Construction and Validation of the Risk Prediction Model for the Recurrence of Unprovoked Venous Thromboembolism
Objective To explore the risk factors for recurrence in patients with unprovoked venous thromboembolism,establish a predictive model,and test the predictive performance of the model.Methods Two hundred cases of unprovoked venous thromboembolism recurrence and 200 cases of non-recurrence unprovoked venous thromboembolism patients who were treated in the First Affiliated Hospital of Zhengzhou University from September 2017 to August 2019 were selected as the modeling group.The risk factors for unprovoked venous thromboembolism recurrence were analyzed and the prediction model was established.The Hosmer-Lemeshow test judges the goodness of fit of the model,and the receiver operating characteristic(ROC)curve tests the discriminative validity of the model.One hundred and sixty-six patients with a history of unprovoked venous thromboembolism who were treated in the same hospital from September 2019 to August 2020 were selected as the verification group to evaluate the clinical application effect of the prediction model.Results A total of factors including the type of first-time venous thromboembolism,male,post-thrombotic syndrome,D-dimer level and other factors were included to construct a predictive model.The Hosmer-Lemeshow test results of the model showed that P=0.087 and the area under the ROC curve of construction group was 0.841,the area under the ROC curve of validation group was 0.740,sensitivity was 69.04％,specificity was 82.26％,accuracy was 78.91％,indicating that the model has good goodness of fit and discrimination validity.Conclusion Factors such as the type of first-time venous thromboembolism,maleness,post-thrombotic syndrome,D-dimer level and other factors can increase the risk of recurrence in patients with unprovoked venous thromboembolism.The constructed model has a good predictive effect,and can provide a reference for medical staff to screen high-risk groups of recurrence in time and take preventive management measures.
NETL
NSTL
万方数据
