research staff when it detected lung congestion. “The system establishes a baseline for each individual patient during a period of stability,” says Abraham. “Then, over time, it detects changes in speech that indicate fluid in the lungs.”
Researchers examined whether the pulmonary congestion alerts predicted heart failure events, defined as at least one worsening symptom that led to hospitalization or an escalation in outpatient medication management. They did this by comparing the date of the alert to dates of subsequent heart failure events. “True alerts” were those issued within a 31-day window prior to the heart failure event. “False Alerts” were those issued outside of the 31-day window before an event.
Professor Abraham notes that a “true alert” would continue to be given daily until the heart failure event occurred.
“In this study, the alerts were not edited by clinicians because the goal was to determine the diagnostic performance of the tool. In real life, notifications would be sent to the patient’s doctor pending a response,” he explains. “Here we observed patients longitudinally for worsening heart failure events and then looked to see if there was an alarm prior to that event – ie a ‘true alarm’. If there was no alert, the system could not predict the event. Alerts not followed by worsening heart failure were ‘false alerts’.”
During the study, patients provided records for an average of 512 days. A total of 49 heart failure events occurred in 37 patients, of which 80 percent were correctly predicted by the app and 20 percent were missed. True alerts were issued an average of 21 days before symptoms worsened. Each patient received an average false alarm every 4.8 months, resulting in an average of 2.5 false alarms per year.
“In this community-based study, a speech analysis app was able to predict most cases of worsening heart failure well in advance, with very few false positives. Weight gain and symptoms come too late to allow for medical interventions that keep patients out of the hospital,” says Abraham. “Future studies will examine whether changing patient management after an alert, such as increasing the dose of diuretics to shed excess fluid, can prevent hospitalizations.”
The results were presented at Heart Failure 2022, a scientific congress of the European Society of Cardiology (ESC) in Madrid, Spain.
South West News Service writer Stephen Beech contributed to this report.
