Better diagnostics on the horizon as AI reshapes early diabetes detection
New algorithms analysing continuous glucose monitor data and electrocardiograms aim to flag metabolic risks years before conventional diagnosis
Current methods for detecting diabetes, such as blood glucose and HbA1c tests, often fail to identify the condition early in specific populations. To address this, researchers are deploying artificial intelligence to analyse existing health data and uncover hidden metabolic patterns that traditional metrics miss.
At Stanford University, an AI algorithm analysing continuous glucose monitor data has demonstrated the ability to identify distinct metabolic patterns indicative of Type 2 diabetes with roughly 90 per cent accuracy. This tool potentially flags risk years before a conventional diagnosis, allowing for preventative measures such as dietary adjustments before significant damage occurs.
Concurrently, scientists at Imperial College London have developed an AI system that predicts Type 2 diabetes risk by analysing electrocardiograms with approximately 70 per cent accuracy. The model, trained on millions of heart tracings, offers a scalable screening method using data already available in hospital records, meaning no new invasive tests are required to assess future risk.
For Type 1 diabetes, a new risk calculator combining age, family history, and autoantibody status aims to enable practical, large-scale early screening. Developed by researchers at the University of Exeter, this tool integrates multiple factors to estimate risk without the cost and time associated with lengthy clinical testing, facilitating earlier intervention before blood sugar levels rise.
Continuous glucose monitors are also becoming cheaper and more accessible in the US, with many now available over the counter. This increased availability paves the way for potential integration into routine preventative healthcare, supporting the goal of keeping people healthy rather than attempting to fix complications later.
While these innovations represent significant advances, experts caution that they are not replacements for gold-standard diagnostic tests like HbA1c. Instead, they serve as complementary tools to flag at-risk patients automatically during routine care, ensuring that those who might otherwise be missed are identified and supported.
