NHS begins world’s first trial of AI tool to identify risk of type 2 diabetes
The NHS in England is launching a world-first trial of a ‘game-changing’ artificial intelligence tool that can identify patients at risk of type 2 diabetes more than a decade before they develop the condition.
More than 500 million people worldwide have this type 2 diabetesand finding new ways to detect people at risk before they develop the condition is a key global health priority. Estimates suggest that by 2050, 1 billion people will have type 2 diabetes.
The condition is a leading cause of blindness, kidney failure, heart attacks, strokes and lower limb amputation. It is often related to being overweight or inactive, or to a family history of type 2 diabetes, although not all diagnosed patients fall into these categories.
Now, doctors and scientists have developed a transformative AI tool that can predict people at risk for the condition as much as 13 years before it starts to develop.
The technology analyzes electrocardiogram (ECG) readings during routine heart scans. It can detect subtle changes that are too small to be noticed by the human eye and can raise early alarms about a patient on the way to developing type 2 diabetes.
It could allow for early interventions and potentially help people not develop the condition at all, for example by making changes to their diet and lifestyle.
The NHS will begin testing the tool at Imperial College Healthcare NHS Trust and Chelsea and Westminster Hospital NHS Foundation Trust in 2025, becoming the first healthcare system in the world to do so, the Guardian has learned.
Those involved in developing the technology, called AI-ECG risk assessment for diabetes mellitus (Aire-DM), hope it can be rolled out across healthcare systems in England and other countries within the coming years.
“AI has enormous potential to transform healthcare, which could lead to substantial improvements in health,” said Dr Libor Pastika, a clinical research trainee at Imperial. “By using AI to unlock insights hidden in ECG data, Aire-DM could be revolutionary in early identification of future risk of type 2 diabetes.
“By offering a low-cost, accessible, non-invasive way to early predict the risk of type 2 diabetes, Aire-DM could open a new opportunity for more targeted, preventative care, helping people avoid the condition and its associated complications. to avoid.”
A team led by Dr Fu Siong Ng and Dr Arunashis Sau from Imperial developed the tool using 1.2 million ECGs from hospital records. They then used data from the UK Biobank, which holds the genetic data and medical records of more than 500,000 participants, to validate the tool’s ability to detect subtle changes in ECGs.
The tool maps small ECG patterns that are more common in those who will develop type 2 diabetes in the future – and then looks for those same patterns in new ECGs.
Telltale signs include variations in the way the heart’s electrical signal propagates, such as small changes in the timing, shape, or patterns of certain electrical waves.
The tool can also identify longer electrical activation times or differences in the way the heart’s electrical signals are reset. Although these changes may seem small, they reflect the early effects of diabetes on the structure and function of the heart, long before symptoms appear.
Tests have already shown that the tool accurately predicts risk about 70% of the time in people of different ages, genders, ethnicities and socio-economic backgrounds.
When the AI predictions were combined with genetic and clinical information, such as age and blood pressure, accuracy improved further, creating an even clearer picture of risk, the researchers said.
Prof Bryan Williams, the chief scientific and medical officer of the British Heart Foundation, a charity that helped fund the development of the tool, said: “This exciting research uses powerful artificial intelligence to analyze ECGs, revealing how AI can notice things that normally cannot be noticed. are observed in routinely collected health data.
“This kind of insight could be a game changer in predicting the future risk of developing type 2 diabetes, years before the condition begins.
“Type 2 diabetes is a rapidly growing health problem that increases the risk of developing heart disease. However, with the right support, it is possible for people to reduce their risk of developing the condition.
“We look forward to seeing how this technology can be integrated into clinical practice, providing the opportunity for early intervention to help reduce risk and even prevent type 2 diabetes and its associated complications.”