Background: The ENCePP Guide on Methodological Standards in Pharmacoepidemiology was recently updated to include an Annex 2 on methods for the evaluation of medicines in pregnancy and breastfeeding, noting the importance of fit-for-purpose data source for this type of research. A comprehensive assessment on the relevance of existing databases to answer common research questions pertaining to use of drugs or vaccines in pregnancy is lacking.
Objectives: To assess the fitness of use of existing databases to address most common research questions on drugs or vaccines safety during pregnancy.
Methods: The structured process to identify fit‐for‐purpose data (SPIFD) framework was adapted to assess the ability of population-based data sources across the big 4 European countries, the UK and select Nordic countries to capture minimal data needs for studying drug or vaccine safety in pregnancy. Data elements included pre-pregnancy comorbidities and parity, characteristics of the pregnancy and pregnant individual, pregnancy and birth complications and outcomes, and were informed by the IMI-ConcePTION project, Annex 2 of the ENCePP Guide and expert opinion. General database characteristics were also collated (data lag, updates, types and size). A ranking was assigned across each data element to assess completeness.
Results: Pregnancies can be linked to infant outcomes in all data sources with some variability in completeness. In Germany, mother-baby pairs must be on the same insurance plan. In France, infant outcomes can be studied if mother-baby pairs receive care in the same hospital. The majority of databases contain most of the pregnancy outcomes (e.g. pregnancy loss, preterm birth) using varied measurement approaches and with different levels of sensitivity. The largest variation in data capture is observed in spontaneous abortion where the Nordic registers have the most complete data. Availability of neonate and infant outcomes, and particularly those requiring longer infant follow-up such as neurodevelopmental disorders, varied across countries. In Italy and Spain, infant follow-up depends on how long the baby remains in the regional health system. Capture of exposure to drugs and vaccines in secondary care settings should be investigated for each study given variability of availability across databases.
Conclusions: In the context of regulatory-decision making for drug or vaccine safety in pregnancy, a multi-country approach is often necessary. Given the heterogeneity of European data availability, feasibility anchored in specific research questions should be conducted for final data source selection and justification. When exposure is limited or rare, Australian, Canadian and US data can contribute to contextualizing European data.
