Trial of AI-assisted voice analysis to identify changes in heart failure

A doctor holding a paper heart

Australian researchers will trial AI-assisted voice analysis to evaluate breathlessness and disease severity in people with heart failure with the aim of providing a cheap and accessible home monitoring tool.

Researchers from the Baker Heart and Diabetes Institute and The Alfred hospital want to recruit 200 Australians with heart failure who can regularly record their voice on a smartphone for up to a year so their patterns of speech and breathing can be analysed by artificial intelligence and then validated by cardiologists.

The most common symptom of heart failure is feeling short of breath — either with rest or exercise — that’s why researchers want to examine novel and user-friendly ways to evaluate breathlessness in people with heart failure.

Head of the new Monash Alfred Baker Centre for Cardiovascular Research, Professor David Kaye says participants will be asked to use an app on their mobile phone developed by the Centre’s clinicians and engineers to record their voice for one minute on a weekly basis, which will be analysed using artificial intelligence.

Professor Kaye says they want to test the software over a long period and validate it against the clinical assessment of cardiologists to see how it stands up. The software has previously been used in a small study which involved a one-off voice recording including of people hospitalised with heart failure. This small study showed good correlation between symptoms and voice changes.

While heart failure is common in older Australians, Professor Kaye says it is not inevitable. However, he says if people develop heart failure, it needs to be managed well or it can result in multiple hospital admissions and complications. Heart failure can damage your liver or kidneys and can lead to pulmonary hypertension or other heart conditions, such as an irregular heartbeat, heart valve disease, and sudden cardiac arrest.

“Our aim is to help people with chronic heart failure access an easier and cheaper home monitoring tool that can save lives and enhance quality of life,” Professor Kaye says.

“We think this tool could be particularly useful for people in rural and remote areas where there may be less opportunity to see clinicians or cardiologists,” he says.

“Artificial intelligence may also assist in identifying more subtle signs of deterioration and respiratory changes which might not be audible by humans or might be missed by less experienced clinicians.”

/Baker Institute Public Release. View in full here.