(A-204) Clearing Real-World Data Quality Hurdles: Characterizing Quality Attributes of the Guardian Research Network in the Context of Regulatory Submission

Background: The quality of real-word data (RWD) directly impacts the value of real-world evidence (RWE) generated for regulatory decision-making. Data owners and investigators must be prepared to provide documentation on data quality assessment when submitting evidence generated from secondary data for regulatory purposes. While robust feasibility is required to justify the relevance of RWD for a specific research question, the reliability of the data, including data collection processes prior to reaching the end user, is of equal importance for drawing valid conclusions.

Objectives: To illustrate fundamental requirements for demonstrating both intrinsic and research-specific data quality to regulators, using a nationally representative network of electronic health records (EHRs) aggregated from 14 United States (US) health systems.

Methods: We first describe the Guardian Research Network (GRN), including data elements collected, timeliness of data availability, representativeness, and access considerations. We then describe how data reliability (accuracy, traceability, timeliness, completeness) and relevance (availability, sufficiency, representatives) are ensured and assessed in GRN, with illustrative examples of data quality controls.

Results: GRN applies quality controls to identify data issues, including logical checks and assessments for missingness. Rigorous validation processes verify the integrity and accuracy of aggregated data, ensuring required standards are met. The GRN data collection system maintains an audit trail, tracking data entry, deletions, and alterations. Data transformations, such as anonymizing dates, are implemented to protect patient privacy. Data is available within 24 hours of entry into the source system, enabling timely analysis and decision-making. GRN has a flexible process for data collection including options for enrichment with primary data collection and curation of unstructured data such as clinical notes, reports, and images.

Conclusions: The complex journey of RWD underscores the need for continued attention to data quality practices for to ensure fitness for regulatory use RWD. GRN’s data quality controls demonstrate this through documentation of data lineage, iterative assessment of data transformations, and harmonization methods. These foundational practices support appropriate study design and transparent communication during regulatory decision-making.