The BETTER project has made significant technical and scientific progress in establishing a privacy-preserving federated learning ecosystem for health data. Key achievements include:
Foundation for Cross-Border Data Integration: Development of a common data schema and FAIRification efforts, along with defined legal and ethical data governance frameworks, including pseudonymisation and GDPR compliance. Prototypes of the BETTER Station have been deployed for secure local data access and interoperability testing.
Data Fusion and AI Tool Development: Initial data harmonisation and pre-integration using synthetic datasets has begun, with early integration of clinical and genomics data successfully demonstrated. Development of AI-based data fusion strategies is ongoing. Initial AI pipelines leveraging federated learning have been prototyped across all three use cases using synthetic data.
BETTER Platform Deployment: Core components of the BETTER platform have been implemented and tested, with a central coordination node operational and local BETTER stations deployed at most medical centers. Federated learning capabilities have been successfully tested in a controlled environment.
Clinical Use Case Progress: Detailed implementation plans for all three clinical use cases (Paediatric Intellectual Disability, Inherited Retinal Diseases, Autism Spectrum Disorders) are finalized, with co-creation workshops completed and WGS pipelines operational.
ELSA Integration: An ELSA-aware co-creation framework has been drafted, and tools for value-sensitive design and post-deployment evaluation are under development with social science and legal experts.
Legal Framework Established: A robust legal framework, including Consortium, Infrastructure User, Data Processing, and Project Agreements, has been developed to ensure GDPR compliance for distributed patient data processing.
