Cardiovascular disease (CVD) is the leading cause of global mortality in type 1 Diabetes (T1D). This risk remains largely under-managed, leading to a future significant CVD burden. Diabetic nephropathy (DN) is considered a risk factor for CVD and overall mortality in patients with T1D. Identifying risk factors that contribute to the development and progression of DN is therefore important in T1D. Early identification of DN may support earlier intervention for patients with T1D and in turn, reduce future risk for CVD. Using the T1D Exchange Registry database of individuals with T1D in the USA, we sought to develop a machine-learning algorithm to predict patients with DN. With performance metrics of F1-score = 0.67, AUC = 0.78, in our Random Forest model, and F1-score= 0.66, AUC = 0.77 in our Logistics Regression model demonstrates the potential to assist in mitigating DN development and progression.
S. Sripada: None. S. Sripada: None. S. Belapurkar: Research Support; Dexcom, Inc., Tandem Diabetes Care, Inc. Other Relationship; T1D Exchange. Research Support; Medtronic.
