Obesity is far more complex than weight alone. New research is helping predict health risks and treatment response for more personalised care
Health
Obesity constitutes one of the most pressing public health challenges of the 21st century. More than one billion people worldwide are currently living with obesity, a chronic and complex disease linked to over 230 health complications including cardiovascular disease, diabetes and cancer.
Despite its growing prevalence, healthcare professionals still struggle to predict which individuals will develop obesity-related complications and who will benefit most from specific treatments. Addressing this knowledge gap was the central mission of the SOPHIA(opens in new window) project, a public-private partnership funded by the Innovative Medicines Initiative (IMI) in collaboration with the European Commission. SOPHIA brings together more than 35 partners from industry, academia, and patient and public organisations from EU and around the world.
Moving beyond a one-size-fits-all approach
For decades, obesity management has relied heavily on body mass index (BMI), a simple measure based on weight and height. While useful at a population level, BMI alone cannot capture the biological complexity of obesity or explain why people with similar body weights often experience different health outcomes.
“Earlier approaches to obesity management viewed all people with obesity as being the same, failing to consider the varying pathogenesis, the individual’s risk for complications and their treatment response,” explains project coordinator Carel le Roux.
SOPHIA sought to address this by identifying distinct obesity phenotypes and improving patient stratification. The project demonstrated that obesity should be viewed through a broader lens that incorporates body composition, physical and physiological biomarkers, genetics and disease risk. This approach will help understand the substantial variation between patients and move towards treatment strategies focused on reducing health risks rather than weight alone.
Harnessing data to predict outcomes
Researchers created a federated database that allowed secure analysis across multiple studies without transferring sensitive patient information. The database brings together data from more than 18 studies and over 90 000 patients.
Using machine learning and large-scale clinical data, researchers focused particularly on predicting the risk of obesity-related complications such as cardiovascular disease and chronic kidney disease. These models are helping to identify which patients are most likely to develop complications and which treatment strategies may be most effective for different patient groups.
Shifting the focus to risk reduction
Among the project’s most significant scientific achievements was the development of a new perspective on obesity treatment. SOPHIA researchers found that absolute values of BMI(opens in new window) and waist-to-height ratio following weight change are more strongly associated with future health risks than the percentage of weight lost.
Based on extensive real-world data, the team proposed that achieving a BMI below 27 kg/m² or a waist-to-height ratio below 0.53 is associated with lower risk of developing conditions such as type 2 diabetes, hypertension, osteoarthritis and cardiovascular disease.
Giving patients a voice
SOPHIA also explored the experiences of people living with obesity. Researchers collected qualitative data from patients and healthcare professionals to better understand the motivations and concerns surrounding obesity management(opens in new window).
One notable outcome was the development of a photovoice educational resource combining photographs and personal testimonies.
“By capturing real-life experiences, we highlighted the emotional and social dimensions of obesity and contributed to efforts to reduce stigma and improve patient engagement,” emphasises project industry lead Alix Feldman.
Looking ahead, SOPHIA will further validate the proposed treatment targets and expand the federated data infrastructure for future studies. The prediction models will continue to be tested in clinical practice, bringing healthcare systems closer to delivering the right treatment for the right patient.
Keywords
Discover other articles in the same domain of application
Pioneering pan-European screening for early detection of type 1 diabetes
4 June 2026
Transforming diabetes care through stem cell-based therapies
3 July 2026
Bioartificial pancreas offers a potential cure for type 1 diabetes
22 April 2026
Making the case for the routine screening of liver fibrosis
21 May 2026
