The Cosmos Collaborative: A Vendor-Facilitated Electronic Health Record Data Aggregation Platform

Area of Science:

• Health Informatics
• Data Science in Healthcare
• Learning Healthcare Systems

Background:

• Learning healthcare systems leverage routinely collected data to generate evidence for practice improvement.
• Aggregating electronic health record (EHR) data from multiple institutions offers greater empowerment than single-institution data.
• Cosmos is a novel tool designed for cross-institutional data aggregation facilitated by a single EHR vendor.

Purpose of the Study:

• To describe the Cosmos initiative for multi-institutional EHR data aggregation.
• To illustrate the potential utility of Cosmos through practical use cases.
• To highlight Cosmos as a facilitator for learning healthcare systems.

Main Methods:

• Cosmos leverages pre-existing agreements, health information exchange networks, and data standards for rapid scaling.
• Data are centrally stored as a limited dataset to maintain privacy.
• A customer-facing query tool employs result limitations to prevent patient re-identification.

Main Results:

• Within two years, Cosmos aggregated EHR data from over 60 million patients.
• Demonstrated utility in chronic disease surveillance (asthma, obesity), syndromic surveillance (influenza, COVID-19), and immunization adherence.
• Showcased application in adverse event reporting and health services research, such as antibiotic usage patterns.

Conclusions:

• Cosmos offers a low-barrier entry for accumulating multi-institutional EHR data, supporting research and quality improvement in learning healthcare systems.
• The platform facilitates within-vendor data standardization and aggregation, serving as a foundation for broader interoperability.
• Identified limitations include vendor specificity, an 'all or none' contribution model, and institutional control over data queries.