Building a diabetes screening population data repository using electronic medical records

Area of Science:

• Health Informatics
• Population Health Management
• Data Science in Healthcare

Background:

• Rapid advancements in information technology have complicated clinical and population health data management.
• Increasing complexity of electronic medical records (EMRs) hinders researchers' ability to access and analyze critical information.
• Effective data management is crucial for advancing health services research and improving healthcare quality.

Purpose of the Study:

• To introduce an outpatient encounter data framework for constructing EMR-based population data repositories.
• To facilitate efficient data access, extraction, and analysis for diabetes screening research.
• To provide a flexible framework applicable to various health service research and quality improvement initiatives.

Main Methods:

• Developed a hybrid data structure combining entity-attribute-value, dimensional, and relational models.
• Implemented a six-layer information processing strategy for EMR data extraction and transformation.
• Ensured compliance with Health Insurance Portability and Accountability Act (HIPAA) and institutional review board (IRB) regulations.

Main Results:

• The framework preserves atomic information transparently and meaningfully for researchers and practitioners.
• Achieved performance levels comparable to conventional data warehouse models.
• The data structure supports compliance with privacy and ethical research standards.

Conclusions:

• The outpatient encounter data framework effectively supports EMR-based population data repositories for research.
• The design is adaptable for diverse health service research, including but not limited to diabetes screening.
• This framework can be a valuable tool for organizations aiming to enhance healthcare quality and efficiency, aligning with 'meaningful use' objectives.