Development of a nomogram for prediction of successful membrane sweeping*.

Objective: To evaluate the association of obstetric, maternal, and fetal variables with successful membrane sweeping and to develop a calculator that can predict spontaneous delivery within 24 hours of membrane sweeping. Methods: This secondary analysis of the STRIP-G Study included all singleton term parturients who underwent membrane sweeping in a tertiary center in October 2011 to July 2013. Primary end point was a 24-hour time interval from sweeping to delivery. Women who delivered without formal induction within the 24-hour interval were included in the "successful sweeping group". Stepwise logistic regression was used to calculate the adjusted odds ratio (aOR) for successful membrane sweeping and to create the calculator. The predictive power of the calculator was evaluated by area under the curve (AUC) of the receiver operating characteristic (ROC) curve and by Nagelkerke R-square. The model was validated by the Hosmer-Lemeshow test and by these validation measures: sensitivity, specificity, and positive and negative predictive value. Results: We analyzed data from 542 women. Parity (aOR = 1.66, 95% confidence interval [CI] 1.1-2.54), cervical dilation (aOR = 3.33, 95%CI 2.04-5.44), and gestational age (aOR = 1.44, 95%CI 1.21-1.72) were independent predictors of spontaneous delivery during the first 24 hours. A cross validation procedure showed that the calculator had a good accuracy (68%). Conclusions: A simple calculator based on maternal age, parity, gestational age, cervical dilatation, effacement and station, can accurately predict the chances of delivery within 24 hours of membrane sweeping. This may assist physicians better counseling of women regarding the likelihood of successful membrane sweeping. [ABSTRACT FROM AUTHOR]