United States: New promising research revealed that people could be diagnosed with diabetes using nothing more than a voice recording from their phone.
Researchers have previously warned that millions of Brits living with the disease, which is characterized by high blood sugar levels, remain undiagnosed.
More about the news
However, as experts said, audio samples of a patient’s voice of only 25 seconds with some simple health information such as age, sex, height, and BMI could define whether a person is a diabetic or not.
About the new invention
Researchers who carried out a trial of the AI model on more than a hundred adults were able to determine that it worked in more than 70 percent of the instances.
Today, charities and experts hailed the ‘innovative’ tool, though they said more studies must be done on the same.
Abir Elbeji, who is an expert in digital health at the Luxembourg Institute of Health and study lead author, said, “Most current methods of screening for type 2 diabetes require a lot of time and are invasive, lab-based, and costly,” DailyMail reported.
Moreover, ‘Combining AI with voice technology has the potential to make testing more accessible by removing these obstacles,” Elbeji added.
“This study is the first step towards using voice analysis as a first-line, highly scalable type 2 diabetes screening strategy,” he continued.
Diabetes is a growing menace
About ninety percent of adults diagnosed with diabetes in the UK suffer from type 2 diabetes. However, yet many do not know they have the disease as, for years, the symptoms have been subtle or absent.
For diagnosis, people need to visit an outpatient GP and appear for urine and blood screenings.
In the study, the researchers examined 607 tapes to determine which acoustic substrate distinguishes non diabetics and diabetics.
All were instructed to read sentential materials that were perhaps read directly from their phone or their laptop.
More about the finding
With this, using signal processing, they were later able to differentiate between small differences in pitch, tone, and intensity, which cannot be distinguishable by the human ear.
Certain sophisticated AI methods were capable of monitoring as many as 6,000 specifics within a subject’s voice. Another was meant to concentrate on 1,024 important characteristics.
The model again had a 66 percent accuracy for the women and 71 percent for the men.
On the same account, the model delivered superior performance in females 60 years and older and in patients with hypertension, as noted by the researchers.
What more are the experts stating?
According to Dr Lucy Chambers, Head of Research at Diabetes UK, “Getting a prompt and accurate diagnosis is crucial for preventing the serious and life-limiting complications of type 2 diabetes,” DailyMail reported.
“Unfortunately, symptoms can be easily overlooked, and about 1.2 million people in the UK are currently living with undiagnosed type 2 diabetes,” she continued.
Therefore, “Using AI to develop convenient and cost-effective type 2 diabetes screening methods will help identify more people who need treatment and support, ultimately improving their quality of life and reducing their risk of long-term diabetes complications.”
“We look forward to further research on innovative AI voice analysis tools for diabetes screening,” she said.