Advancing Real-World Evidence Generation: Growth and Lessons Learned from the Early Years of Cosmos

Area of Science:

• Healthcare Informatics
• Clinical Research Data Platforms
• Health Data Science

Background:

• Cosmos is a vendor-facilitated platform developed with healthcare systems utilizing Epic Systems Corporation electronic health records (EHRs).
• It enables participating organizations to advance research and implement insights at the point of care.
• This manuscript details the platform's infrastructure evolution, data growth, and usage patterns.

Purpose of the Study:

• To outline the evolutions in the Cosmos platform's infrastructure.
• To detail the growth of available data within the platform.
• To describe the usage patterns and applications of the Cosmos platform for research.

Main Methods:

• Data is securely transferred via Health Level Seven Clinical Document Architecture to the Cosmos host.
• Deduplicated data is stored in relational SQL databases, accessible remotely.
• Analysis is performed using a no-code tool for aggregate data and a secure virtual environment for de-identified line-level data.

Main Results:

• Over 330 organizations contribute data from Canada, Lebanon, Saudi Arabia, and the United States.
• The platform contains over 300 million deduplicated patient records, with 39% having more than 5 years of clinical data.
• Cosmos data has supported over 100 peer-reviewed studies across various disciplines.

Conclusions:

• The expanding Cosmos platform, with its data breadth and tooling, supports diverse study types.
• Cosmos exemplifies vendor-facilitated collaboration for extensive research enablement.
• Limitations include workflow variations, documentation errors, and potential bias from a single EHR vendor user base.