A Computational Prediction Model for Uterine Rupture Risk in Women Undergoing Trial of Labour After Caesarean in Nigerian Tertiary Hospitals

Uterine rupture during trial of labour after caesarean (TOLAC) is a catastrophic obstetric complication, and the absence of a validated rupture risk prediction tool adapted to Nigerian clinical and demographic parameters limits safe TOLAC candidate selection in tertiary hospitals. This dissertation develops and validates a computational prediction model for uterine rupture risk during TOLAC in Nigerian tertiary hospitals. A retrospective multi-site cohort enrolled 600 women who underwent TOLAC at teaching hospitals in Lagos, Kano, Enugu, Port Harcourt, Abuja, and Sokoto. Clinical predictors including inter-delivery interval, prior caesarean indication, number of prior caesareans, induction versus spontaneous labour, and cervical dilation at admission were extracted from case notes. A training set of 420 cases was used to develop logistic regression and gradient boosting machine models, compared on AUC, calibration, and decision curve analysis. External validation was performed on a prospective cohort of 180 TOLAC cases at Obafemi Awolowo University Teaching Hospital. Uterine rupture occurred in 3.8 percent of TOLAC cases. The gradient boosting model achieved AUC of 0.87 in the training set and 0.83 in external validation. Inter-delivery interval below 18 months (OR 5.1), more than one prior caesarean (OR 4.4), and induction with oxytocin (OR 3.8) were the strongest predictors. A web-based TOLAC risk calculator was developed and is available for clinical use. The dissertation provides an original computational tool for Nigerian TOLAC risk stratification. Keywords: uterine rupture, trial of labour after caesarean, prediction model, Nigeria, TOLAC