A Model to Predict Vaginal Delivery and Maternal and Neonatal Morbidity in Low-Risk Nulliparous Patients at Term

Academic Article


  • Objective: This study aimed to develop and validate a model to predict the probability of vaginal delivery (VD) in low-risk term nulliparous patients, and to determine whether it can predict the risk of severe maternal and neonatal morbidity. Methods: Secondary analysis of an obstetric cohort of patients and their neonates born in 25 hospitals across the United States (n = 115,502). Trained and certified research personnel abstracted the maternal and neonatal records. Nulliparous patients with singleton, nonanomalous vertex fetuses, admitted with an intent for VD ≥ 37 weeks were included in this analysis. Patients in active labor (cervical exam > 5 cm), those with prior cesarean and other comorbidities were excluded. Eligible patients were randomly divided into a training and test sets. Based on the training set, and using factors available at the time of admission for delivery, we developed and validated a logistic regression model to predict the probability of VD, and then estimated the prevalences of severe morbidity according to the predicted probability of VD. Results: A total of 19,611 patients were included. Based on the training set (n = 9,739), a logistic regression model was developed that included maternal age, body mass index (BMI), cervical dilatation, and gestational age on admission. The model was internally validated on the test set (n = 9,872 patients) and yielded a receiver operating characteristic-area under the curve (ROC-AUC) of 0.71 (95% confidence interval [CI]: 0.70-0.72). Based on a subset of 18,803 patients with calculated predicted probabilities, we demonstrated that the prevalences of severe morbidity decreased as the predicted probability of VD increased (p < 0.01). Conclusion: In a large cohort of low-risk nulliparous patients in early labor or undergoing induction of labor, at term with singleton gestations, we developed and validated a model to calculate the probability of VD, and maternal and neonatal morbidity. If externally validated, this calculator may be clinically useful in helping to direct level of care, staffing, and adjustment for case-mix among various systems.
  • Authors

    Published In

    Digital Object Identifier (doi)

    Author List

  • Costantine MM; Sandoval G; Grobman WA; Bailit JL; Reddy UM; Wapner RJ; Varner MW; Thorp JM; Caritis SN; Prasad M
  • Start Page

  • 786
  • End Page

  • 796
  • Volume

  • 39
  • Issue

  • 7