Complementing Medical Records with Precalculated Data Items to Facilitate Decision Support and Phenotyping

Area of Science:

• Computer Science
• Medical Informatics
• Health Informatics

Background:

• The Arden Syntax standard facilitates clinical decision support via Medical Logic Modules (MLMs).
• Early Arden Syntax versions had limited data types; later versions introduced static object types for complex data structures like Electronic Medical Records (EMRs).
• Modifying EMR structures at runtime was not supported in standard Arden Syntax versions.

Purpose of the Study:

• To enhance the Arden Syntax by enabling dynamic modification of object data structures at runtime.
• To reduce Medical Logic Module (MLM) complexity and improve knowledge encoding by pre-calculating data items.
• To explore benefits for clinical decision support, phenotyping queries, and Electronic Medical Record (EMR) data preprocessing.

Main Methods:

• Modified the object data type in an experimental Arden Syntax version (PLAIN) to allow runtime structure extension.
• Implemented explicit creation of new attributes and implicit attribute creation upon value assignment.
• Integrated a mapper for arbitrary data structures, including EMRs, within the PLAIN version.

Main Results:

• Successfully complemented EMRs with patient-individual recommendations for expiratory tidal volume in the ventilation section.
• Demonstrated that pre-calculated data items can be added to EMRs without altering MLM logic.
• Showcased the ability to extend object structures dynamically at runtime.

Conclusions:

• Extending object structures at runtime simplifies MLM development and maintenance for clinical decision support.
• Dynamic data structure modification offers advantages for phenotyping queries and EMR data preprocessing.
• The enhanced Arden Syntax approach facilitates more efficient and flexible clinical informatics solutions.