Comparison of risk assessments for venous thromboembolism during the puerperium

Abstract

ObjectiveTo compare the predictive value of different risk assessment methods for puerperium venous thromboembolism (VTE). MethodsThis study included 55 women with and 165 women without puerperal VTE. Using the cases, 11 assessment methods were compared. ResultsThe area under the curve (AUC) value of the 11 assessments was highest for the modified Caprini risk assessment model for pregnancy (a modified risk scoring method from Caprini, AUC = 0.805). Pairwise comparison of the AUC values of the 11 assessment methods indicated no significant difference among the five methods with AUC values > 0.7. Among them, the modified Caprini, the risk scoring method recommended by the Swedish Guidelines (Swedish method), and the risk scoring method recommended by the Shanghai consensus (Shanghai method) performed better than the other six methods with AUC values < 0.7 (P < 0.05). The sensitivities of the five methods for predicting a high risk of VTE were 69.09–94.55% and the specificities were 25.45–77.58%. The sensitivity of the modified Caprini was higher than those of the risk management method from the Chinese consensus (Chinese consensus method), Royal College of Obstetricians and Gynaecologists risk assessment scale (RCOG), and Swedish method (P < 0.05), but the specificity was only 25.45%. No significant difference in sensitivity was detected among the Swedish, Shanghai, RCOG, and Chinese consensus methods, whereas the specificity of the Swedish method was higher than that of the Shanghai, RCOG, and Chinese consensus methods. ConclusionThe predictive value of different risk assessment methods for puerperium VTE varies greatly. Considering the sensitivity and specificity, the Swedish method may have better clinical application value among the 11 methods.