Data Harmonization, Standardization, and Quality Assessment in Distributed Health Data Networks: Lessons from Around the World

Background: The growing importance of real-world data for regulatory decision making has spurred the creation of federated health data networks (FHDNs) that facilitate access to large patient populations and enable multi-site unified analyses to address questions that require more than a single data repository.

Objectives: The goal of the symposium is to discuss lessons learned, challenges, and opportunities related to FHDN data standardization, harmonization, curation, and quality assessment for regulatory decision making

Description: The operation of FHDNs raises important challenges related to data standardization, harmonization, characterization, curation, and assessment of fitness-for-use. The panel will describe their experiences and lessons learned leading data quality and curation activities in their respective networks. Lisa Lix (15 minutes) will detail the data curation and data quality assessment activities of the Canadian Network for Observational Drug Effect Studies (CNODES) network, which has access to data from a total population of over 40 million. CNODES consists of a network of 8 sites (7 Canadian provinces with access to administrative data and the UK Clinical Practice Research Datalink which comprises electronic health records) and over 100 investigators who together conduct research to support medical product safety assessments for Health Canada and the Canadian provinces. Recent activities includes harmonization of key CNODES study variables and documentation about data quality assessment practices. Judy Maro (15 minutes) will detail the 12+ years of experience of FDA Sentinel in conducting data curation across the Sentinel network, including discussion of how data curation has evolved and expansion of the data model to additional data domains and to support international collaborations. Matvey Palchuk (15 minutes) will describe the approach used by TriNetX, a global federated network of electronic medical record data, to standardize, harmonize, and assess data quality in network spanning 16 countries, including 44 sites in Europe, 73 in the US, 4 in Asia, and 5 in Latin America. Robert Reynolds (5 minutes) will describe approaches to assessing data sources, fitness-for-purpose, and transparency for internal decision-making and regulatory submissions from the perspective of a product sponsor. Xavier Kurz (5 minutes) will describe the fitness-for-purpose assessment for data partners, the data quality framework and the analytical tools applied in the DARWIN EU® federated data network initiated in 2022 by the European Medicines Agency. Jeffrey Brown will provide a brief introduction (5 minutes) to the speakers and the symposium and moderate the discussion (30 minutes).