Federating European REgistries for Stroke: Principles, Data Model and Future Perspectives for a Federated Analysis of Multiple Prospective Stroke Registries.

backgroundAcute ischemic stroke is a leading cause of death and disability. Despite strong evidence supporting reperfusion therapies and Stroke Unit care, access and quality of stroke services remain heterogeneous across Europe. Although national stroke registries provide valuable real-world data, fragmentation, limited interoperability, and data protection constraints have restricted multinational analyses and benchmarking.

methodsThe Federating European REgistries for Stroke (FERES) initiative establishes a GDPR-compliant federated framework for secondary use of stroke registry data. FERES harmonizes heterogeneous datasets through a stroke-specific Common Data Elements (CDE) model and performs analyses locally within each registry using the Medical Informatics Platform, sharing only aggregated, non-identifiable outputs. To validate the framework, a predefined showcase analysis comparing anterior versus posterior circulation acute ischemic stroke was executed in both centralized and federated modes using an identical harmonized dataset.

resultsFERES connected five national registries from Austria, Greece, Ireland, Italy, and Switzerland within a unified federated infrastructure. At the time of analysis, 149,772 patient events from two registries were accessible for federated querying, with three additional registries technically integrated and in advanced onboarding. The harmonized ontology comprised 945 standardized variables spanning the stroke care pathway. Federated execution reproduced centralized pooled-data results across descriptive statistics, hypothesis testing, effect sizes, and multivariable regression models with only minimal numerical discrepancies.

conclusionsFERES demonstrates that large-scale, multinational stroke research and benchmarking can be conducted in Europe using a privacy-preserving federated approach, providing a scalable foundation for cross-border real-world evidence generation and quality improvement.