Privacy-Preserving Workflow for the Cross-Border Federated Analysis of Clinical Data.

Area of Science:

• Health Informatics
• Data Security
• Epidemiology

Background:

• Analyzing sensitive patient data across different jurisdictions poses significant privacy challenges.
• Sharing individual-level data is often impossible due to legal and ethical restrictions.

Purpose of the Study:

• To develop and validate a privacy-preserving federated data analysis workflow.
• To enable secure cross-border analysis of sensitive clinical data without exposing patient-level information.
• To showcase the integration of data harmonization within a federated analysis framework.

Main Methods:

• A federated data analysis workflow was designed using open-source software and DataSHIELD libraries.
• The workflow utilizes the R language for statistical computing and ensures data remains at remote sites.
• Data harmonization techniques were implemented and showcased within the secure workflow.

Main Results:

• The federated workflow successfully performed data analysis on sensitive clinical datasets from the Pan-European SARS-Cov-2 cohort (ORCHESTRA project).
• Results were validated against centralized analysis, confirming equivalent outcomes without compromising patient privacy.
• The study demonstrated the viability of a cross-border federated analysis framework.

Conclusions:

• Federated data analysis is a viable approach for secure, cross-border research involving sensitive health data.
• Maintaining data security, usability, and availability requires careful attention to data harmonization, anonymization, and IT infrastructure.
• This privacy-preserving methodology achieves analytical results comparable to traditional centralized approaches.