A groundbreaking study presented at the American Diabetes Association’s 2026 Scientific Sessions in New Orleans has unveiled a machine learning model capable of predicting the long-term risk of developing type 2 diabetes. This model, based on electronic health records, successfully identifies patients at the highest risk up to a decade in advance, potentially transforming early intervention strategies.

Study Details and Findings

The study analyzed data from over three million adults aged 18 to 70, who received care from Kaiser Permanente Northern California between 2012 and 2024. The median age of participants was 39, with a slight majority of 55% being female. Researchers employed a hazard-based super learning approach, integrating multiple survival-analysis models to estimate each patient’s diabetes risk over one, three, and ten years.

Key factors such as age, weight, blood glucose levels, medical history, and medications were considered, alongside publicly available data like access to healthy food and walkable areas. During a median follow-up of 5.4 years, the study reported a type 2 diabetes incidence of 10.7 per 1,000 person-years. The model demonstrated impressive accuracy, with an area under the curve of 0.886 for the training model and 0.883 for the validation model.

Implications for Healthcare

Lead author Luis A. Rodriguez, PhD, MPH, RD, emphasized the model’s potential to enhance current diabetes prevention efforts. By identifying high-risk individuals often missed by traditional screening methods, healthcare providers can focus on those who would benefit most from early prevention and treatment. The model’s high sensitivity and specificity over a ten-year follow-up period underscore its effectiveness.

The researchers plan to test the model in clinical settings to evaluate its impact on increasing engagement in diabetes prevention programs and reducing the incidence of the disease. This proactive approach could significantly shift how healthcare systems address the growing diabetes epidemic in the United States, where over 60% of adults have risk factors for type 2 diabetes.

About the ADA Scientific Sessions

The ADA’s 2026 Scientific Sessions, held in New Orleans, is the world’s largest scientific meeting focused on diabetes research, prevention, and care. The event attracts thousands of leading physicians, scientists, and healthcare professionals globally, offering a platform for unveiling cutting-edge research and treatment advancements.

Original reporting: KTBS 3 (Shreveport) — read the source article.