A new artificial intelligence tool, developed by researchers at the University of Cambridge and Queen Mary University of London, can predict an individual’s risk of developing obesity-related diseases, offering a more personalised approach to deciding who should receive scarce weight-loss treatments on the NHS.
The tool, named Obscore, is designed to address a growing problem: around two-thirds of adults in England are overweight or obese, a crisis that costs the health service an estimated £11.4 billion each year. At present, access to medications such as tirzepatide (Mounjaro) and semaglutide (Ozempic) is strictly rationed and largely determined by body mass index (BMI) – typically requiring a BMI of 35 or above, or 30 with at least one weight-related condition, with even higher thresholds for some drugs. The researchers argue that this blunt measure overlooks many people who face significant health risks.
Writing in the journal Nature Medicine, the team describes how they applied a form of artificial intelligence known as interpretable machine learning to data from nearly 200,000 participants of the UK Biobank project, a large-scale prospective study that has collected biological samples and health data from half a million people aged 40–69 since 2006. All of the participants used to develop the tool had a BMI of 27 or higher, placing them in the overweight or obese category. The AI analysed a wealth of information per participant – the Biobank holds more than 10,000 variables each – to identify 20 health, lifestyle and demographic features that could predict the 10-year risk of developing 18 different obesity-related complications, ranging from gout to stroke.
These features include age, sex, total cholesterol and creatinine levels. The tool then places each person into one of five equal-sized risk categories, from low to high, for each specific condition. Crucially, the analysis showed that people of the same age, sex and BMI could have very different risk profiles. For conditions such as type 2 diabetes, a considerable proportion of those in the highest risk category were classified as overweight rather than obese. “These constitute a population of individuals who may be overlooked if we only look at BMI and not other risk factors,” said Kamil Demircan, a co-author of the study from Queen Mary University of London. Professor Claudia Langenberg, director of the university’s Precision Healthcare University Research Institute, is also a lead author.
The tool was validated using data from the UK Biobank itself as well as two independent health studies. The researchers also applied a version of Obscore to data from a randomised controlled trial of the weight-loss drug tirzepatide, confirming that those predicted to be at highest risk for obesity-related conditions experienced a similar degree of weight loss to other participants. Tirzepatide has been shown in trials to produce an average weight loss of up to 22.9% and to reduce the risk of progressing from pre-diabetes to type 2 diabetes by 94%, as well as lowering the risk of major cardiovascular events in certain patients.
Professor Nick Wareham, of the University of Cambridge, a co-author of the study, said the purpose of the tool was not to extend the use of particular therapies but to enable more rational resource allocation. “It’s about developing and validating a score that can help with more rational resource allocation,” he said. “So, can we prescribe therapy to those people who are most likely to need it and most likely to benefit from it – which is what we should do within the NHS.”
However, independent experts have urged caution about the tool’s readiness for everyday clinical use. Professor Naveed Sattar, a professor of cardiometabolic medicine at the University of Glasgow who was not involved in the work, said many of the obesity-related conditions covered by Obscore are closely interrelated, and robust, simpler risk scores already exist for some of them. He noted that several of the metrics used in the study – such as certain blood biomarkers – are not routinely available within the NHS. “Overall, this work represents a thoughtful attempt to move towards more holistic risk prediction across multiple obesity-related conditions,” Sattar said. “But substantial further development and validation will be required before such an approach can be translated into routine clinical practice.”