2023-10-24 06:02:00
“The detection of this disease is essential when we know that it represents a real driver of cardiovascular diseases,” recalls Dr Antoine Bondue, president of the scientific committee of the Belgian Cardiological League and head of the cardiology department at the Erasmus Hospital. Nearly 10% of the population is currently affected in Belgium but it is estimated that this figure is twice as high given that many patients are unaware of it, they are in fact falling under the screening radar. Today, too many diabetics are unaware.”
Faced with this situation, doctors are calling for more prevention and screening to ensure better monitoring of patients. And that’s good, a new detection weapon might enrich the Belgian arsenal. Indeed, according to a study recently published in Mayo Clinic Proceedings: Digital Health, we learn that artificial intelligence would be able to determine if a person suffers from diabetes, just by analyzing the sound of their voice.
A record number of Belgians with diabetes
“We lack screening tools in Belgium”
During the tests, patients had to record messages six times a day for two weeks. In total, 18,000 audio files were analyzed to determine acoustic differences between recordings of diabetics and non-diabetics.
At the same time, the scientists combined the examination of six to ten seconds of voice with data on health (age, sex, height and weight). Using all this information, they created an artificial intelligence model that might distinguish whether this person has type 2 diabetes.
Concretely, the results are slightly more promising for women than for men with an accuracy of 89% compared to 86%. According to Jaycee Kaufman, head of the study at Klick Labs, this discovery might “transform” the way diabetes is screened. However, further research with larger and more diverse cohorts will be needed to validate the effectiveness of the technique.
Currently, to diagnose type 2 diabetes, a blood test taken on an empty stomach is necessary to measure blood sugar levels. “Artificial intelligence can be an interesting ally in the years to come because it must be said, we lack tools in Belgium to detect diseases and do more prevention,” he continues. We have diabetes screening and detection strategies, particularly in schools, but this remains insufficient given the needs within the population. Moreover, with the voice, we can diagnose certain diseases, it is an important source of information in the field of health”.
According to many scientists, the use of voice recordings might help identify other diseases, such as Covid.
1698130712
#seconds #detect #diabetes #Voice #technology #revolutionize #healthcare #screening
