Support research data management in health studies: library and research working together in a data steward pilot

This pilot project within a medicine-oriented consortium of the National Research Data Infrastructure (NFDI4Health) aims to explore the impact of two library-based data stewards working closely together with people who conduct(ed) health studies and to serve as a positive practical example for other libraries or infrastructure institutions wishing to establish data stewardship together with discipline specific research projects. It also aims to explore the relationship between generic research data management (RDM) knowledge at library or infrastructure level as well as for various disciplines within a research institution, and to then determine how this knowledge can be combined to benefit both[1]. Regarding the tasks and embedding of data stewards, the so-called embedded or discipline-specific data steward comes into play. In addition to general issues of research data management, these individuals are primarily concerned with the specific tasks associated with their particular discipline. This involves managing medical research data, primarily from clinical and epidemiological studies. The profiles of a discipline-specific data steward and an infrastructure-related data steward are based on the profiles developed as part of the ‘DataStew’ project, in which five profiles for data stewardship were derived based on areas of expertise and training opportunities[2]. The pilot project describes the initial situation, the roles of data stewards on both infrastructure and research levels, the goals of collaboration between the two, and the approach taken. The experiences gleaned by the two libraries implementing a data steward for epidemiological and a data steward for clinical studies will be illustrated and the main features of the two approaches will be given. The initial situation is characterised by the different expertise of the infrastructure and discipline specific data stewards: the expertise in the libraries comprises basic research data management concepts like the FAIR data principles or metadata. Further, they are involved in the development of standards within the NFDI consortium, e.g. a publication guideline, a metadata schema for data publication or templates for data management plans. Discipline-specific data stewards from research, on the other hand, have less experience with generic RDM, but bring discipline-specific expertise, either from epidemiological or clinical research, when it comes to methods and workflows. They assessed the RDM in place for the studies conducted by their research institute, highlighting which RDM strategy was already in place and which other parts of RDM were missing. Additionally, their role involves testing the established standards and services, as well as soliciting feedback from research and the broader research community for the relevant study. This pilot aims to combine the knowledge of the infrastructure provider and the research party. The initial aim was for both to familiarise themselves with and gain a thorough understanding of each other's perspective. The infrastructure data steward would collaborate with the researcher or discipline-specific data steward to design an infrastructure that caters to the needs of the researchers at their institution. The discipline-specific data steward, being well-versed with the specific requirements of the researchers, offers valuable input on the necessary infrastructure. The two data stewards then jointly develop training formats covering basic RDM concepts and consortium specific services, such as a central search hub (CSH). The aim was to educate researchers on RDM in general, but especially on data publication, e.g. equip them with the knowledge needed to input their studies (= study metadata) into the CSH. An important result of the pilot involving two use cases of different study types was that it is essential to find the right contact person involved with research. While in the epidemiological study it was clear from the outset that a person within the study would be selected as a counterpart to the data steward, it proved difficult to find the right contact person (counterpart) for the partner „clinical trial center”. The main reason for this is that a clinical trial center is a service provider or infrastructure partner like the library, although with different service portfolios. As an overall result it is evident that both data steward profiles are necessary, because synergy effects are created through their cooperation. The extent to which fundamental knowledge, can be merged into one individual in the future relies on the availability of training opportunities. One option could be to provide training for researchers in the role of a data steward on the fundamentals of RDM. This could be accomplished through a certification course akin to the one presently available in North Rhine-Westphalia. It seems more difficult to train a generic data steward in the basics of a discipline, as this would require at least a bachelor's degree in that discipline. Liaising between an infrastructure data steward and a research data steward therefore seems appropriate given the current situation. [1] Fürst, J., Dierkes, J., Lindstädt, B., Perrar, I., & Shutsko, A. (2024). Demonstrator of the Data Stewardship Service. Zenodo. https://doi.org/10.4126/FRL01-006483497 [2] Eva Seidlmayer, Fabian Hoffmann, Jens Dierkes, Birte Lindstädt, Ralf Depping, Konrad U. Förstner: Forschung unterstützen - Empfehlungen für Data Stewardship an akademischen Forschungsinstitutionen : Ergebnisse des Projektes DataStew, 2023, DOI 10.4126/FRL01-006441397, page 72 ff