Objective To construct a risk predicting model for cardio/cerebrovascular disease in middle-aged and elderly patients with type 2 diabetes based on China Health and Retirement Longitudinal Study(CHARLS). and conduct external model validation.

Methods Retrospective cohort study was conducted in middle-aged and elderly patients with type 2 diabetes from CHARLS in 2011 (training group) and middle-aged and elderly patients with type 2 diabetes from CHARLS in 2015 (validation group). The onset of cardio/cerebrovascular disease was the outcome event. Software R 4.2.3 was used for data analysis, Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis was used to screen the predictors, and Nomogram was used to construct the risk prediction model, receiver operating characteristic curve (ROC) and calibration curve were used for model validation.

Results A total of 507 study participats were included in the training group, in whom 198 suffered from cardio/cerebrovascular disease later. A total of 368 study participants were included in the validation group, in which 74 suffered from cardio/cerebrovascular disease later. LASSO regression analysis indicated that being women, excessive waist circumference, longer duration of type 2 diabetes, sleep duration, and hypertension were the risk factors for cardio/cerebrovascular disease in patients with type 2 diabetes. Based on the above five factors, a risk predicting model was constructed and presented as the Nomogram. The area under ROC curve in the training group and in the validation group were 0.689 95% confidence interval (CI): 0.641−0.737, 0.779 (95%CI: 0.717−0.841). The calibration curves indicated that the predicted value was close to the actual one, calibration degree was well fitted.

Conclusion The risk predicting model in this study has a good predictive value, which can be used to guide community medical workers for precise interventions.