To explore the construction and validation of a prediction model for the prognosis of traumatic brain injury (TBI) based on Rotterdam CT score and coagulation function indicators.

A total of 108 TBI patients admitted to Neurosurgery Department of Lingbi Hospital of the Second People's Hospital of Anhui Province from January 2020 to January 2022 were selected as the study objects. CT image examination and Rotterdam CT score were performed on all patients within 24 h of admission, and coagulation function index was also detected. Based on the GOS score at 6 months after treatment, patients were divided into a good prognosis group (4-5 score) and a poor prognosis group (1-3 score). Binary Logistic regression analysis was used to screen the independent factors influencing the prognosis of TBI patients, and a prediction model of TBI prognosis was constructed based on Rotterdam CT score and coagulation function index [prothrombin time (PT), activated partial thromboplastin time (APTT), D-dimer (D-D)]. Receiver operating characteristic (ROC) curve was established to evaluate the predictive efficiency of the model. At the same time, the clinical data of 46 TBI patients admitted to our hospital from January to December 2019 were collected as a validation set to verify the accuracy of the prediction model.

Among 108 patients, there were 73 patients in the good prognosis group and 35 patients in the poor prognosis group. The severity of the disease, combined hypoxemia, Rotterdam CT score, and coagulation function indicators (PT, APTT, D-D) in the poor prognosis group were higher than those in the good prognosis group, with statistical significance (P<0.05). Logistic regression analysis showed that disease condition, Rotterdam CT score, PT, and D-D were independent influencing factors for poor prognosis in TBI patients (P<0.05). The predictive model for poor prognosis of TBI is Y=3.078×Rotterdam score CT score-0.847×PT+1.211×D-1.672×severity severity of the disease-3.195. The prediction model predicted poor prognosis with 0.886 sensitivity, 0.836 specificity, 0.722 Jorden index, 0.262 cut-off value and 0.937 area under curve. The Hanley&McNeil method was used to compare the predictive model and the validation curve, and it was found that there was no statistically significant difference in the AUC comparison of the ROC curves between the two models (P<0.05).

For patients with TBI, the prediction model based on Rotterdam CT score and coagulation function has good accuracy and is worthy of promotion.