Establishment And Verification Of Prediction Model Of Venous Thromboembolism After Gynecological Surgery

Objectives:To establish a clinical prediction model by screening and analyzing the risk factors of venous thromboembolism(VTE)after surgery in gynecology,and to validate and evaluate the model.Methods:The case data of gynecological inpatients undergoing surgery in the First Affiliated Hospital of Kunming Medical University from September 2016 to September 2022 were collected for retrospective study.The clinical data were analyzed,independent risk factors were screened out,a line graph model was established,and the prediction model was externally verified and the prediction effect was evaluated.Results: 1.Univariate analysis results showed that: Compared with patients who developed VTE after the training concentration,there was a significant difference in age(P<0.001),Malignant tumor(P<0.001),operation duration(P<0.001),hypertension(P<0.001),BMI(P=0.001),plasma D-dimer(P<0.001)and plasma Fib(P=0.045)were statistically different.2.Results of multi-factor analysis showed that: Age(OR=1.04,95%CI:1.01 ～1.07,P=0.014),malignancy(OR=2.28,95%CI:1.05 ～ 4.92,P=0.036),operation duration ≥ 3h(OR=3.58,95%CI:1.20 ～ 11.86,P=0.027),hypertension(O R=2.16,95%CI:1.01 ～ 4.60,P=0.045),BMI(OR=1.04,95%CI:1.07 ～ 4.71,P=0.031)and plasma D-dimer(OR=1.00,95%CI:1.00 ～ 1.19,P=0.044)were Vtes after gynecological surgery Were independent risk factors for VTE.3.Age,malignancy,operation duration≥3h,hypertension,BMI,and plasma D-dimer were used as predictors to establish a predictive model with an area under receiver operating characteristic curve(ROC)of 0.811(95%CI:0.750-0.871).Hosmer-Lemeshow goodness of fit test p = 0.734 > 0.05.4.The area under ROC curve of the external validation set was 0.772(95%CI:0.684 ～0.859).Conclusions: 1.Age,malignant tumor,operation duration,hypertension,BMI,and plasma D-dimer were the risk factors for postoperative VTE.2.The individual prediction model expression of VTE after gynecological surgery was Y=-6.194+0.039 × age+0.824 × malignant tumor +0.994 × operation duration(1 ～ 3h)+1.276 × operation duration(≥ 3h)+0.771 × hypertension +0.088 ×D-Dimer+0.811×BMI.3.The Logistic prediction model has a good degree of differentiation and calibration in the training set and verification set,which has certain clinical prediction value for early detection and diagnosis of VTE after gynecological surgery.