International journal of general medicine
Development and Internal Validation of a Risk Prediction Model for Visual Impairment Progression in Elderly Patients with Diabetic Retinopathy: A Single-Center Retrospective Study
By Liang H, Hu P, Qin L and 1 others.
Abstract
Objective
To identify predictors of visual impairment progression in elderly patients with diabetic retinopathy (DR) and to establish a risk prediction model.
Methods
This single-center retrospective study included 443 elderly patients with DR treated between January 2022 and January 2025. According to the occurrence of visual impairment within 12 months after discharge, patients were divided into progression and non-progression groups. Multivariate logistic regression analysis was used to identify predictors, and a nomogram-based risk prediction model was constructed. Receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) were used to evaluate model performance.
Results
Multivariate Logistic regression analysis demonstrated that duration of diabetes, educational level, duration of DR, presence of other underlying diseases, DR stage, baseline glycated hemoglobin (HbA1c), and vascular endothelial growth factor A (VEGFA) level were independently associated with visual impairment progression and were included as predictors in the final model. ROC curve analysis showed that the area under the curve (AUC) was 0.932 in the training cohort and 0.925 in the validation cohort, indicating good discriminative performance. Bootstrap validation demonstrated good agreement between the calibration curve and the ideal curve, and no significant difference was observed between the predicted and observed values in either cohort (χ2=7.355 and 2.960, P=0.499 and 0.937, respectively), suggesting favorable calibration accuracy. DCA showed positive net benefit across threshold probabilities of 0.03-0.96 in both cohorts.