Background: Real-world data (RWD) from different countries are increasingly used to support regulatory, health technology assessment (HTA), and population-level evidence generation. However, cross-country analyses are challenged by differences in data provenance, healthcare systems, coding practices, completeness, and clinical workflows. The Observational Medical Outcomes Partnership (OMOP) common data model (CDM) is widely used to harmonise heterogeneous RWD sources, but its ability to improve comparability of downstream epidemiological analyses relative to native source data across countries requires empirical evaluation. Methods: We examined RWD from Denmark, Finland and Portugal in their ability to capture epidemiology of female breast cancer (BC) and amyotrophic lateral sclerosis (ALS), exemplifying, respectively, a common disease with established treatment modalities and high survival and a rare fatal disease with scarce treatment options. To enable head-to-head comparison on a semantic level, data were mapped to the OMOP CDM. Data in the native format were used for comparison. In a downstream analysis, we examined disease epidemiology, patient characteristics, treatment, and survival.
Background: Real-world data (RWD) from different countries are increasingly used to support regulatory, health technology assessment (HTA), and population-level evidence generation. However, cross-country analyses are challenged by differences in data provenance, healthcare systems, coding practices, completeness, and clinical workflows. The Observational Medical Outcomes Partnership (OMOP) common data model (CDM) is widely used to harmonise heterogeneous RWD sources, but its ability to improve comparability of downstream epidemiological analyses relative to native source data across countries requires empirical evaluation. Methods: We examined RWD from Denmark, Finland and Portugal in their ability to capture epidemiology of female breast cancer (BC) and amyotrophic lateral sclerosis (ALS), exemplifying, respectively, a common disease with established treatment modalities and high survival and a rare fatal disease with scarce treatment options. To enable head-to-head comparison on a semantic level, data were mapped to the OMOP CDM. Data in the native format were used for comparison. In a downstream analysis, we examined disease epidemiology, patient characteristics, treatment, and survival. Results: OMOP conversion enabled a common analytical framework across countries and supported semantically aligned comparisons of key epidemiological and clinical variables. However, cross-country comparability was influenced by differences in data provenance, population coverage, coding practices, availability of clinical details, treatment capture, and healthcare-system-specific workflows. Iterative comparison with native data and external clinical evidence was necessary to identify mapping issues, assess information loss, and ensure high semantic fidelity of the converted data. Overall, OMOP-based estimates were highly consistent with native-data analyses and existing clinical expectations, but residual discrepancies reflected both source-data heterogeneity and decisions in the Extract, Transform, Load (ETL) workflow design. Conclusions: OMOP CDM conversion facilitates semantically meaningful cross-country analyses of RWD by mapping heterogeneous source data to a common structure and standardised vocabularies. However, CDM conversion does not eliminate heterogeneity in the underlying data-generating processes and cannot substitute for study-specific data quality and fitness-for-purpose assessment. Robust use of harmonised RWD for regulatory, HTA, or population-level evidence generation requires iterative benchmarking against native data, clinical expertise, and data-science expertise to support valid interpretation across countries.