Validation of a predictive model for coronary artery disease in patients with diabetes

BACKGROUND: No reliable model can currently be used for predicting coronary artery disease (CAD) occurrence in patients with diabetes. We developed and validated a model predicting the occurrence of CAD in these patients. METHODS: We retrospectively enrolled patients with diabetes at Henan Provincial People's Hospital between 1 January 2020 and 10 June 2020, and collected data including demographics, physical examination results, laboratory test results, and diagnostic information from their medical records. The training set included patients (n = 1152) enrolled before 15 May 2020, and the validation set included the remaining patients (n = 238). Univariate and multivariate logistic regression analyses were performed in the training set to develop a predictive model, which were visualized using a nomogram. The model's performance was assessed by area under the receiver-operating characteristic curve (AUC) and Brier scores for both data sets. RESULTS: Sex, diabetes duration, low-density lipoprotein, creatinine, high-density lipoprotein, hypertension, and heart rate were CAD predictors in diabetes patients. The model's AUC and Brier score were 0.753 [95% confidence interval (CI) 0.727–0.778] and 0.152, respectively, and 0.738 (95% CI 0.678–0.793) and 0.172, respectively, in the training and validation sets, respectively. CONCLUSIONS: Our model demonstrated favourable performance; thus, it can effectively predict CAD occurrence in diabetes patients.