Enhancing Data Protection via Auditable Informational Separation of Powers Between Workflow Engine Based Agents:...

Area of Science:

• Health Informatics
• Data Privacy Engineering
• Medical Research

Background:

• German best practice standards mandate strict data separation for secondary patient data use.
• Ensuring identifying data (IDAT), pseudonyms (PSN), and medical data (MDAT) are never simultaneously accessible is crucial.
• Existing systems may struggle to meet these stringent data protection requirements.

Purpose of the Study:

• To describe a software solution that adheres to German best practice standards for the secondary use of patient data.
• To implement a system ensuring informational separation of powers for data provisioning.
• To facilitate compliant and secure access to pseudonymized patient data for research.

Main Methods:

• A distributed agent architecture involving a clinical domain agent (CDA), trusted third party agent (TTA), and research domain agent (RDA).
• Utilizing an off-the-shelf workflow engine for distributed CDA and RDA operations.
• Integrating the gPAS framework within the TTA for pseudonym generation and persistence.
• Implementing agent interactions via secured REST-APIs.

Main Results:

• The system successfully implemented dynamic interaction between three software agents.
• Deployment across three university hospitals demonstrated a seamless rollout.
• The workflow engine facilitated auditability of data transfer and pseudonymization with minimal effort.
• The solution effectively met technical and organizational data protection requirements.

Conclusions:

• A distributed agent architecture, powered by workflow engine technology, is an efficient method for compliant patient data provisioning.
• The described solution successfully meets the rigorous requirements for secondary data use in Germany.
• This approach enhances data security and privacy while enabling valuable medical research.