Project Summary Postpartum hemorrhage (PPH) is the most common complication during childbirth, occurring after 10% of all deliveries, and is a significant contributor to maternal morbidity. While acute and active management of PPH is required to prevent morbidity, pregnant individuals without anemia are better equipped to tolerate delivery- associated blood loss without incurring morbidity. The primary cause of anemia in pregnancy is iron deficiency. There is insufficient evidence to support universal iron deficiency screening during routine obstetric care in the US. Currently, most individuals may only be screened for iron deficiency if they are anemic. This two-step screening process can lead to 1) failure to diagnose iron deficiency if the patient does not undergo the secondary work-up for anemia and 2) failure to adequately treat iron deficiency given the time it takes for oral therapy, the historical standard treatment, to replete iron stores and its side effects that can limit regular use. Thus, innovative strategies are needed to address iron deficiency and anemia in pregnancy, especially for the 10% of individuals who will have a PPH and are at an increased risk for severe maternal morbidity. The primary objective of this Phased Innovation Award (R21/R33) is to develop and test a clinical decision support (CDS) tool that proposes a novel iron deficiency screening and management algorithm for individuals at high risk for PPH. In Phase I (R21), structured data known at the end of the second trimester will be used to develop a machine learning- based predictive model to identify those at high risk for PPH. We will then build a CDS tool within the electronic health record (EHR) that will prompt providers to proactively screen and treat iron deficiency for patients at high risk for PPH. In Phase II (R33), we will test this CDS tool’s efficacy in reducing the prevalence of anemia before delivery via a randomized controlled trial. This phase will also monitor the acceptability of the CDS tool among obstetric providers in the intervention group. Based at Massachusetts General Hospital, this work is led by an experienced, multidisciplinary team of researchers with expertise in machine learning, informatics, practical application of risk stratification tools, and clinical obstetrics. Ultimately, the proposed work seeks to improve maternal health outcomes and accelerate the application of artificial intelligence-aided clinical tools in obstetrics at the point of care. Specifically, it will demonstrate how a digital health tool, which uses a personalized risk assessment, can be integrated into clinical workflows within prenatal care and offers an actionable, resource- conscious strategy to address the ongoing public health crisis related to maternal morbidity in the US.