Lessons learned from evaluating the NFDI4Health metadata schema for health studies using the Research Data Alliance's FAIR data maturity model

ObjectivesThe German National Research Data Infrastructure for Personal Health Data (NFDI4Health) has developed a metadata schema (MDS) for harmonizing health study descriptions, along with a platform for content discovery. This work evaluates how well the MDS promotes FAIR data principles (Findability, Accessibility, Interoperability, Reusability) and serves as a proxy for assessing study FAIRness. Materials and MethodsUsing the Research Data Alliances FAIR data maturity model, we assessed the scope of evaluable FAIRness indicators for NFDI4Healths MDS (version 3.3). Results29 of 41 FAIRness indicators can be evaluated (Findability: 7/7, Interoperability: 7/12, Accessibility: 7/12, Reusability: 8/10). The remaining indicators relate to the research datas format and access procedures. DiscussionThe MDS provides a viable basis for evaluating study FAIRness. To enable full indicator coverage, additional metadata elements should be incorporated. ConclusionNFDI4Healths MDS effectively promotes FAIR-compliant study descriptions, although adjustments could strengthen its utility for assessing research data FAIRness. Lay summaryMaking research data easily Findable, Accessible, Interoperable, and Reusable (FAIR) is crucial for accelerating discoveries in health research. The German National Research Data Infrastructure for Personal Health Data (NFDI4Health) is an initiative and network that develops services, tools, and training programs to promote and support the FAIR sharing of health research data in Germany. As part of this effort, NFDI4Health has created a standardized framework, referred to as metadata schema (MDS), to systematically collect and structure information about health studies. The collected information can be accessed through a dedicated platform provided by NFDI4Health. In light of ongoing developments of the platform, it is important to assess whether the MDS content may serve as a proxy for assessing the FAIRness of the included studies. Therefore, we assessed the scope of evaluable FAIRness indicators based on the FAIR data maturity model by the Research Data Alliance. Our assessment shows that 29 of 41 FAIRness indicators can currently be evaluated. The remaining indicators relate to the data format and access procedures. However, they could be easily covered by introducing new elements to the MDS and encouraging data holders to provide the required information. Overall, the MDS effectively supports FAIR-compliant descriptions of health studies.