gms | German Medical Science

Introduction: The collaborative efforts of the German Medical Informatics Initiative (MII) [1] [BS1] especially on methods of informed consent and information model standardization [1] pave the way on making healthcare data findable and accessible for research [2]. One of the core infrastructural assets though are the (Medical) Data Integration Centers (MeDICs) in nearly all German University Medical Centers [3], [4]. However, their architectural approaches and implementations differ strongly between the different sites, due to different starting situations, site-specific conditions and different consortium agreements. While sites with strong medical informatics expertise may mainly reuse site-specific workflows and existing workflows, new sites are building their MeDIC from scratch. Sharing and discussion on the different architectural design decisions and implementation strategies may in particular help sites that are at the beginning of the MeDIC setup, but also may give new insights for established MeDICs.

Methods: In this workshop, we discuss the MeDIC design of a first and a second generation HiGHmed MeDIC [5] in comparison to a new site benefiting from previously achieved workflows and components. Furthermore, we include a different architecture approach within the MIRACUM consortium [6].

All contributors follow the same presentation scheme (i) ETL approach for connecting data sources, (ii) data storage approach, (iii) normalization approach, (iv) the discoverability approach and (v) what we [the MeDIC] would do differently next time.

Results: The workshop will give insight into architectural design decisions, the implementation strategies and their reasoning. We hypothesize that local organization models (e.g. closer to routine-IT or closer to research institute) pre-existing re-usable infrastructures elements like a clinical data warehouse and strong presence of Medical Informatics departments have a major impact, on the internal processes. On the normalization level, we see consortia-specific infrastructure elements. Due to the efforts on the MII information model started in 2017 [1], the presentation of the MeDIC data to the outside world is uniform for each MeDIC (s. iv).

Discussion: The initial explorative approach for the MII offers a broad spectrum of different architectural solutions to prepare locally structured healthcare data accessible for research. Within the workshop, we try to identify additional technical or educational components to facilitate the onboarding for Data Integration Centers to reach the MII goals effectively and efficiently.

Conclusion and Outlook: The degrees of freedom for the specific MeDIC enabled well-adapted site specific solutions. In the second phase of the MII the need for a common information model enabled interoperability solutions beyond specific architectural questions in order to make data findable and sharable within the consortia and the outside world. We hope that this workshop will help to build stable and sustainable MeDICs for nationwide sharing of clinical data.

Acknowledgement: This project was funded by the BMBF Medical Informatics Initiative (01ZZ1802A-Z, AB).

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.