Analysis of Clinical Record Data for Anticoagulation Management within an EHR System

Area of Science:

• Health Informatics
• Clinical Data Management

Background:

• Generic electronic health record (EHR) information models are developed based on extensive requirements gathering from clinical communities.
• New-generation interoperability standards often incorporate these comprehensive models, aiming for broad applicability.

Purpose of the Study:

• To evaluate the properties of a generic EHR information model that were actually required and utilized during the import of an existing clinical application.
• To assess the gap between theoretical requirements and practical usage in EHR data representation.

Main Methods:

• Developed a generic EHR repository and system within EU Projects Synapses and SynEx.
• Deployed a web application for anticoagulation therapy management in a London hospital, interfacing with the EHR system.
• Imported five years of cumulative clinical data from a legacy anticoagulation management application to critique the generic EHR's utilized components.

Main Results:

• Imported records of 3,226 patients, utilizing over 900,000 Record Components (average 289 nodes per patient).
• Approximately two-thirds of components were value-containing leaf nodes, with the remainder being structural container nodes.
• A significant portion of the model's properties were found to be unused or infrequently used, with only about 0.5% of Record Components ever revised.

Conclusions:

• The development of generic EHR information models has been requirement-driven without prioritization, leading to potentially excessive complexity.
• This evaluation demonstrated that a small fraction of the generic EHR model is sufficient to represent substantial clinical data.
• Further evaluations across diverse clinical systems are necessary to balance theoretical requirements with practical needs for optimal EHR information model complexity.