Applying Fiduciary Concepts to Health RDM - Mapping and Evaluation

Background By applying "data trust" or "data fiduciary" models to the research use of health data, ethical fiduciary concepts gain importance in the health research sector. Back in 2017 the German Ethics Council already suggested data fiduciaries as intermediaries between data subject and data user to foster trust and to prevent misuse in the field of data use for biomedical research [1]. Since then, the implementation of data fiduciaries has been recommended multiple times by experts and ethics councils [2], [3], [4] and is again mentioned in the coalition agreement of the upcoming German government to ensure trust in data management and high data quality [5]. Considering the essential elements of a fiduciary relationship is necessary in the design of potential data fiduciary models (e.g. German research data centers) and data initiatives (e.g. MII, GHGA, 1+million genomes). Challenges Despite multiple recommendations for implementing data fiduciaries, there is no mutual understanding of the term and the underlying concept. The term is currently used indiscriminately for complex proxy constructions as well as simple IT applications from the field of Personal Information Management Systems (PIMS) to support data subjects in managing their own data. An analysis and mapping of concepts and existing models show that data fiduciaries in the field of health data can be summarized under the following six functional categories: - balancing interests as a neutral intermediary, - storing and organizing data, - deciding on data access, - ensuring data protection, - analyzing data, - ensuring economic participation of data subjects. The various and different usages of the fiduciary concept are problematic, as some models do not meet the criteria for fiduciary relationships. Therefore, data subjects' expectations generated by the trust-provoking labelling "fiduciary" might be disappointed and models labelled as data fiduciaries suffer from legal and ethical uncertainty regarding the extent of their duties. Therefore, the foundation of the concept of data fiduciaries and the identification and elimination of misnomer is essential to successfully implement the concept of data fiduciaries in health research. This is the focus of our DFG founded project "DAta FIDUciaries in MEdical Research: ethical and legal foundations and implementations (DAFIDUMER)" [6]. Approach and Discussion The fiduciary concept is characterized by the following attributes [7]: A fiduciary is entrusted with discretionary power over the practical interests of another person (the beneficiary). The fiduciary is typically characterized by special expertise, which they use for the benefit of the beneficiary. The relationship between fiduciary and beneficiary is characterized by vulnerability and dependency of the beneficiary and based on a special trust. The fiduciary is bound by special duties – the fiduciary duties. Against the backdrop of this theory, we evaluate whether the above-mentioned categories display these fiduciary qualities. Conclusion In the current discussion, models are labelled as data fiduciaries that do not meet the corresponding requirements. This includes simple repositories or storage services, pseudonymization services and some PIMS.