OBJECTIVES:

To assess the challenges and opportunities for observational multi-country studies based on electronic health records (EHRs).

METHODS:

Desk research on publications from selected networks at the forefront of multi-country collaborative research studies (EHDEN, https://www.ehden.eu/; EMIF, http://www.emif.eu/; OHDSI, https://www.ohdsi.org/; SIGMA, https://sigmaconsortium.eu/).

RESULTS:

Multi-country studies allow the generalisability of real-world findings beyond specific geographic borders, permit the evaluation of heterogeneity and provide the ability to replicate findings across countries of interest. However, multi-country studies involve challenges: EHRs in different countries are structured in multiple ways, employ different coding terminologies (e.g., ICD-9/10, SNOMED CT, ICPC) and reflect inherently different healthcare systems organised to serve specific populations. Together, these factors generate considerable variation and limit the interpretability of heterogeneous findings. However, standardisation techniques have emerged to address these challenges, including the adoption of standards for study development (PASS protocol templates, ENcEPP registration), the use of a common data model for data extraction and analysis (e.g., EMIF Catalogue, OMOP Common Data Model) and statistical methods for data comparison (negative case-control) and for data pooling (e.g., meta-analytic techniques). Using standardisation techniques can allow for robust multi-country studies which provide comprehensive and timely information. Nevertheless, some challenges remain including the potential loss of information at the local level and heterogeneity across multiple countries.

CONCLUSIONS:

Researchers seeking accurate answers to a research question need to consider the potential benefits and drawbacks of observational multi-country studies based on EHRs. As industry professionals and researchers, we must advocate for the use of standardised methodology for multi-country studies to ensure the reporting of reliable, valid and generalisable findings.