Semantic Harmonization Workflow of Health Data: Application in the Post-COVID Network Netherlands

Semantic interoperability in multi-source health datasets is a key barrier to data integration and reuse in biomedical research. FAIR principles promote standardization, but operationalizing them for heterogeneous clinical and patient-reported data remains challenging. We introduced a structured workflow that yields metadata-rich codebooks to support semantic interoperability. It integrates manual curation, expert-driven semantic annotation using biomedical ontologies and contextual suffixes for measurement nuances. The workflow was applied to the Post-COVID Network Netherlands (PCNN) dataset. In a representative PCNN subset, we harmonized variables across domains like demographics, COVID-19 history, and questionnaires. Mappings are designed for expert review to ensure accuracy and reproducibility.