New AI-Powered Tool Predicts Obesity-Related Complications Beyond BMI

Body mass index (BMI) remains the standard metric for assessing obesity-related health risks, despite its well-documented limitations. Researchers at Queen Mary University of London have developed a more comprehensive tool to predict 18 serious complications linked to obesity. Unlike traditional methods, this model incorporates BMI alongside family medical history, dietary habits, existing illnesses, and socioeconomic factors extracted from medical records.

The new tool addresses a critical gap in obesity treatment. Currently, medications like GLP-1 agonists—originally designed for type 2 diabetes—are prescribed based on BMI alone or in combination with other conditions. However, these drugs have demonstrated benefits beyond weight loss, including reducing risks for cardiovascular disease, kidney disease, liver disease, sleep apnea, and osteoarthritis. Despite their potential, determining which patients are optimal candidates for these lifelong, costly treatments has remained uncertain.

An integrated model for 18 obesity-related complications

“We really wanted to have an integrated model that enables us to look at not one, but 18 different obesity-relevant complications,” said Claudia Langenberg, co-author of the study published in Nature Medicine on August 1, 2024. Langenberg serves as director and professor of medicine and population health at the Precision Healthcare University Research Institute of Queen Mary University of London. She shared these insights during a media briefing on July 30, 2024.

The research team’s model leverages advanced machine learning techniques to analyze vast datasets, identifying patterns that predict individual risk profiles more accurately than BMI alone. By considering a broader range of factors, the tool aims to personalize obesity treatment strategies, ensuring patients receive the most appropriate interventions.

Potential impact on obesity treatment

Obesity is a global health crisis, contributing to numerous chronic conditions and placing significant strain on healthcare systems. Traditional risk assessment tools often fail to capture the full spectrum of complications associated with obesity, leading to suboptimal treatment decisions. The new AI-driven model could transform how clinicians evaluate and manage obesity by providing a more nuanced understanding of patient risk.

For example, patients with a high risk of cardiovascular complications might benefit from early intervention with GLP-1 medications, even if their BMI falls within a borderline range. Conversely, individuals with lower risk profiles could explore alternative treatments, reducing unnecessary exposure to costly drugs and their potential side effects.

The study highlights the importance of moving beyond BMI as the sole determinant of obesity-related health risks. By integrating multiple data points, the model offers a more holistic approach to patient care, aligning treatment strategies with individual needs.