PURPOSE: To determine a predictive model for supporting decisions relating to the prognosis for women presenting with preterm labor and intact membranes. METHOD: We conducted a prospective observational cohort study on 70 pregnant women at between 22 and 34 weeks of gestation. Transvaginal sonographic evaluation on the cervix was performed once on women who had completed a course of parenteral tocolysis. The sonographic parameters of cervical length measurement and presence of cervical glandular area were obtained. The outcome variable was occurrences of preterm delivery (<35 weeks). Using a univariate logistic regression model, the coefficients of each independent variable were first estimated. To construct the predictive model, multivariate logistic regression containing all the selected variables that might be related to preterm delivery was used as the starting point. Accuracy, sensitivity, specificity and predictive values were used to measure associations of predicted probabilities and to check the ability of the model to predict outcomes. The predictive analyses were based on logistic regression models, with calculation of odds ratios and 95 % confidence intervals. RESULTS: The incidence of preterm delivery was 32.80 % (23/70). After validation, the predictive model proposed showed accuracy of 87.88 %, sensitivity of 78.26 % and specificity of 93.02 %. CONCLUSION: The model presented good accuracy with correspondence between predictions and observations, and has the capacity to become a useful tool for management of pregnant women with preterm labor and intact amniotic membranes.
PURPOSE: To determine a predictive model for supporting decisions relating to the prognosis for women presenting with preterm labor and intact membranes. METHOD: We conducted a prospective observational cohort study on 70 pregnant women at between 22 and 34 weeks of gestation. Transvaginal sonographic evaluation on the cervix was performed once on women who had completed a course of parenteral tocolysis. The sonographic parameters of cervical length measurement and presence of cervical glandular area were obtained. The outcome variable was occurrences of preterm delivery (<35 weeks). Using a univariate logistic regression model, the coefficients of each independent variable were first estimated. To construct the predictive model, multivariate logistic regression containing all the selected variables that might be related to preterm delivery was used as the starting point. Accuracy, sensitivity, specificity and predictive values were used to measure associations of predicted probabilities and to check the ability of the model to predict outcomes. The predictive analyses were based on logistic regression models, with calculation of odds ratios and 95 % confidence intervals. RESULTS: The incidence of preterm delivery was 32.80 % (23/70). After validation, the predictive model proposed showed accuracy of 87.88 %, sensitivity of 78.26 % and specificity of 93.02 %. CONCLUSION: The model presented good accuracy with correspondence between predictions and observations, and has the capacity to become a useful tool for management of pregnant women with preterm labor and intact amniotic membranes.