AI-Powered Voice Analysis Shows Promise for Rapid Type 2 Diabetes Screening

A novel tool leveraging artificial intelligence to analyze speech patterns could diagnose type 2 diabetes in mere seconds, according to a recent study. Researchers believe this technology offers a new avenue for diabetes testing, with voice recordings potentially captured via phone calls or mobile applications.

Millions in the UK are living with diabetes, and a significant portion remain undiagnosed due to the condition's often slow-developing symptoms, such as fatigue and increased thirst. Limited access to routine check-ups and standard blood tests can also hinder early detection.

The technology, developed by researchers from the tech company thymia and RMIT University in Melbourne, Australia, identifies changes in speech that have been associated with type 2 diabetes. These changes can include vocal strain, hoarseness, and difficulties with breath control. High blood sugar levels, common in diabetes, can affect the vagus nerve, which influences the voice box muscles, leading to a rough or scratchy voice. Additionally, increased stomach acid reflux in people with diabetes can irritate vocal cords, and reduced lung function can impact speech clarity.

To train the AI model, researchers utilized over 63,000 voice samples from more than 21,000 individuals across the UK and US. The model was then tested using 20-second recordings of participants reading aloud. In an analysis of 7,319 UK participants, the AI tool assigned a higher risk score to individuals who reported having type 2 diabetes 80% of the time. While the model performed consistently across various ages and genders, its accuracy was lower for Black patients, a disparity researchers attribute to the limited number of Black participants in the dataset.

A subsequent analysis focused on a subgroup of 801 individuals who had undergone diabetes blood tests within three months of providing their speech recording. The AI tool correctly identified individuals with a higher risk of diabetes in 75% of these cases.

Traditionally, type 2 diabetes is diagnosed through blood tests measuring average blood sugar levels over two to three months. These tests are typically offered to individuals experiencing symptoms or those aged between 40 and 74 during routine health checks.

Giedre Cepukaityte, a research scientist at thymia, highlighted the tool's potential to transform diabetes screening. "This is the largest real-world study of speech-based screening for type 2 diabetes to date which also checks the model's predictions against blood test results as well as against what people reported about their own diagnosis," she stated. "A speech sample can be taken over the phone or through an app, so we can reach far more of the people who need a blood test than current pathways do, particularly those who never get to a health check."

Cepukaityte emphasized that the AI tool is not a substitute for blood tests and should not deter individuals from seeking medical attention if they suspect they have diabetes. "Our next step is to test the model in clinical settings and to understand how well it works for every group of people, because a screening tool has to work for everyone," she added.