# Machine Learning Addresses Drug-Safety Evidence Gaps in Pregnancy > Source: > Published: 2026-05-27 16:20:52.245998+00:00 # Machine Learning Addresses Drug-Safety Evidence Gaps in Pregnancy The JMIR News and Perspectives article by Michelle Falci reports that women, and especially pregnant women, have been systematically excluded from clinical research following historical FDA guidance and related policies. According to JMIR, a survey found that **79%** of women take at least one medication during pregnancy, with **63.2%** making at least one medication change and **36.5%** modifying or stopping medication without clinician input. The article notes global efforts such as a World Health Organization task force aiming for ethical inclusion of pregnant and breastfeeding women in clinical research by **2030**, and it highlights the FDA Pregnancy and Lactation Labeling Rule (PLLR) as an incomplete but recent step. JMIR reports researchers are turning to large datasets and machine learning, with an emphasis on interpretability and causal inference, to help close these evidence gaps. ### What happened The JMIR News and Perspectives article by Michelle Falci documents longstanding exclusion of women, particularly pregnant women, from clinical trials, tracing the trend back to post-thalidomide policy and guidance, and noting subsequent policy interventions such as the **NIH Revitalization Act**. According to JMIR, a survey of over 1,200 women found **79%** take at least one medication during pregnancy, **63.2%** made at least one medication change, and **36.5%** altered medication use without health care provider input. JMIR also reports that the **World Health Organization** established a task force with a goal of timely, ethical inclusion of pregnant and breastfeeding women in clinical research by **2030**, and that the **FDA Pregnancy and Lactation Labeling Rule (PLLR)** has improved labeling but leaves many evidence gaps. ### Technical details (reported) Per JMIR, researchers advocate using **large datasets** combined with machine learning approaches while prioritizing **interpretability** and **causal inference** to avoid black-box conclusions. The article frames these methods as tools to extract signals about safety from observational sources rather than as replacements for randomized evidence. ### Editorial analysis - technical context For practitioners: applying machine learning to pregnancy safety is primarily a causal-inference and data-quality problem. Industry-pattern observations suggest that effective work in this area typically combines target-trial emulation, propensity methods, negative controls, and model explainability techniques to mitigate confounding and selection bias. Privacy, small-cohort sizes for specific exposures, and linkage across registries and electronic health records create operational and statistical challenges that influence model design and evaluation. ### Context and significance closing pregnancy evidence gaps affects clinical decision making, regulatory labeling, and patient safety. Machine learning and real-world evidence (RWE) pipelines can reduce latency between signal detection and actionable estimates, but their outputs require careful triangulation with epidemiologic methods to be credible for clinicians and regulators. ### What to watch For practitioners: indicators to follow include broader inclusion of pregnancy cohorts in longitudinal EHR datasets and registries, advances in causal ML that explicitly quantify uncertainty, updates from the WHO task force on timelines toward **2030**, and regulatory guidance clarifying acceptability of RWE-based safety assessments. ## Scoring Rationale This is a notable application-area story: it highlights concrete, practitioner-facing use of ML to address a persistent evidence gap in clinical safety. The technical and data challenges mean this is important for ML practitioners working in health, but it is not a frontier-model or regulatory watershed. Practice interview problems based on real data 1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with. [Try 250 free problems](/problems)