Implementing partnership-driven clinical federated electronic health record data sharing networks

Area of Science:

• Health Informatics
• Data Science
• Software Engineering

Background:

• Federated data sharing architectures are crucial for network-based research.
• Establishing these architectures involves multiple data owners and semantic alignment.
• End-user focus and system utility are key considerations.

Purpose of the Study:

• To describe critical factors, shared outcomes, and results from two federated data sharing case examples.
• To detail the findings from developing pilot architectures for translational health sciences research.
• To identify essential elements for successful federated data sharing system development.

Main Methods:

• Utilized a spiral model for implementation and evaluation across two pilot architectures.
• Employed cross-collaboration between two network development teams.
• Structured development through iterative cycles and knowledge sharing.

Main Results:

• The spiral model and iterative cycles were effective in meeting network design goals.
• Building a trusted environment for data sharing architectures was resource-intensive.
• Adaptive use cases were essential for defining and testing system utility.

Conclusions:

• An iterative, cyclical development model fosters trust with data partners.
• This approach supports the refinement of federated data sharing architecture design.
• Measurable success was achieved in developing new federated data sharing architectures.