Raising the Bar for Real-World Data in Oncology: Approaches to Quality Across Multiple Dimensions

Area of Science:

• Oncology Research
• Real-World Data (RWD) Science
• Health Informatics

Background:

• Electronic health records (EHRs) are a vital source of real-world data (RWD) for oncology research.
• Ensuring the fitness for use of EHR-derived RWD is critical for reliable research outcomes.
• The complexity of EHR data necessitates transparent and systematic quality curation methods.

Purpose of the Study:

• To describe the application of data quality dimensions in curating EHR-derived oncology RWD.
• To enhance transparency in the quality assessment of oncology RWD.
• To ensure the fitness for use of EHR-based RWD for real-world evidence generation.

Main Methods:

• A targeted review of data quality dimensions from major health regulatory and policy organizations was conducted.
• Quality processes applied to Flatiron Health's EHR-derived oncology RWD were characterized against these dimensions.
• Specific quality dimensions included relevance (availability, sufficiency, representativeness) and reliability (accuracy, completeness, provenance, timeliness).

Main Results:

• Key quality dimensions identified were relevance and reliability, with specific subdimensions.
• Flatiron Health RWD quality processes were aligned with these dimensions.
• Relevancy is optimized by dataset size and variable scope; reliability is addressed through validation, verification, provenance tracking, and timely data refreshes.

Conclusions:

• Systematic processes across the data lifecycle are essential for high-quality, scaled EHR-based RWD.
• Optimizing data quality requires understanding data sources, curation, and use case needs.
• Adherence to established quality dimensions enhances transparency and fitness for real-world evidence generation.