Heart problems can be heard in someone’s voice before they even know they have a problem.
A new study has found that heart conditions that need treatment can be identified by how someone sounds when they speak.
Researchers at Mayo Clinic in the United States have created an artificial intelligence algorithm that picks up small changes in speech.
Among a group of 108 people who were classified as at high risk of heart disease, nearly six in 10 who scored higher on the audio test were in the hospital within two years.
“We can’t hear these special features ourselves,” said Jaskanwal Deep Singh Sara, MD, MD, cardiology at Mayo Clinic and lead author of the study.
“This technology uses machine learning to identify something that cannot easily be quantified for us using our human brains and human ears.”
It was found that people with elevated vocal vital signs were 2.6 times more likely to develop plaques in the arteries of the heart.
They were also three times more likely to show evidence of plaque buildup on medical exams than those with a low score.
At the start of the study, X-rays assessed the condition of the heart’s arteries.
The subjects were then asked to record three 30-second audio samples using the Vocalis Health smartphone app.
The first was a processed script, the second was a positive personal experience, and the third was a negative personal experience.
The AI-based system was trained to analyze more than 80 features of audio recordings, such as frequency, amplitude, pitch and tempo, based on a training set of more than 10,000 sound samples collected in Israel.
From previous studies the researchers found six traits associated with coronary artery disease, which were then used to make one score for each person.
One-third of the patients had a high score – and at risk – with two-thirds having a low score.
Dr Sarah said: “Telemedicine is non-invasive, cost-effective and efficient, and has become increasingly important during the pandemic.
“We are not suggesting that voice analysis technology will replace clinicians or replace current approaches to healthcare delivery, but we do believe that there is a significant opportunity for voice technology to serve as an adjunct to current strategies.
“Providing an audio sample is very intuitive and enjoyable even for patients, and can become a scalable way for us to enhance patient management.
“It’s certainly an exciting area, but there is still a lot of work to be done.
“We have to know the limits of the data we have, and we need to do more studies in more diverse populations, larger trials and more future studies like this one.”
This article originally appeared on The Sun and is reproduced here with permission.