A machine learning model can help forecast patients’ risk for 18 obesity-related complications and guide treatment prioritization, according to a study published April 30 in Nature Medicine.
Researchers analyzed data from about 200,000 adults with overweight or obesity to develop the model, called OBSCORE, which uses 20 clinical variables to predict 10-year risk across conditions including cardiovascular disease, Type 2 diabetes and chronic kidney disease.
The model outperformed traditional approaches such as body mass index alone, more accurately identifying patients at elevated risk. Investigators found some individuals classified as overweight — not obese — still fell into high-risk categories, underscoring limitations of BMI-based screening.
OBSCORE stratified patients into risk groups with wide variation in outcomes. For example, it predicted 10-year cardiovascular mortality ranged from 0.1% in the lowest-risk group to 5.7% in the highest-risk group.
The researchers said the framework could support more targeted use of obesity treatments and improve resource allocation, particularly as demand rises for GLP-1 therapies and other weight-loss interventions.
