Tan HS, Liu N, Sultana R, et al. - The performance of machine learning was compared with regression techniques in detecting parturients at elevated risk of breakthrough pain during labor epidural analgesia. They conducted a single-center retrospective study including data from 20,716 parturients who received patient-controlled epidural analgesia. Employing the training cohort comprising of randomly selected 80% of the cohort, they developed three prediction models for breakthrough pain using random forest, XGBoost, and logistic regression; this was followed by validation against the remaining 20% of the cohort (validation cohort). Breakthrough pain incidence of 14.2% was reported. Of 31 candidate variables, 30, 23, and 15 variables were included in random forest, XGBoost and logistic regression models, respectively. Similar performance was observed for the models, with area-under-the-curve 0.763-0.772. The fewest variables were required for the model derived using logistic regression. Findings thereby indicate no improvement in the prediction of breakthrough pain in correlation with machine learning (random forest and XGBoost) vs multivariable regression.