Development and validation of the SickKids Enterprise-wide Data in Azure Repository (SEDAR)

Area of Science:

• Health Informatics
• Data Management
• Cloud Computing

Background:

• Electronic health record (EHR) data presents challenges for direct utilization due to complexity and varied user needs.
• The Hospital for Sick Children (SickKids) sought to enhance EHR data accessibility and usability.

Purpose of the Study:

• To describe the development of sequentially transformed schemas for electronic health record (EHR) data utilization.
• To enable data access for multiple user types through a cloud-based repository.

Main Methods:

• Utilized Microsoft Azure for the SickKids Enterprise-wide Data in Azure Repository (SEDAR).
• Developed three sequential schemas: Filtered Schema, Curated Schema, and SK-OMOP Schema.
• Implemented nightly data updates and validation processes.

Main Results:

• Successfully transferred and updated Epic Clarity data in Azure.
• Achieved improved data structure and query performance with the Curated Schema.
• SK-OMOP Schema facilitated research and machine learning with 80% concept mapping coverage.

Conclusions:

• The sequential schema approach effectively addresses diverse EHR data access requirements.
• Further validation at other institutions is recommended to assess generalizability.