Predicting Postpartum Hemorrhage Risk With Second-Trimester Data.

Area of Science:

• Obstetrics and Gynecology
• Health Informatics
• Predictive Analytics in Healthcare

Background:

• Postpartum hemorrhage (PPH) remains a leading cause of maternal morbidity and mortality.
• Accurate, early risk stratification is crucial for timely intervention and improved patient outcomes.
• Existing tools often lack the ability to stratify risk prenatally using readily available data.

Purpose of the Study:

• To develop and validate a risk stratification tool for postpartum hemorrhage (PPH).
• To utilize structured data from electronic health records (EHR) available before the third trimester.
• To enable proactive predelivery planning and health optimization for high-risk pregnancies.

Main Methods:

• Retrospective cohort study of 17,201 patients delivering between 2017 and 2023.
• Development using logistic regression with stepwise backward selection on training data (2017-2022).
• Validation in independent testing (2017-2022) and temporal validation (2023) cohorts using data known by 24 weeks gestation.

Main Results:

• The final model demonstrated good discrimination with an area under the curve of 0.689 (testing) and 0.659 (validation).
• The tool successfully identified high-risk patients with screen-positive rates around 20% and positive predictive values exceeding 20%.
• Model performance was consistent across testing and temporal validation datasets, confirming robustness.

Conclusions:

• A reliable prenatal risk stratification model for PPH can be developed using structured EHR data available by 24 weeks gestation.
• This tool facilitates predelivery planning and targeted management for individuals at high risk of PPH.
• The model offers an advantage over post-admission tools by enabling earlier intervention strategies.