Medication Use in Mothers with Mental Illness (MUMMI) - Project Summary / Abstract In 2020, approximately 20% of women were prescribed medication for their mental health, and about 10% received a psychotropic drug during pregnancy. However, the safety profiles of many medications in pregnancy are still evolving and often fail to consider the broad spectrum of maternal and child outcomes. Therefore, the goal of the proposed project is to examine the effects of psychotropic medication use in pregnancy on obstetric complications, maternal general health during pregnancy, neonatal outcomes, and offspring neurodevelopmental disorders, with careful consideration of the effects of untreated maternal mental illness. This research will leverage a large-scale (>300M patients), U.S.-based and nationally representative electronic health record dataset to conduct a target trial emulation. This approach adopts the core principles of a randomized clinical trial to effectively enroll all eligible pregnant women and “assign” them to treatment strategies (continue or discontinue their psychotropic medication), while balancing baseline and time-varying covariates between comparison groups. This framework is specifically designed to mitigate biases pervasive in conventional epidemiological approaches, including sample selection based on future information (e.g., restricting analyses to individuals with complete follow-up data or livebirths), which can distort observed relationships; and improper alignment of follow-up time, which can classify individuals as exposed during intervals when the outcome was not yet possible. Causal effects will be estimated using modern, doubly robust, machine-learning-based methods which avoid imposing parametric assumptions in prediction modeling, further reducing bias and increasing robustness of study findings. The training plan includes supplemental coursework in causal inference, applied statistics, and pharmacokinetics. Additionally, its robust mentoring structure includes an interdisciplinary team of experts spanning perinatal epidemiology, causal inference, maternal-fetal medicine, psychiatry, bioethics, pharmacotherapy, and health informatics. Each member is highly committed to supporting this important research and brings valuable insight from their respective specialties. The proposed work is positioned to produce valuable insights for clinicians, researchers, pregnant women, and the general public regarding the safety of medications during pregnancy and their impact on maternal and child outcomes.
