JMIR Publications

ABSTRACT

Background:

Aligned with the European Health Data Space (EHDS), the HealthDCAT-AP metadata model has been designed to support the semantic description of health datasets and to make them findable and interoperable across EU-level catalogues. However, context-specific requirements, driven by disease areas, data models, should be incorporated on a per-project basis.

Objective:

As part of the IDERHA project, which is establishing a federated data space for the secondary use of heterogeneous health data with an initial focus on lung cancer, we aim to tailor a custom metadata model based on the HealthDCAT-AP specifications to address these needs.

Methods:

We derived IDERHA metadata profile based on HealthDCAT-AP specification by: (i) adopting all mandatory classes and properties (Dataset, Catalog, Distribution, Data Service, Agent) and selecting a preferred subset of recommended and optional properties based on working-group input and compatibility with current ELIXIR Luxembourg data-catalog metadata specifications and data holders’ requirements; (ii) modeling legal bases and conditions of use using DPV and mapping them to DISH concepts; (iii) extending the profile with OMOP CDM–derived statistics and data-quality indicators; and (iv) adding clinical and imaging measures. The metadata model is represented with a JSON Schema, and Protégé 5.6 was used to develop a semantic metadata model incorporating the ontologies.

Results:

The resulting model organizes metadata into five categories: (1) HealthDCAT-AP-derived dataset descriptors; (2) OMOP statistics and data quality; (3) data privacy/conditions of use (DPV-compatible); (4) clinical statistics; and (5) study metadata. Implementation is provided via a JSON Schema for provider adoption and an OWL ontology that imports DCAT-AP v3, DPV, and DQV, and adds IDERHA-specific classes and data properties. This custom profile enables coherent, machine-readable descriptions across participating IDERHA data providers and supports publishing and further discoverability in the IDERHA data catalog.

Conclusions:

The proposed custom metadata model is aligned with EHDS goals by ensuring interoperability with other catalogues and improving dataset discoverability through the IDERHA community data catalog. The approach and metadata profile can be reused for other diseases or domain-specific community data infrastructures.