Skip to content

1.3. Research questions and reading guide

1.3.1. Core design and research questions

This document elaborates the architecture of a hybrid SPE that can be implemented as part of a nationally covering basic infrastructure for the secondary use of data. The following design and research questions are central to this:

  • Can a unambiguous functional and technical specification be given of a hybrid SPE based on data stations and processing hubs, that is consistent with existing reference architectures?
  • In what way could data stations be standardised to achieve maximum interoperability between different networks (a data station connected to multiple processing hubs)?
  • In what way can data stations be used for data pooling to a central TRE? What options are there for achieving standardisation in this area?
  • Are there gaps or contradictions in the current EHDS approach for secondary use that obstruct the use of a hybrid SPE, particularly in relation to data stations for primary use? If so, what solution directions are there?
  • What is the maturity of the existing reference implementations relative to the design described above:
    • KIK-V as a reference implementation of data stations for federated analysis;
    • PLUGIN/vantage6 as a reference implementation of data stations for i) federated analysis, ii) federated learning and iii) data pooling.

1.3.2. Reading guide

This document is structured as follows. The Process section first describes the basic processes of secondary use, from data request to data use and publications, using the Use Case 3.0 modelling technique. The relevant concepts within the architecture of a data space are explained, in order to provide a clear picture of the context of a data station and the processing hub. Where necessary, reference is made to an Archimate model of the data space. The Information chapter addresses semantic and syntactic interoperability as a prerequisite for the (re)use of health data. The Application section goes into greater detail on the various application components of a hybrid SPE, including the data station and processing hub per se. We follow the FAIR hourglass model as the definition of the layers. The specifications from TEHDAS2 are used as a starting point and further elaborated. The Infrastructure section then describes the physical technology and standards relevant to realising the application components.

Having explained the architecture from various perspectives, we turn in Implementations to existing implementations of data stations and federated processing hubs. By analysing in particular KIK-V and PLUGIN as reference implementations, we aim to critically reflect on the proposed specification of a data station for secondary use. Finally, Discussion sets out the key findings in relation to the research questions and makes suggestions for further development in the future.

Comments