AI voice technology to target diabetes risk

Type 2 diabetes (T2D) is a global epidemic, affecting 496 million people worldwide and 1.3 million Australians. Alarmingly, it is often diagnosed 5 to 7 years after the first symptoms appear, by which time complications have already set in, contributing to excess mortality.

In a bid to shorten this diagnostic delay, a pioneering research collaboration between RMIT University and digital health company DDM Health is turning its focus to artificial intelligence (AI) voice analysis technology. Facilitated by the Digital Health Cooperative Research Centre (DHCRC), the project aims to develop an innovative AI model that analyses voice data to assess the risk of type 2 diabetes.

Reflecting on the broader impact of diabetes, Professor Barbora de Courten, project lead at RMIT University, noted “Type 2 diabetes is a global epidemic, affecting around 537 million people worldwide, and global spending on diabetes is USD$966 billion annually, which equates to 11.5% of total health expenditure.”

He continued, “This project offers a new frontier in non-invasive disease risk assessment and will pioneer the use of advanced machine learning algorithms to identify diabetes risk through voice patterns, pushing the boundaries of healthcare technology.”

Supporting the rationale behind this innovative approach, research from the Luxembourg Institute of Health presented at last year’s European Association for the Study of Diabetes (EASD) meeting demonstrated the promising potential of voice analysis in detecting undiagnosed type 2 diabetes cases. By employing an average of 25 seconds of recorded voice along with basic health data such as age, sex, body mass index (BMI), and hypertension status, the AI model achieved an accuracy of 66% in women and 71% in men when determining the presence of T2D.

Arjun Panesar, founding CEO of DDM Health, stated “This project will showcase the power of AI to transform healthcare by enabling early detection of diabetes through non-invasive voice analysis. We’re developing a cutting-edge tool to predict diabetes risk by leveraging advanced AI and machine learning that will integrate seamlessly into global health applications.”

Traditional diagnostic methods for diabetes typically require a medical consultation and at least one blood test, if not a series of tests. The hope is that this innovative vocal biomarker approach will offer a cost-effective, non-invasive tool that could prompt individuals to seek timely medical advice, ultimately enabling earlier detection and improved patient outcomes.

Highlighting the far-reaching benefits of such technology, DHCRC CEO Annette Schmiede, stated “The development and commercialisation of a voice-based diabetes risk prediction tool has the potential to provide a novel type 2 diabetes detection system for wide adoption within Australia, including remote, underserved communities.”

She further added, “As well as improved patient outcomes, this project has the potential to significantly reduce healthcare costs associated with type 2 diabetes and more broadly further strengthen and showcase the capability of the digital health industry in Australia.”

In reimagining healthcare across the entire patient journey, Health Industry Hub^TM is the only one-stop-hub uniting the diversity of the Pharma, MedTech, Diagnostics & Biotech sectors to inspire meaningful change.

The Health Industry Hub^TM content is copyright protected. Access is available under individual user licenses. Please click here to subscribe and visit T&Cs here.