Development and internal validation of a nomogram to predict temporary acute agitated delirium after surgery for chronic subdural hematoma in elderly patients: an analysis of the clinical database

BackgroundThis study aimed to develop a nomogram for predicting temporary acute agitated delirium after surgery in patients with chronic subdural hematoma (CSH) without neurological compromise and hospitalized in the neurosurgery.MethodsWe included 289 patients with chronic subdural hematoma (CSH) from the medical information system of Yuebei People’s Hospital of Shaoguan City, Guangdong Province, and collected 16 clinical indicators within 24 h of admission. We used the least absolute shrinkage and selection operator (LASSO) regression to identify risk factors. We established a multivariate logistic regression model and constructed a nomogram. We performed internal validation by 1,000 bootstrap samples; we plotted a receiver operating curve (ROC) and calculated the area under the curve (AUC), sensitivity, and specificity. We also evaluated the calibration of our model by the calibration curve and the Hosmer–Lemeshow goodness-of-fit test (HL test). We performed a decision curve analysis (DCA) and a clinical impact curve (CIC) to assess the net clinical benefit of our model.ResultsThe nomogram included alcoholism history, hepatic insufficiency, verbal rating scale for postoperative pain (VRS), pre-hospital modified Rankin Scale (mRS), and preoperative hematoma thickness as predictors. Our model showed satisfactory diagnostic performance with an AUC value of 0.8474 in the validation set. The calibration curve and the HL test showed good agreement between predicted and observed outcomes (p = 0.9288). The DCA and CIC showed that our model had a high predictive ability for the occurrence of postoperative delirium in patients with CSDH.ConclusionWe identified alcoholism, liver dysfunction, pre-hospital mRS, preoperative hematoma thickness, and postoperative VRS pain as predictors of postoperative delirium in chronic subdural hematoma patients. We developed and validated a multivariate logistic regression model and a nomogram.