Researchers at the Technical University of Denmark have developed a free online self-assessment tool, MEDWACS, which can identify individuals at risk of prediabetes or type 2 diabetes using simple measurements, potentially years before symptoms appear.
A new artificial intelligence tool developed by researchers at the Technical University of Denmark may help people spot prediabetes or undiagnosed type 2 diabetes from home, potentially years before symptoms drive them to a clinic. The work, published in the Journal of Clinical Epidemiology, centres on MEDWACS, a free online self-assessment designed to flag risk rather than deliver a diagnosis.
According to the study, the model was trained on more than 30 years of data from a large US health survey and then tested on separate groups in the United States and South Korea. The researchers say the system uses seven straightforward measurements and questions, allowing users to assess their risk with items such as a bathroom scale, tape measure and blood pressure monitor. The unusually specific thigh-length measure was included because, the researchers argue, it may reflect early-life nutrition and muscle mass, both of which can shape how well the body handles glucose.
The team says MEDWACS performed as well as or better than established screening methods, and that its results held up across different populations. That matters because many existing screening approaches depend on a clinic visit or laboratory testing, which can leave people unaware of risk until the disease has progressed. The researchers say the tool is intended to prompt follow-up blood tests where needed, not to replace a clinician’s judgement.
The broader appeal lies in accessibility. A home-based check could lower the threshold for screening among people who feel well and might otherwise never seek testing, yet still carry significant risk. The researchers believe the model could also work well in Denmark, though they say local validation studies are still needed before any widespread use. Similar efforts have explored AI-based diabetes screening using non-laboratory health data and even electrocardiograms, reflecting growing interest in low-cost ways to identify the disease earlier.
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
