Study on the Prediction Model of Wound Infection After Spinal Fusion and Internal Fixation Based on Logistic Regression.

With the development of the times, spinal problems are not only one of the diseases that older people pay close attention to, but also gradually spread among teenagers. Therefore, it is very important to predict the possibility of wound infection in patients after spinal fusion and internal fixation. The method is to statistically analyze the clinical data of patients with clinical spinal disease, and to propose individualized treatment and recovery plan for each patient's pathological characteristics and postoperative recovery, so as to realize humanized service and minimize the possibility of wound infection. In this paper, Logistic logistic regression, SMOTE algorithm and confusion matrix are used to model the probability of infection after spinal fusion and internal fixation. In the positive confirmation analysis part, the information data of 449 clinical cases were selected for analysis, and 14 variables such as gender, age, number of internal fusion fixation segments, past medical history, intraoperative blood transfusion and bleeding volume were selected as research indicators to explore the related factors of postoperative infection. The classification method adopts two classifications. The two types of data are 'postoperative infection' and 'postoperative non-infection'. In the statistical description of the data, it is found that age, selection of internal fusion fixation segments, preoperative hospitalization days, cerebrospinal fluid leakage, and preoperative ASA score are all important factors affecting the incidence of postoperative infection.