Experiences of Transforming a Complex Nephrologic Care and Research Database into i2b2 Using the IDRT Tools

Area of Science:

• Biomedical Informatics
• Health Data Science
• Clinical Data Management

Background:

• Secondary use of Electronic Health Record (EHR) data is crucial for disease research.
• The informatics for integrating biology and the bedside (i2b2) platform facilitates patient cohort selection.
• Integrating diverse clinical data into i2b2 requires robust data extraction and transformation processes.

Purpose of the Study:

• To evaluate the feasibility of adapting data from TBase, a nephrology transplantation documentation system, into the i2b2 schema.
• To assess the suitability of the Integrated Data Repository Toolkit (IDRT) for Extract, Transform, and Load (ETL) processes involving complex EHR data.
• To determine if data relationships from a relational schema (TBase) can be accurately represented in the entity-attribute-value (EAV) model of i2b2.

Main Methods:

• Utilized TBase as the source system for nephrologic transplantation data.
• Employed the Integrated Data Repository Toolkit (IDRT) for the Extract, Transform, and Load (ETL) process.
• Identified and analyzed relevant entities within the TBase schema for cohort identification purposes.
• Evaluated data structure transformations required for mapping TBase to the i2b2 EAV schema.

Main Results:

• Data entities from TBase exhibited varied structures necessitating tailored ETL handling.
• The IDRT demonstrated limitations with large datasets and specific modern EHR data structures.
• The adapted TBase data in i2b2 successfully supported common clinical cohort identification queries.

Conclusions:

• Modeling relational data into the i2b2 EAV schema is achievable for clinical research.
• While IDRT has limitations, the ETL approach proved effective for TBase data integration into i2b2.
• This methodology enables efficient patient cohort identification from EHR data for secondary use in research.