Mapping Clinical Documents to the Logical Observation Identifiers, Names and Codes (LOINC) Document Ontology using...

Area of Science:

• Health Informatics
• Clinical Documentation
• Data Management

Background:

• Electronic Health Record (EHR) systems generate vast amounts of clinical documentation.
• Categorizing this data for clinical care and research presents significant challenges.
• Existing natural language processing (NLP) methods for text categorization face limitations in scalability and metadata utilization.

Purpose of the Study:

• To present a framework for mapping clinical notes to the LOINC document ontology.
• To address the shortcomings of NLP techniques in terms of computational scalability and metadata integration.
• To enable efficient and accurate categorization of clinical documentation within healthcare institutions.

Main Methods:

• A Bag of Words approach is utilized for mapping notes to the LOINC document ontology.
• Preliminary manual value-set mapping is performed.
• An automated pipeline leverages structured EHR metadata to align notes with the document ontology dimensions.

Main Results:

• The framework achieved 73.4% coverage of Electronic Health Record documents.
• Over 132 million notes were mapped in under 2 hours.
• The method demonstrated an order of magnitude improvement in efficiency compared to NLP-based approaches.

Conclusions:

• The proposed framework offers a scalable and efficient solution for categorizing clinical documentation.
• This approach enhances the usability of EHR data for clinical care and research.
• Institutions can effectively map their notes to the LOINC document ontology, improving data accessibility.