Understanding data differences across the ENACT federated research network

Area of Science:

• Health Informatics
• Clinical Research
• Data Science

Background:

• Federated research networks facilitate medical research by exchanging electronic health record (EHR) data.
• Poor data quality in EHRs can significantly hinder research progress.
• Existing data quality methods often rely on rigid standards, limiting adaptability.

Purpose of the Study:

• To develop a novel, data-centric method for identifying data quality issues in federated research networks.
• To create a privacy-preserving system applicable even to sites not yet fully integrated into the network.
• To enhance the reliability of EHR data for clinical trial accrual.

Main Methods:

• Distributed high-performance patient counting scripts (Integrating Biology at the Bedside - i2b2) across ENACT sites.
• Aggregated site-contributed patient counts at the ENACT Hub to generate network statistics.
• Developed the Data Quality Explorer (DQE) web application to visualize and investigate data quality relative to network benchmarks.

Main Results:

• Thirteen ENACT sites contributed patient counts, with ten actively using DQE for data quality analysis.
• The system successfully generates network statistics from aggregated patient counts.
• The Data Quality Explorer (DQE) facilitates data quality investigation for participating sites.

Conclusions:

• Implemented a novel metric for data quality investigation in federated networks using patient counting and network statistics.
• The developed end-to-end pipeline is privacy-preserving and adaptable to various research networks.
• This approach offers a low-barrier, evolving solution to improve EHR data quality for clinical research.