Data Quality-Driven Improvement in Health Care: Systematic Literature Review.

Area of Science:

• Health Informatics
• Data Quality Management
• Real-World Evidence

Background:

• High-quality data is fundamental for advancing real-world evidence and learning health care systems.
• Despite recognized issues with data quality (DQ), standardized best practices for assessment and improvement remain elusive.

Purpose of the Study:

• To systematically review how research studies define, assess, and improve the quality of structured real-world health care data.
• To identify current methodologies and frameworks used in data quality improvement initiatives.

Main Methods:

• Conducted a systematic literature search of Embase and PubMed (1945-2023) for studies on structured real-world data quality.
• Standardized data quality concepts using the Data Management Association (DAMA) DQ framework.
• Identified 39 relevant articles reporting data quality improvement initiatives after screening by two independent authors.

Main Results:

• Significant heterogeneity observed in study settings, DQ assessment methods (largely manual, focusing on completeness), and DQ frameworks used.
• Limited use of standardized methodologies for quality improvement, with Plan-Do-Study-Act (PDSA) and Define-Measure-Analyze-Improve-Control (DMAIC) cycles being common.
• Most effective DQ improvements resulted from combined interventions including feedback, IT solutions, training, workflow enhancements, and data cleaning.

Conclusions:

• An urgent need exists for standardized guidelines in data quality improvement research to enable synthesis and comparison of findings.
• Frameworks like PDSA and DAMA, alongside root cause analysis, can guide sustainable DQ improvement.
• Standardizing DQ assessment, analysis, and improvement methodologies is crucial for enhancing effectiveness, comparability, and generalizability of research.