Estimating success of vaginal birth after caesarean section in a regional Australian population: Validation of a prediction model

Background Following a primary caesarean section (CS), women must decide between attempted vaginal birth after caesarean (VBAC) and elective repeat caesarean section (ERCS) in subsequent pregnancies. Both options carry potential morbidity and mortality for mother and child, with the most feared being uterine rupture and its consequences. In attempts to reduce morbidity, several predictive nomograms have been developed to assist in delivery mode decisions. Aim To assess the validity of the predictive nomogram developed by Grobman et al. in our regional Australian population. Materials and Methods In our retrospective analysis, patients at term, with one previous CS who had a trial of labour were assigned a 'Grobman score' based on antenatal details. Outcomes were noted and patient groups analysed according to percentage deciles of estimated VBAC success, compared with actual VBAC success rates. Results A total of 395 women underwent trial of labour after a single prior CS, with a VBAC success rate of 83%. The Grobman model displayed adequate calibration and the re-calibrated model good calibration with the slope coefficient of 0.87 (95% CI 0.54-1.19) and intercept 0.19 (95% CI -0.34-0.72). Discrimination was moderate with receiver operating characteristic area of 0.71 (95% CI 0.67-0.76). Conclusion This analysis supports further validation studies in larger Australian settings, and suggests that use of the original Grobman predictive nomogram may be appropriate in Australia.