Evaluating a Hierarchical Clinical Event Linkage Model for Clinic-Specific Databases

Area of Science:

• Clinical Informatics
• Database Design
• Health Data Management

Background:

• Clinical data often lacks explicit context, hindering retrieval and interpretation.
• Existing database models may not adequately capture the hierarchical nature of clinical events.

Purpose of the Study:

• To present a relational database model designed to preserve the contextual meaning of clinical event attributes.
• To evaluate the model's effectiveness in representing and retrieving clinical data.

Main Methods:

• Developed a relational database model incorporating hierarchical linkages and qualifier codes for clinical events.
• Conducted a retrospective analysis of 302 forms derived from the model.
• Assessed the representation and retrieval rates of 17,899 data elements across 19 clinics.

Main Results:

• The model successfully represented 91% of requested clinical data elements.
• A significant portion (62%) of represented data elements were used only once.
• Data retrieval was achieved for clinicians and researchers across multiple clinics.

Conclusions:

• The developed relational database model effectively preserves the contextual meaning of clinical data.
• The low reuse rate of data elements underscores the specificity of clinic-specific databases.
• Explicitly stored, unambiguous clinical data is crucial for effective data management and research.