Best Practices to Design, Plan, and Execute Large-Scale Federated Analyses-Key Learnings and Suggestions from a Study...

Area of Science:

• Health Informatics
• Observational Research
• Open Science

Background:

• Federated network studies enable data privacy by keeping data local while sharing analytical code and results.
• The Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) facilitates standardized observational research across diverse health systems.
• Federated analyses using OMOP CDM are crucial for generating robust, reproducible, large-scale health evidence.

Purpose of the Study:

• To share key lessons and strategies for conducting complex, large-scale, multidatabase federated analyses.
• To provide best practices for leveraging the OMOP Common Data Model in federated research networks.
• To document the experience of conducting large-scale federated analyses within the European Health Data and Evidence Network (EHDEN).

Main Methods:

• Conducted large-scale federated analyses involving 52 databases across 19 countries.
• Utilized the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) for data standardization.
• Employed meticulous planning, community building, standardized analytics, and strategic division of responsibilities.

Main Results:

• Successful execution of large-scale federated analyses across 52 databases from 19 countries.
• Identification of essential strategies including planning, collaboration, communication, and standardized analytics.
• Positive feedback from data custodians, highlighting the value of network engagement and shared learning.

Conclusions:

• Meticulous planning, strong community engagement, and standardized analytics are critical for successful large-scale federated network studies.
• Federated analyses using the OMOP CDM are a powerful tool for generating robust health evidence.
• Continuous improvement and shared learning are vital for the future of federated health data research.