ABSTRACT Pregnant people are a vulnerable population: healthcare professionals must exercise caution in prescribing many common pharmaceuticals to expectant patients, given potential risk of injury to a developing fetus. Arguably, teratogenicity is the most serious manifestation of fetal toxicity, as teratogens lead to fetal malformation and are implicated in physical and mental disabilities throughout the life course of an affected child. Consideration of teratogenicity, however, is a largely nonsystematic process. While regulatory agencies like the United States Food and Drug Administration have established discrete teratogenicity scores for evaluating drug safety, these classification criteria provide little concrete distinction among score classes, making it difficult for pharmaceutical scientists to systematically determine the teratogenic potential of a drug. This effect ripples to the bedside, as providers lack robust data on most drugs to inform their judgements of safety, efficacy, and risk in recommending therapies to expectant patients. Consequently, treatment of many diseases during pregnancy remains understudied and uncertain, for fear of causing harm. To address these shortcomings and improve quality of care for pregnant patients and their unborn children, we propose the development of a holistic, generalizable framework for the identification of teratogenic drug exposures via a new paradigm of clinical pharmacovigilance. Therefore, given our inability to evaluate drug safety in pregnant patients in real time, the objective of our investigation is to develop an observational platform to uncover new drug safety information through strategic, associative analysis of medication history within the electronic health record (EHR) of a pregnant patient and developmental diseases within the EHRs of their neonates. This approach will harness 48,434 linked maternal and neonatal EHRs within the Research Derivative, a databank of identified EHRs at Vanderbilt University Medical Center, across which we plan to apply and appropriately validate a medication-wide association study to uncover new, data-driven associations between maternal medication exposures and teratogenic health outcomes. To date, we have employed this technique to discover 129 associations between maternal drug exposure and neurodevelopmental diseases from existing primary care data, which recapitulate teratogens well-known to clinical practice that also serve as positive controls for our model. Now, we hope to expand our platform to encompass more advanced techniques of clinical trial emulation; if successful, we anticipate that discovering new toxicants through retrospective data modeling will greatly enhance our goal of systematically clarifying cases of suspect drug safety in pregnancy. In the long run, we hope this approach will allow clinicians to embrace a more accurate knowledge base in their evaluation of the risks and benefits in prescribing drugs with unclear teratoge...