Objective To construct a nomogram risk prediction model for postoperative delirium in patients undergoing video-assisted thoracic surgery (VATS) based on anesthesia factors. Methods The clinical data of 172 patients (modeling set) who underwent VATS were retrospectively analyzed. According to whether delirium occurred after surgery, patients were divided into delirium group (n=43) and non-delirium group (n=129). The independent influencing factors of postoperative delirium were analyzed, and the prediction model was constructed based on regression analysis method. The corresponding nomogram risk prediction model is drawn by using R language, and its prediction efficiency is internally verified by using receiver operating characteristic (ROC) curve and calibration curve. In addition, 74 VATS patients were selected as the verification set, and ROC curve and calibration curve were drawn with verification set data to conduct external verification of the prediction efficiency of the model. Results The results of multivariate Logistic regression analysis showed that age, intraoperative blood loss, anesthesia method, midazolam's usage, compound use of anesthesia, anesthesia duration, and stroke history were all independent influencing factors for postoperative delirium in patients undergoing VATS (all P<0.05). The area under the curve (AUC) of the nomogram risk prediction model based on the above 7 independent influencing factors was 0.905, and the sensitivity and specificity corresponding to the optimal cut-off value of 0.286 were 0.860 and 0.853, respectively, which indicating the model has good distinguish performance. The analysis results of the calibration curve show that the average absolute error is 0.045, and the calibration curve fits the ideal curve. The model has good calibration performance and is reliable and stable. The results of ROC curve and calibration curve of verification set were good, indicating that the model has good external prediction efficiency. Conclusion The occurrence of postoperative delirium in patients undergoing VATS is affected by age, anesthesia method, midazolam's usage, anesthesia duration and other factors. The nomogram risk prediction model based on the above anesthesia factors and clinical data factors has good predictive efficacy, which can provide reference for the prevention of delirium in patients.

无