Improving Postpartum Hemorrhage Risk Prediction Through Machine Learning

Posted on November 15, 2021 | Jazmin Juarez

Sema4 is committed to providing science-driven insights and solutions to address the most pressing medical needs. In collaboration with clinicians from the Mount Sinai Health System, we recently published two studies demonstrating the utility of machine learning to predict clinical outcomes for postpartum hemorrhage (PPH), the world's leading cause of maternal death.

PPH is a preventable medical emergency resulting in excessive blood loss during or after delivery, with or without clinical symptoms. With one to 10 percent of deliveries at risk, PPH is among the top five causes of maternal death globally. It accounts for around a third of maternal deaths and often occurs in patients with no known risk factors for hemorrhage. While PPH is preventable, inconsistent diagnostic guidelines and risk assessment tools make the condition difficult to recognize, diagnose, and adequately address promptly to prevent maternal mortality.

The two new studies, led by Li Li, MD, SVP of Clinical Informatics at Sema4, report our success in leveraging real-world, large-scale clinical data to improve the detection of PPH risk factors. Both papers are currently available online and will soon appear in print in a special "Informatics for Sex- and Gender-Related Health" issue of the Journal of the American Medical Informatics Association (JAMIA).

In the first study, we reviewed over 70,000 de-identified health records from a diverse cohort across five Mount Sinai Health Systems Hospitals using Centrellis®, our advanced health intelligence platform. This longitudinal data was used to create and validate an accurate PPH digital phenotyping algorithm. The same patient cohort was then used in the second study to develop a more confident risk assessment tool to predict the risk of PPH clinically and to compare it to existing risk models.

"These two new papers are among the first to use large-scale, comprehensive real-world data to predict clinical outcomes," said Eric Schadt, PhD, Founder and Chief Executive Officer of Sema4 and joint corresponding author, in a recent press release on the papers. "By implementing our predictive model into the clinical standard of care, healthcare providers may be able to improve PPH risk assessment and medical management for their pregnant patients resulting in better health outcomes."

Our phenotyping algorithm detected additional clinical variables that could better identify PPH with 89% accuracy compared to standard clinical tools with 67% accuracy. Furthermore, we identified 24 risk markers - including five new potential markers - that could alert clinicians to higher PPH risk, along with inflection points in laboratory and vital sign values that also suggest increased risk. These newly-identified predictors are readily obtained from routine blood tests and vital sign parameters but are not currently utilized in standard risk assessment tools.

Following further evaluation, this tool may enable reproductive health care providers to predict and treat potential high-risk symptoms before they occur, allocate resources appropriately, and ultimately reduce PPH morbidity and mortality.

_ _ _

To learn more about how Sema4 is using data-driven insights to improve maternal health outcomes, please see our recent blog post featuring Dr. Li and her work on preeclampsia, the second leading cause of maternal mortality worldwide. For more information on Sema4 Elements, our portfolio of information-driven genomic solutions, digital tools, and services that enable providers to treat patients holistically during their reproductive and generational health journey, please click here.

Public link

Please use the above public link if you want to share this noodl on another website.