Establishment of Bp-ann Based Risk Assessment Model for Lower Extremity Deep Venous Thrombosis in Critically Ill Patients : Mullti-centre, Prospective, Descriptive Correlational Study

Abstract

Background The objective of this study was to establish a risk assessment model for lower extremity deep venous thrombosis in critically ill patients and compared with Caprini, Padua and Wells risk assessment model to evaluate its efficacy. Methods We conducted a pooled analysis of prospective cohort studies. The outcomes of interest were lower extremity deep vein thrombosis group and Non-lower extremity deep vein thrombosis group were determined by univariate analysis, and SPSS was used to establish the back propagation artificial neural network prediction model. ROC curve was used to evaluate the predictive effectiveness of the model. Medcalc15.2 was used to compare the predictive capabilities of different models. Results 600 cases of intensive care unit patients were selected in this study. The prevalence of lower extremity venous thromboembolism after ICU admission was 12.5%. The results of univariate analysis that showed 16 statistically significant difference influencing factors. The ROC curve area of BP-ANN risk assessment model was 0.828, showing good predictive efficacy. In addition to the ROC curve area of BP-ANN risk assessment model was higher than Caprini, Padua and Wells model. Conclusion BP-ANN risk assessment model can play an auxiliary role in predicting the occurrence of lower extremity venous thromboembolism in critically ill patients. This model can provide a reference for medical staff to take preventive management measures.