A Customizable Data Quality Tool for Global Observational Research Networks.

Area of Science:

• Clinical Data Management
• Health Informatics
• Observational Studies

Background:

• Multi-site observational clinical data analysis requires robust data quality evaluation.
• Existing data quality tools often demand significant technical expertise, limiting their adoption by research consortia.
• Generalizing data quality solutions is crucial for improving the reliability of collaborative research.

Purpose of the Study:

• To generalize an established data quality checking and report generation tool for easier implementation across diverse research networks.
• To reduce the technical expertise needed at user sites for data quality assessment.
• To provide a scalable solution for data quality and reporting in multi-site research consortia.

Main Methods:

• Leveraged REDCap (Research Electronic Data Capture) software to store data model details, variable expectations, and user group information.
• Developed an application using the REDCap API to retrieve stored details and assess dataset conformance, logical consistency, and completeness.
• Built the Harmonist Data Toolkit using freely available REDCap and R/Shiny platforms, with code accessible on GitHub.

Main Results:

• The generalized Harmonist Data Toolkit successfully reduced the need for local programming expertise.
• All five collaborating consortia reported benefits in detecting inconsistencies and generating informative data reports and visualizations.
• The toolkit effectively assessed datasets for conformance to data models, logical consistency, and completeness.

Conclusions:

• The Harmonist Data Toolkit provides a valuable, accessible solution for data quality and report generation in research consortia.
• The toolkit addresses the need for standardized data quality assessment in multi-site observational studies.
• Freely available platforms like REDCap and R/Shiny enable the development of powerful, generalizable research tools.