Prediction of severe adverse neonatal outcomes at the second stage of labor using machine learning: a retrospective cohort study.

OBJECTIVE To create a personalized machine-learning model for prediction of severe adverse neonatal outcomes (SANO) during the second stage of labor. DESIGN Retrospective Electronic-Medical-Record (EMR) based study. POPULATION OR SAMPLE A cohort of 73,868 singleton, term deliveries that reached the second stage of labor, including 1,346 (1.8%) deliveries with SANO. METHODS A gradient boosting model was created, analyzing 21 million data points from antepartum features (e.g. gravidity and parity) gathered at admission to the delivery unit, and intrapartum data (e.g. cervical dilation and effacement) gathered during the first stage of labor. Allocation of deliveries to high and low risk groups based on the Youden index to maximize sensitivity and specificity. Main Outcome Measures SANO was defined as either umbilical cord pH levels ≤ 7.1 or 1-minute or 5-minute Apgar score ≤7. RESULTS The model for prediction of SANO yielded an area under the receiver operating curve (AUC) of 0.761 (95% CI 0.748-0.774). A third of the cohort (33.5%, n=24,721) were allocated to a high-risk group for SANO, which captured up to 72.1% of these cases (OR=5.3 (95% CI: 4.7-6.0) high-risk vs. low-risk group). CONCLUSION Data acquired throughout the first stage of labor can be utilized to predict SANO during the second stage of labor using a machine learning model. Stratifying parturients at the beginning of the second stage of labor in a "time out" session, can direct a personalized approach to management of this challenging aspect of labor, as well as improve allocation of staff and resources.