Construction and validation of a risk prediction model for venous thromboembolism in hospitalized children
Objective To establish a risk prediction model for the occurrence of venous thromboembolism(VTE)in hospitalized children and to validate the application of the model.Methods Retrospective analysis was conducted on 62 children who developed VTE during hospitalization in our hospital from January 2022 to July 2023 and were selected in a VTE group.In the ratio of 1∶4,248 children from the same department and time who did not develop VTE were selected in a non-VTE group.Logistic regression analysis was used to develop the risk prediction model.Hosmer-Lemeshow test and receiver operator characteristic(ROC)curves were used to test the goodness of fit and predictive effectiveness of the model.Bootstrap self-help method was used for the internal validation of the model by re-sampling the sample l,000 times.Results In our study,the incidence of VTE was 0.43‰.Logistic regression analysis showed that the length of central venous access device(PICC)stay ≥3 d,stay with fully implantable intravenous port,chemotherapy,ICU length of stay ≥4 d,and bracing duration＞72 h were the independent risk factors for the development of VTE in hospitalized children(P＜0.05).Hosmer-Lemeshow test showed that x2=5.291,P=0.152.Sensitivity was 0.857,specificity was 0.218,and area under the R0C curve was 0.823 by Bootstrap self-help method resampling method.Conclusion The risk predictive model of VTE risk in hospitalized children constructed in this study has good predictive performance,which can be used for clinical medical staff to identify children at high risk of VTE at an early stage and take preventive measures.
NETL
NSTL
万方数据
