1266-P: Validation of a Predictive Model of Diabetes in People with Overweight and Obesity

Abstract

Obesity is the most common chronic disease in the U.S. A substantial subset of people with obesity are at high risk for type 2 diabetes mellitus (T2DM) and many of their risk factors for T2DM are modifiable. It is important to identify this at-risk subset to appropriately direct treatment and resources for earlier intervention and the prevention of T2DM. The original model was developed using EMR data of 338,653 adults < 65 years old with BMI ≥ 25 kg/m2 and without baseline T2DM between 2000 and 2019. Variables were selected from a literature-based pool of risk factors for T2DM using LASSO penalized Cox regression. A Cox model (adjusted for the competing risk of death) was developed using an 80% training dataset with selected variables and internally validated on a held-out 20% dataset of EMR data. This model was then externally validated on a dataset of 103,419 participants of the Nurses’ Health Study (NHS) and Health Professionals Follow-Up Study (HPFS). Over a mean of 7.0 years of follow-up, 30,100 (8.9%) patients in the EMR dataset developed T2DM (1.3% annual incidence). The predictive model included BMI (HR 1.47 for 5 kg/m2; 95% CI 1.46 - 1.49), age, history of hypertension, systolic blood pressure and family history of diabetes. The model achieved a Harrell’s C-statistic of 0.76 on the held-out EMR dataset and the calibration slope for 10-year estimates was 1.09. In the NHS-HPFS external validation dataset, 12,402 (12.0%) individuals developed T2DM over a mean 14.2 years of follow-up (0.85% annual incidence). On this dataset, the model achieved Harrell’s C-statistic of 0.72 and the calibration slope for 10-year estimates of 0.78. It is feasible to use predictive modeling to identify people with overweight and obesity at high risk of T2DM. This approach could be utilized to guide population management and clinical treatment decisions. Disclosure A.Turchin: Consultant; Proteomics International, Covance, Research Support; AstraZeneca, Edwards Lifesciences Corporation, Eli Lilly and Company, Novo Nordisk, Stock/Shareholder; Brio Systems. M.Shubina: None. F.J.Morrison: None. N.Ahmad: Employee; Eli Lilly and Company. L.M.Neff: Employee; Eli Lilly and Company, Research Support; Aegerion Pharmaceuticals, Stock/Shareholder; Eli Lilly and Company. H.Kan: Employee; Eli Lilly and Company, Stock/Shareholder; Eli Lilly and Company. Funding Eli Lilly and Company