New test could screen for type 2 diabetes in seconds using just speech, study shows
Published: | Updated:
A tool that uses artificial intelligence to scan for changes in speech could be used to diagnose type 2 diabetes in just seconds, a new study has found.
Researchers say the technology 'opens a new route for diabetes testing' - with voice recordings able to be collected over the phone or by using an app.
More than six million people in the UK are currently living with diabetes, according to Diabetes UK.
Just 4.7 million have an official diagnosis, however - while another third of diabetes patients are living with the condition without knowing it.
The slow onset of diabetes symptoms, such as fatigue and excessive thirst, can contribute to delayed diagnoses, research shows.
But a lack of access to routine check-ups and testing can also mean fewer at-risk patients are offered standard blood tests that identify the condition.
Now, a new tool promises to change the way type 2 diabetes is diagnosed.
Developed by researchers at tech company thymia and RMIT University in Melbourne, Australia, the technology uses artificial intelligence to detect changes in speech which have been linked to type 2 diabetes.
A tool that uses artificial intelligence to scan for changes in speech could be used to diagnose type 2 diabetes in just seconds, a new study has found
Vocal strain, increased hoarseness and an inability to control breath can all signal the condition, research shows.
A rough or scratchy voice quality is common in people with poor blood sugar control, as high blood sugar can harm the vagus nerve, which controls the muscles in the voice box.
People with diabetes also have a higher rate of stomach acid reflux, which irritates and inflames the vocal cords - causing hoarseness - while reduced lung function lowers the airflow needed for clear speech.
To be able to pick up on subtle changes in speech, the new tool was trained using more than 63,000 voice samples from more than 21,000 people in the UK and US.
Researchers then tested the model using 20-second recordings of people reading Aesop's fables out loud.
The study included 7,319 people in the UK and found that the speech model gave a higher risk score to those who reported having type 2 diabetes 80 per cent of the time.
While performing well across different ages and genders, the tool's performance was lower when analysing the speech of black patients - which researchers suggest was likely due to the low number of black patients taking part.
The second analysis included a sub-group of 801 people who took diabetes blood tests at home within three months of the speech recording.
The AI tool gave these people higher risk scores 75 per cent of the time.
Diabetes is usually diagnosed through a blood test that determines average blood sugar levels over the previous two to three months.
The test is available to those having symptoms, or during routine health checks which are offered to people aged between 40 and 74.
Giedre Cepukaityte, a research scientist at thymia who will be presenting the findings at the European Association for the Study of Diabetes (EASD) in Milan said the tool 'has the potential to change what screening looks like'.
'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 said.
'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.
'Our model opens a new route to screening for diabetes. It is not a replacement for a blood test, and it should never stop anyone who thinks they need one from getting one.
'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.'
KioskNews shows a cleaned-up reading view extracted from the publisher’s page — the original always lives on their site, not ours.