Leveraging Multi-National Observational Study in Post-Marketing Safety Assessment: Challenges and Strategies

The use of multi-national observational study in post-marketing safety assessment has been rising in recent years in parallel with the rapid development and adoption of electronic healthcare data (e.g., administrative claims, electronic health records) and novel statistical analysis methods to handle these data. Secondary use of routinely collected electronic health information has long been available to conduct pharmacoepidemiologic studies using data from millions of patients. Certain observational studies or surveillance activities, especially those investigating rare exposure or outcome or those designed to study specific patient subgroups (e.g., elderly, pediatric) or newly approved medical products, necessitate a multi-national approach. Other instances utilizing such study design include but not limited to (1) postmarketing study requested by multiple regulatory authorities; (2) multiple data systems chosen to complement each other (e.g., databases with long-term clinical outcome data combined with another that includes lab and radiology findings to allow case adjudication and/or algorithm validation); (3) multiple data sources needed to verify and replicate study findings. In this article, we share examples of multi-national postmarketing studies and discuss key pitfalls related to the design and analysis of such studies as well as strategies to mitigate biases.