Web services for data warehouses: OMOP and PCORnet on i2b2

Area of Science:

• Health Informatics
• Biomedical Data Science

Background:

• Healthcare organizations often struggle with supporting the same data across multiple research data models.
• This fragmentation hinders the adoption of valuable research tools and projects.
• A need exists for tools and projects to be independent of the underlying data model.

Purpose of the Study:

• To introduce a reusable application programming interface (API) for healthcare research.
• To demonstrate the adaptability of the i2b2 API for supporting diverse patient-centric data models.
• To enable research tools to function independently of specific data models.

Main Methods:

• Developed a methodology to extend the i2b2 API for querying additional data models.
• Utilized i2b2's multi-fact-table querying feature.
• Created data-model-specific i2b2 ontologies and mapped them to query non-standard table structures.

Main Results:

• Successfully implemented and validated the methodology to query the OMOP and PCORnet data models using the i2b2 API.
• Modeled and queried ancillary data model columns as i2b2 modifiers.
• Demonstrated the capability of i2b2's REST API to query multiple healthcare data models.

Conclusions:

• The i2b2 API can be adapted to query diverse healthcare data models, including OMOP and PCORnet.
• This approach separates data models from software tooling, enhancing flexibility in healthcare research.
• The methodology is available as open-source Docker images and on the i2b2 community wiki.