Objective:To investigate the construction and clinical application value of perioperative risk prediction model of lower extremity deep vein thrombosis in patients with femoral fracture.Methods:A total of 1424 patients with femoral fracture surgery admitted to our hospital from January 2019 to December 2023 were selected as the study objects.Age,gender,hemoglobin,D-dimer,platelets,fibrinogen,hypertension,diabetes,smoking history,Caprini score,ADL score,BM1 and whether anticoagulant drugs were used were also extracted.70％of the research objects were randomly assigned to the model group and 30％to the verification group.The model group and the verification group were used to construct and verify the prediction model respectively.The suspected risk factors of lower extremity deep vein thrombosis in perioperative period of femur fracture patients were screened by single factor x2 test,and the independent risk factors of lower extremity deep vein thrombosis in perioperative period of femur fracture patients were determined by Logistic regression analysis.Based on independent risk factors,a prediction model was established to predict the incidence of lower extremity deep vein thrombosis in perioperative period of femur fracture patients.The differentiation degree,calibration degree and clinical validity of the prediction model were tested respectively.SPSS and R software were used for statistical a-nalysis.Results:There was no significant difference in the incidence of LEDVT between model group and verification group(P＞0.05).Univariate screening and multivariate analysis showed that age,combined hypertension,Caprini score,ADL score,no use of anticoagulants and platelet,fibrinogen and D-dimer levels were all risk factors for perioperative LEDVT in femur fracture patients(P＜0.05).The ROC curve AUC drawn according to LEDVT risk prediction model was 0.812,and the corresponding sensitivity and specificity were 94.17％and 94.95％,respectively.The goodness of fit test found that the test value of the model group was P=0.535,and that of the verification group was P=0.302.The results of DC A decision curve analysis show that the net benefit of this model is higher than that of the extreme curve in the threshold probability range of 0～0.96,and the clinical effect and application value are better.Conclusion:The LEDVT prediction model built based on the clinical characteristics and laboratory index detection results of patients with femoral fracture has high sensitivity and specificity,which can provide data reference for perioperative LEDVT risk assessment of surgical patients,and help formulate and improve the follow-up treatment plan.
Femoral FracturePerioperative PeriodLower Extremity Deep Vein ThrombosisRisk FactorsRisk Prediction Model
维普
万方数据
