A person talking for just 20 seconds could show whether or not they have type 2 diabetes, research in nearly 30,000 individuals suggests.
AI-based analysis of short voice recordings could transform screening for the condition, which currently involves blood tests and primary care consultations.
Through the use of digital and telephone platforms, it would enable type 2 diabetes to be diagnosed earlier and thereby reduce the risk of complications.
The research was presented at the European Association for the Study of Diabetes 62nd annual meeting in Milan, Italy.
The researchers trained an AI model to detect these changes using 63,283 voice samples from 21,129 people in the U.K. and U.S. who reported whether or not they had been diagnosed with diabetes.
The model was then validated using 20-second recordings of people reading Aesop's fables in 7319 adults from the U.K., of whom two-thirds were women, nearly half were 40 years or older, and 217 of whom reported having type 2 diabetes.
Results showed that the model gave a higher risk score to people who reported having type 2 diabetes than others 80% of the time, which the researchers deemed clinically useful.
【MORE】