Developing and Validating a Prediction Model for Postpartum Hemorrhage at a Single Safety-net Tertiary Care Center [11M

Abstract

INTRODUCTION: The aim of the current study is to develop and validate a prediction model for postpartum hemorrhage (PPH) based on antenatal and intrapartum risk factors in a diverse patient population at a tertiary safety-net center. METHODS: In this observational cohort of 10,025 pregnant women who delivered at Denver Health Medical Center (DHMC) between April, 2016 and March, 2019, ante- and intrapartum characteristics were compared between women with PPH and those without. A predictive multivariable model was estimated using logistic regression with a backwards stepwise approach eliminating variables when P>.05 among the prediction cohort of women delivering before April 2018. The model was validated among women, who delivered after April, 2018. RESULTS: A total of 9,774 women met inclusion criteria, 6,525 in the prediction cohort and 3,249 in the validation cohort. The prevalence of PPH in the prediction and validation cohorts was 6.1% and 6.8%, respectively. Maternal age, maternal body mass index (BMI, kg/m2), parity, gestational age, history of prior cesarean, history of PPH, abnormal placentation, neonatal weight, chorioamnionitis, platelet count on admission and intrapartum magnesium administration were included in the final prediction model. The area under the curve (AUC) for the prediction model was 0.81 (95% CI 0.79-0.84) and the AUC for the validation cohort was 0.89 (95% CI 0.86-0.91). CONCLUSION: The model performed well and can be used to identify women at risk for PPH. Further studies are necessary to evaluate the effect of the model use on decreasing PPH and associated maternal morbidity.