Improving biomedical data interoperability with semantic web technology

Health data is essential for advancing clinical research and improving patient care. However, poor data interoperability limits the reuse and integration of existing data, which is more important in the context of rare diseases. This thesis addresses these challenges by applying the FAIR Data Principles—ensuring data is Findable, Accessible, Interoperable, and Reusable—and utilizing Semantic Web technologies to improve data quality and harmonization, and hence interoperability.
The work first identifies 98 challenges faced by European rare-disease registries in implementing the FAIR principles, grouped into training, legal, modelling, implementation, and community-related issues. Solutions emphasize the role of FAIR data stewards and the need for standardized practices. An automated method to assess the quality of RDF data is then introduced, focusing on Resolvability, Parsability, and Consistency. This method, applied to healthcare data models and rare-disease data, uncovers common quality issues and offers practical recommendations to enhance the utility of linked data.
To improve semantic interoperability, the thesis proposes a novel harmonization method using SSSOM and RDF to align data elements across five major health data standards: HL7 FHIR, OMOP, CDISC, Phenopackets, and openEHR. Demonstrated through real-world use cases, this approach facilitates cross-standard data integration and querying, supporting more effective data reuse.
The thesis concludes that achieving interoperability requires coordinated technical, semantic, and organizational efforts. By combining FAIR stewardship, automated quality assurance, and semantic harmonization, this work contributes to building more interoperable, high-quality biomedical data infrastructures—ultimately enhancing research and healthcare outcomes.