DATA ARCHITECTURE
A Federated Network of Data Integration Centres
The Clinnova data architecture is anchored by local Data Integration Centres (DICs), which provide a robust, interoperable, and scalable data infrastructure built on open-source components. DICs play a vital role in collecting, aggregating, cleaning, and converting isolated health data into interoperable datasets, making them ready for analysis. By offering access to clinical, omics, and patient-generated data, these DICs lay the groundwork for precision medicine within Clinnova.
To develop advanced automated analysis models, we use sophisticated AI algorithms trained on real patient data. The success of this training depends on gathering a diverse and representative dataset across the entire Clinnova network. However, maintaining patient privacy is paramount, which is why Clinnova employs federated learning.
In this federated design, each DIC retains full control over the data of local study participants, adhering to the principle of data protection by design and default. Instead of sharing patient data across institutions, each partner site trains models locally. Only the trained models, not the data, are shared and combined with those from other institutions. This approach enables collective learning while safeguarding patient privacy.
The Clinnova DICs are also aligned with national and European initiatives, such as the German Medical Informatics Initiative (MII), Gaia-X, and the European Health Data Space, ensuring that this infrastructure not only serves the project’s immediate goals but also contributes to broader efforts in digital health.
