Metadata-driven Clinical Data Loading into i2b2 for Clinical and Translational Science Institutes

Area of Science:

• Clinical Informatics
• Data Management
• Biomedical Research

Background:

• Clinical and Translational Science Award (CTSA) recipients require research data marts from clinical data warehouses.
• Existing extract, transform, and load (ETL) processes are often duplicative and difficult to maintain.
• i2b2 and SHRINE technologies are commonly used in research data networks.

Purpose of the Study:

• To develop a novel, metadata-driven ETL process for creating customized research data marts.
• To reduce the burden of maintaining separate ETL logic for diverse data mart requirements.
• To enhance the flexibility and efficiency of data integration within research networks.

Main Methods:

• Developed a metadata-driven ETL framework integrated with i2b2's ontology database schema.
• Extended the open-source Eureka! Clinical Analytics software to support the new ETL capability.
• Configured the system for different data marts using separate configuration files.

Main Results:

• Successfully created i2b2 data marts for multiple research projects.
• Loaded over 147 million facts for 1.2 million patients into the Accrual for Clinical Trials (ACT) network.
• Demonstrated the system's ability to handle large-scale data integration efficiently.

Conclusions:

• The metadata-driven ETL process offers a flexible and efficient solution for populating research data marts.
• This approach streamlines data integration for CTSA programs utilizing i2b2 and SHRINE.
• The Eureka! software extension provides a scalable tool for managing complex research data needs.