Initializing a hospital-wide data quality program. The AP-HP experience.

Area of Science:

• Health Informatics
• Data Science
• Clinical Research

Background:

• Electronic health record (EHR) data is crucial for learning healthcare systems.
• The AP-HP Clinical Data Repository aggregates EHR data from 37 hospitals for research.
• Established data quality (DQ) programs are essential for reliable EHR data utilization.

Purpose of the Study:

• Describe the AP-HP DQ program.
• Report lessons learned from two DQ campaigns initiated in 2017.
• Assess the impact of DQ initiatives on EHR data quality.

Main Methods:

• Conducted DQ campaigns focusing on patient identification (PI) and healthcare services (HS) domains.
• Employed a 5-phase approach: scope definition, measurement, analysis, improvement, and control.
• Utilized semi-automated DQ profiling on large datasets (8.8M patients, 13,099 consultation agendas).
• Defined 17 DQ measures and classified issues using a unified DQ reporting framework.

Main Results:

• Identified 11 DQ issues across PI and HS datasets, categorized as completeness, conformance, and plausibility.
• Root causes included data originator errors, ETL issues, and EHR tool limitations.
• Developed 16 action plans for improvement and monitoring.
• Observed significant DQ measure improvements despite partial implementation.

Conclusions:

• DQ assessments in hospital information systems are underreported.
• AP-HP DQ campaigns demonstrate the value of systematic DQ programs.
• A unified DQ reporting framework facilitates clear communication of findings.
• Need for dedicated tooling to automate and expand DQ program scope.
• Per-study DQ checks are necessary to ensure EHR data suitability for specific research.