Targeted Data Quality Analysis for a Clinical Decision Support System for SIRS Detection in Critically Ill Pediatric...

Area of Science:

• Health Informatics
• Data Science
• Clinical Decision Support

Background:

• Data quality issues in clinical decision support systems (CDSSs) can lead to incorrect decisions.
• Assessing local data quality is crucial for successful CDSS adoption and preventing failures.

Purpose of the Study:

• To develop a shareable set of applicable measurement methods (MMs) for targeted data quality assessment.
• To determine the suitability of local data for a specific CDSS.

Main Methods:

• Derived task-specific MMs using GUI-based analysis (openCQA), case analysis of false CDSS decisions, data-driven learning, and a systematic check using the HIDQF framework.
• Expressed data quality knowledge using 5-tuple-formalization for MMs.

Main Results:

• Identified task-specific dataset characteristics for targeted data quality assessment.
• Defined 394 MMs organized into 13 data quality knowledge bases.

Conclusions:

• Created a shareable set of MMs for targeted data quality assessment in CDSS-based detection of systemic inflammatory response syndrome (SIRS) in critically ill pediatric patients.
• Promotes targeted data quality assessment as a standard practice in health care research for data-consuming systems.