Axon Registry® data validation: Accuracy assessment of data extraction and measure specification

Area of Science:

• Health Informatics
• Clinical Data Management
• Quality Improvement

Background:

• The Axon Registry® is a clinical quality data registry.
• Accurate data extraction from electronic health records (EHRs) is crucial for reliable quality reporting.

Purpose of the Study:

• To perform a data validation study assessing the accuracy of the Axon Registry® data extraction process.
• To evaluate the agreement between data abstracted by the Axon Registry® and an independent auditor.

Main Methods:

• An external auditor (IQVIA) abstracted data from EHRs at nine diverse clinical sites.
• IQVIA independently calculated quality measure performance rates using American Academy of Neurology specifications.
• Agreement and discordance between Axon Registry® and IQVIA data were analyzed.

Main Results:

• High match rates (≥92%) were found for discrete data elements (e.g., demographics).
• Lower match rates (<44%) were observed for free-text elements (e.g., plan of care).
• Patient-level measure performance agreement was 76% (κ = 0.53, p < 0.001).

Conclusions:

• Variable concordance exists between Axon Registry® data and independently abstracted data.
• Data validation is essential for clinical quality registries using automated EHR data extraction.
• Remediation strategies are needed to enhance data accuracy for reliable quality reporting.