Improving data harmonization in pregnancy safety studies: Demonstrating the feasibility of standardized data analysis of pregnancy reports–A contribution from the conception project.

Introduction: Aligned with the Innovative Medicines Initiative (IMI) ConcePTION project’s goal of advancing drug safety research in pregnancy, we explored the feasibility of employing a standardized list of core data elements (CDEs) and statistical outputs to harmonize data collection and analysis across various study partners. Methods: Study partners from public research institutions and pharmaceutical companies extracted pregnancy safety data on disease-modifying therapies for multiple sclerosis from their databases, mapped them to the CDE framework, and analysed datasets using a common statistical analysis plan (SAP). Questionnaires were used to gather contextual information on data collection practices, including data definitions and infant follow-up, and to evaluate the feasibility and practicality of the proposed methodology. Results: The study demonstrated the feasibility of harmonizing pregnancy safety data using standardized CDEs and a common SAP. All six study partners successfully mapped their data into the CDE framework and produced standardized outputs. Challenges were identified in aligning with recommended definitions, such as classifying ‘prospective cases’ based on differing criteria—pregnancies reported while ongoing, before fetal malformation diagnosis, or before any prenatal screening—and categorizing congenital anomalies using the EUROCAT system. The feasibility questionnaires revealed strong support for the methodology, with all six study partners reporting access to the necessary expertise and all six finding the statistical guidance easy to follow. Conclusions: This study demonstrates the feasibility of employing standardized CDEs and a common SAP to describe pregnancy exposure and outcome data across diverse data providers. The methodology shows promise for improving consistency in pregnancy safety research but requires addressing challenges such as resource constraints and variability in data collection for broader adoption.