Building a predictive model for successful vaginal delivery in nulliparas with term cephalic singleton pregnancies using decision tree analysis.

AIM: To establish a model predicting successful vaginal delivery (VD) in nulliparas with term cephalic singleton pregnancies.

METHODS

We retrospectively identified 6799 term nulliparas with cephalic singletons (6416 VD and 383 cesarean section [CS] due to dystocia) who entered labor (cervical dilation ≥2 cm) between September 2014 and August 2015. Using VD as the dependent variable and age, maternal body height, educational attainment, gravidity, gestational age, pre-pregnancy body mass index (BMI), BMI upon admission for delivery, gestational weight gain, gestational hypertension and gestational diabetes as the independent variables, predictors of VD success were identified using a multivariate binary logistic regression and then ranked with decision-tree analysis.

RESULTS

While multiple factors are associated with improved VD success, we found body height, gestational age, and intrapartum BMI to be the best predictors of successful VD. Our predictive model has a classification accuracy, sensitivity and specificity of 76.6%, 96.7% and 16.4%, respectively, and it was subsequently confirmed by both internal and external validation.

CONCLUSION

Our predictive model indicates body height, gestational age and intrapartum BMI as the major predictors of successful VD in low-risk patients.