In a world in which medicine is benefiting from digital progress and ever-increasing amounts of data are being generated, routine data from medical care must be made available efficiently, securely, and in a way that promotes innovation, made available for medical research, and used to answer medical questions. The NUM-DIZ project builds on the preliminary work of the Medical Informatics Initiative (MII), within which data integration centers (DIC) have been established at most German university hospitals to support data provision and cross-site data integration and analysis. As part of the NUM-DIZ project, the established DICs continuously expand their service portfolio and tap into new data sources.

Project goals

The infrastructure components, datasets, and data usage regulations established at the NUM-DIZ partners' DICs serve, in conjunction with central components from NUM-RDP, the German Research Data Portal for Health, and other central components developed in the MII, to enable Germany-wide data usage projects with the (potential) database of patient data collected in clinical care by 38 NUM-DIZ partners and thus make essential contributions to gaining medical knowledge. With NUM-RDP, NUM-RACOON, and other NUM infrastructure projects, NUM-DIZ will provide a national resource in which high-quality health data, patient partnerships, and research expertise can quickly deliver trustworthy answers that can improve healthcare.

Locally, the DIC will evolve into a comprehensive research infrastructure service provider supporting a variety of medical research scenarios for research at the site, but especially for networked research across sites. Each NUM-DIZ partner is building an efficient technical, personnel, and organizational infrastructure. As a result, infrastructures are available at all German university hospitals to enable networked research with health data.

Challenges and Implementation

One of the major challenges is the secure processing of health data, which, due to its nature, constitutes particularly protected personal data. The infrastructure in NUM-DIZ protects this data through standardized pseudonymization steps, encrypted transmission, and additional measures that protect against intentional or unintentional re-identification of individuals.

The four consortia DIFUTURE, HiGHmed, MIRACUM, and SMITH have developed concepts for secure and interoperable health data management as part of the Medical Informatics Initiative. For this purpose, DICs have been or are being established at each university hospital, which access routine healthcare data in a standardized and harmonized manner.

These technical infrastructures, the trained personnel, and the established data provision, governance, and use processes are now being consolidated and made permanent in NUM-DIZ.

NUM-DIZ is based on five years of fundamental development and harmonization activities within the framework of the MII. During this time, the essential prerequisites for efficient cooperation between decentralized DIC components and central MII infrastructures were created both at the individual university medicine sites with the establishment of the DIC infrastructures and in the cross-consortia harmonization work to be able to use routine data for research across all locations in Germany. The results achieved in the MII working groups include

• A broad-based patient information and consent form (broad consent) for the use of data collected in the care processes of German university hospitals for medical research (accepted by the Conference of Federal and State Data Protection Commissioners).
• The harmonized definition of an MII core data set includes basic modules on patient demographics, case data, diagnoses, procedures, laboratory data, medication data, consent information, and extension modules.
• The establishment of uniform data use and data governance processes throughout Germany (based on uniform data use regulations, uniform templates for data use contracts, and local data use commissions, https://www.medizininformatik-initiative.de/de/nutzungsordnung)
• A participation framework agreement that regulates the legal framework conditions for all stakeholders in the research data infrastructure who are involved in these processes
• A harmonized interface definition between all technical components of the NUM-DIZ partners and the central FDPG.

During the COVID-19 pandemic, the NUM-DIZ partners have supplemented this as part of the NUM-CODEX project with harmonized IT tools for the prospective collection of research data (tailored to the German Corona Consensus Dataset, GECCO) as well as components for the secure transfer of patient data to a central research database (subject to patient consent).

Based on this preliminary work, 24 German university hospitals are connected to the FDPG. They can support cross-site data usage processes based on the MII core dataset base modules.