Machine Learning Prediction Model:
Diabetic Nephropathy
Poster presentation at the American Diabetes Association (ADA) 83rd Scientific Session
Presented at the 83rd ADA Scientific Sessions in San Diego, California, USA: June 2023
Shared with leading clinicians, researchers, and industry experts in diabetes care
Our research focuses on understanding and addressing the interconnected risks between cardiovascular disease (CVD) and type 1 diabetes (T1D), working toward a future where people with type 1 diabetes can live longer, healthier lives with reduced cardiovascular disease burden.
Risk Detection
CVD is the leading cause of death in people with T1D, yet often goes unrecognized and undertreated
Diabetic nephropathy (DN), a diabetes-related kidney disease, is a major but overlooked risk factor
Our ML models identify children with T1D at risk of DN
Conclusion
Using the T1D Exchange National Registry dataset, we developed predictive models with Random Forest and Logistic Regression and ScikitLearn
Both models performed well (AUC ~0.77–0.78), showing promise for enhanced early detection and intervention in at-risk patients
