Federated electronic data capture (fEDC): Architecture and prototype

Area of Science:

• Clinical Informatics
• Health Data Management
• Collaborative Research Technologies

Background:

• Structured documentation is crucial in clinical research and patient care, often using electronic case report forms (eCRF).
• Integrating eCRF systems with electronic health records (EHR) can prevent duplicate data entry.
• Collaborative studies face challenges integrating eCRF with local EHR systems due to security and data protection restrictions.

Purpose of the Study:

• To propose a novel federated electronic data capture (fEDC) system architecture.
• To address limitations in integrating joint internet-based eCRF systems with local EHR systems in multi-institutional collaborations.
• To ensure reliable form definitions, EHR integration, local data control, and adherence to data sharing principles.

Main Methods:

• Developed a federated electronic data capture (fEDC) architecture.
• Ensured form definitions are shared via the Operational Data Model (ODM) format.
• Enabled sites to run their own fEDC instances compliant with local standards.
• Validated the fEDC architecture with an open-source prototype at five German university hospitals.

Main Results:

• The fEDC architecture facilitates integration with local electronic health record systems.
• Metadata sharing using the Operational Data Model (ODM) ensures consistent form definitions across sites.
• Patient data remains under local control until explicit transfer for analysis, enhancing privacy.
• A working open-source prototype was successfully validated at five university hospitals.

Conclusions:

• The fEDC architecture offers a novel approach to improve collaborative data capture in clinical research.
• Reduced data entry efforts and increased data quality are key benefits.
• Lowered barriers for EHR integration and enhanced patient privacy are significant advantages.
• The system supports widely accepted data sharing principles for both data models and captured data.