HealthData@EU Pilot - Deliverable 9.1 Use cases report

This deliverable provides an overview of the experiences, challenges, and findings of the five use cases developed within the HealthData@EU Pilot. These use cases were designed to explore the feasibility of reusing health data across multiple European countries for secondary purposes. The report presents a general overview of the use cases and a detailed analysis of each one.

Overview of the Use Cases

• Predicting nationwide cardio-metabolic disease trajectories using machine learning
  • Lead: HDH / University of Helsinki
  • Participating Nodes: University of Helsinki (FI), AHeaD (FR), Copenhagen University (DK), Norwegian Institute of Public Health (NO)
  • Brief Overview: This use case applied machine learning to predict a 5-year cardiovascular risk. It encountered challenges related to the interpretation of data minimization in France and Norway.
• Test use, hospitalisation, and vaccination adherence in vulnerable sub-populations
  • Lead: Sciensano
  • Participating Nodes: Sciensano (BE), Central Denmark Region (DK), Croatian Institute of Public Health (HR), Finnish Institute for Health and Welfare (FI)
  • Brief Overview: This use case analyzed COVID-19 testing, hospitalization, and vaccination rates in vulnerable groups. Issues were encountered in accessing socio-economic data.
• Identifying the risks of coagulation disorders in patients with COVID-19
  • Lead: EMA
  • Participating Nodes: Croatian Institute of Public Health (HR), Danish Health Data Authority (DK), Finnish Institute for Health and Welfare (FI), Health Data Hub (FR), DARWIN EU Network
  • Brief Overview: This use case investigated the venous/arterial thromboembolic risks in COVID-19 patients, using the OMOP CDM for analysis.
• Antimicrobial resistance (AMR) surveillance
  • Lead: ECDC
  • Participating Nodes: Sciensano (BE), Croatian Institute of Public Health (HR), Finnish Institute for Health & Welfare (FI)
  • Brief Overview: This use case tested the EHDS architecture for communicable disease surveillance, focusing on data concordance, legal barriers, and federated data processing.
• Identifying genomic signatures of colorectal cancer 
  • Lead: ELIXIR / BSC
  • Participating Nodes: BBMRI-ERIC, Sciensano (BE), Danish Health Data Authority (DK), Norwegian Institute of Public Health (NO), National Institute of Oncology (HU)
  • Brief Overview: This use case used genomic and clinical data for colorectal cancer analysis but faced significant delays in access, eventually leading to transfer to the GDI project.

Key Challenges and Findings

• Data Access & governance
  • Data minimization interpretation: This posed a challenge in France and Norway, restricting access to some datasets.
  • Socio-economic data restrictions: The Sciensano use case faced significant challenges in accessing this data.
  • Genomic data complexity: The ELIXIR use case encountered major delays and feasibility issues in accessing genomic data.
• Technical & interoperability challenges
  • Heterogeneous data formats: Different nodes used varying data formats (e.g., OMOP CDM vs native formats).
  • Federated data processing limitations: The ECDC use case highlighted scalability concerns related to federated data processing.
• Legal & ethical considerations
  • Varying national regulations: Different regulations impacted data access and timelines.
  • Data privacy and security concerns: Cross-country challenges emerged in ensuring compliance with data privacy and security standards.