Cathrine Axfors

Research problem and specific questions: Most prescription drugs lack enough evidence on their safety for use during pregnancy. Post-marketing studies commonly focus on few selected drugs and indications, while unexpected safety signals can be missed, as may the chance to generate timely evidence on safe use. Our overall aim is to systematically gather new evidence for the safety of a comprehensive range of prescription drugs during pregnancy, using national health registers. Specific aims are to (1) implement a medication-wide association study of drug safety in pregnancy and evaluate its performance, and (2) apply statistical triangulation methods and structured qualitative review to identify relevant safety signals and evidence of safe use.Data and method: We will systematically investigate the safety of the full range of prescription drug classes (i.e., “medication-wide”) in Swedish national register data (&gt

1.5 million pregnancies 2006-2024). We will refine a hypothesis-generating statistical method (TreeScan™) suitable for hierarchical outcomes like diagnoses and that adjusts for multiple comparisons. The models will use a predefined set of important fetal/neonatal outcomes, e.g., congenital anomalies, and propensity score stratification with predefined covariables for confounding adjustment.Results from hypothesis-generating methods need replication. Therefore, found signals will be investigated also with in-depth pharmacoepidemiological models. An interdisciplinary expert panel with, e.g., pharmacoepidemiologists, obstetricians, and patient representatives, will perform a structured qualitative review to identify signals and non-signals of relevance. They will consider, e.g., statistical robustness, consistency with prior evidence, and clinical and biological plausibility.Societal relevance and utilisation: Missing drug safety evidence in pregnancy is a critical public health concern. It can cause both unsafe use and undertreatment of chronic and acute illness. This project will yield new systematic evidence on drug safety for the benefit of pregnant women and their children. It will also provide a novel active surveillance method for Swedish and international implementation.Plan for project realisation: The applicant will lead a team with a postdoctoral researcher and statistician and coordinate the work of the international expert panel. The budget includes salaries and costs for data, travels, and meetings.

Most prescription drugs lack enough evidence on their safety for use during pregnancy. Our overall aim is to systematically gather new evidence for the safety of a comprehensive range of prescription drugs during pregnancy, using national health registers.We will investigate the safety of every prescription drug class (i.e., “medication-wide”) in Swedish national register data 2006-2024, regarding important fetal/neonatal outcomes (e.g., congenital anomalies). We will refine a hypothesis-generating statistical method (TreeScan, suitable for diagnostic outcomes) with multiple comparison adjustment and propensity score stratification.Results from hypothesis-generating methods need corroboration. Therefore, found signals will be investigated also with in-depth pharmacoepidemiological models. Also, an interdisciplinary expert panel will qualitatively review results, considering, e.g., statistical robustness, consistency with prior evidence, and clinical and biological plausibility.The applicant will lead a team of a postdoc and statistician (medication-wide analysis, yrs 1-2, corroboration analyses, yrs 3-4) and lead the expert panel’s work (qualitative review, yrs 3-4).Missing drug safety evidence in pregnancy can cause both unsafe use and undertreatment of chronic and acute illness. This project will yield new systematic evidence on drug safety to inform clinical decision-making. It will also provide a novel surveillance method for Swedish and international implementation.