Electronic health record data quality variability across a multistate clinical research network

Area of Science:

• Health Informatics
• Data Science
• Clinical Research

Background:

• Electronic health record (EHR) data quality issues can impact research outcomes and decision support.
• Current methods for evaluating EHR data quality lack a standardized best practice.
• Assessing EHR data quality variability across multiple healthcare systems is crucial.

Purpose of the Study:

• To quantify data quality concerns within the PCORnet Clinical Research Network.
• To compare a rule-based framework with the existing PCORnet data curation process.
• To explore clinical care variability and quality using testosterone therapy prescribing data.

Main Methods:

• A rule-based framework, adapted to the PCORnet Common Data Model, was applied at 13 clinical sites.
• Data quality assessments were performed across eight states.
• Results were compared against the current PCORnet data curation procedures.

Main Results:

• The rule-based framework detected significant data quality discrepancies across all participating sites.
• The framework identified additional data errors with high specificity, aiding technical remediation.
• Analysis revealed variability in clinical care, particularly in testosterone therapy prescribing.

Conclusions:

• Rule-based methods effectively quantify significant discrepancies in EHR data quality across healthcare systems.
• Data errors often originate from medication and laboratory data sources.
• The developed framework can support clinical care variability and quality improvement programs.