Dr. Chengxi Zang, assistant professor of population health sciences, has received an R01 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development to bridge critical gaps in trials for pregnant women using electronic health records (EHR), machine learning (ML), and artificial intelligence (AI) techniques.
Pregnant women are routinely excluded from most clinical trials, whether due to ethical or safety concerns, with nearly 95 percent of clinical trials submitted to the FDA excluding pregnant women. Even existing obstetric trials are limited by safety concerns, small sample sizes, short time frames, early discontinuation, and rigorously controlled conditions that do not reflect real-world experiences.
Given that almost all pregnant women take at least one medication during pregnancy, this gap in evidence presents challenges for clinicians who advise treatment for pregnant women. These gaps are exacerbated by the lack of data on acute- and post-acute concerns of COVID-19 acquired during pregnancy, as well as the safety of using Selective Serotonin Reuptake Inhibitors (SSRIs), glucagon-like peptide 1 (GLP-1) agonists, and sodium-glucose co-transporter-2 (SGLT2) inhibitors during the perinatal period.
To address these critical gaps in evidence, Dr. Zang and his team propose novel causal ML approaches and AI-based tools to generate real-world evidence (RWE) for pregnant women and to inform pregnancy trial eligibility criteria design. They will leverage two large real-world data (RWD) resources, the INSIGHT and OneFlorida+ clinical research networks (CRNs) contributing to the national Patient-Centered Clinical Research Network (PCORnet). The rapid adoption of EHR systems allows these RWD to reflect the characteristics and outcomes of patients being treated in real-world settings available for research.
Furthermore, advancements in AI and causal ML offer new opportunities to generate RWE and eligibility criteria design for pregnancy trials. Dr. Zang’s team will develop natural language processing-assisted algorithms and tools to extract pregnant cohorts, key patient characteristics, and outcomes relevant to pregnant women from RWD. They will then develop a target trial emulation framework to estimate real-world safety and effectiveness outcomes of interventions from distributed CRNs.
Finally, Dr. Zang’s team will develop an RWE-based trial design system to quantify the impact of individual eligibility criteria on patients’ clinical outcomes regarding both effectiveness and safety. This system will use causal AI and follow a user-centered design to help investigators design eligibility criteria considering real-world patient outcomes and population characteristics. The system will also be evaluated and disseminated by obstetrical trialists.
Dr. Zang is working alongside Dr. Fei Wang, associate dean of AI and chief of the Division of Health Informatics and AI; Dr. Judy Zhong, chief of the Division of Biostatistics; Dr. Yiye Zhang, associate professor of population health sciences in emergency medicine; Dr. Rochelle Joly, assistant professor of clinical obstetrics and gynecology; Dr. Qian Yang, assistant professor of information science at the Cornell University Ann S. Bowers College of Computing and Information Science; Dr. Jie Xu, assistant professor of health outcomes and biomedical informatics at the University of Florida, and Dr. Ann M. Bruno, assistant professor in the Department of Obstetrics and Gynecology at the University of Utah Health.
The long-term goal for this research is to bridge the evidence gap for pregnant women, such that future pregnancy studies are safer, more inclusive, and more relevant to real-world patients. With real-world data, AI, and privacy-preserving methods, Dr. Zang and colleagues strive to improve evidence-based care for pregnant women and their infants.
- Highlights