External validation of a prediction model on vaginal birth after caesarean in a The Netherlands: a prospective cohort study

Objectives: Discussing the individual probability of a successful vaginal birth after caesarean (VBAC) can support decision making. The aim of this study is to externally validate a prediction model for the probability of a VBAC in a Dutch population. Methods: In this prospective cohort study in 12 Dutch hospitals, 586 women intending VBAC were included. Inclusion criteria were singleton pregnancies with a cephalic foetal presentation, delivery after 37 weeks and one previous caesarean section (CS) and preference for intending VBAC. The studied prediction model included six predictors: pre-pregnancy body mass index, previous vaginal delivery, previous CS because of non-progressive labour, Caucasian ethnicity, induction of current labour, and estimated foetal weight 90th percentile. The discriminative and predictive performance of the model was assessed using receiver operating characteristic curve analysis and calibration plots. Results: The area under the curve was 0.73 (CI 0.69-0.78). The average predicted probability of a VBAC according to the prediction model was 70.3% (range 33-92%). The actual VBAC rate was 71.7%. The calibration plot shows some overestimation for low probabilities of VBAC and an underestimation of high probabilities. Conclusions: The prediction model showed good performance and was externally validated in a Dutch population. Hence it can be implemented as part of counselling for mode of delivery in women choosing between intended VBAC or planned CS after previous CS.