Risk assessment tools to predict postpartum hemorrhage

Abstract

Postpartum hemorrhage (PPH) is a leading cause of maternal morbidity and mortality, and accurate risk assessments may allow providers to anticipate and prevent serious hemorrhage-related adverse events. Multiple category-based tools have been developed by national societies through expert consensus, and these tools assign low, medium, or high risk of hemorrhage based on a review of each patient's risk factors. Validation studies of these tools show varying performance, with a wide range of positive and negative predictive values. Risk prediction models for PPH have been developed and studied, and these models offer the advantage of more nuanced and individualized prediction. However, there are no published studies demonstrating external validation or successful clinical use of such models. Future work should include refinement of these models, study of best practices for implementation, and ultimately linkage of prediction to improved patient outcomes.

Keywords: artificial intelligence; machine learning; maternal morbidity; maternal mortality; postpartum hemorrhage (PPH); projections and predictions.