A Neuro-ontology for the neurological examination

Area of Science:

• Medical Informatics
• Computational Linguistics
• Clinical Data Management

Background:

• Electronic health records (EHRs) require converting free text into machine-readable codes for data analytics and machine learning.
• Capturing complex clinical information, such as neurological examinations, in a structured format is a significant challenge.

Purpose of the Study:

• To assess the feasibility of creating machine-readable codes for neurological examinations using the UMLS Metathesaurus.
• To develop and evaluate a target ontology for encoding neurological examination findings.

Main Methods:

• A target ontology was constructed using 1100 concepts from the UMLS Metathesaurus.
• A dataset of 2386 test phrases from 419 published neurological cases was created.
• Test phrases were mapped to the developed ontology.

Main Results:

• All 2386 test phrases were successfully mapped to 601 unique UMLS concepts.
• The 1100-concept ontology demonstrated sufficient coverage for neurological examination concepts.
• Component ontologies like HPO, SNOMED CT, and OMIM lacked the necessary depth and breadth for this task.

Conclusions:

• An ontology derived from a subset of the UMLS Metathesaurus can effectively convert neurological examination deficits into machine-readable codes.
• Utilizing a curated subset of UMLS concepts improves ontology manageability and allows for refinement of hierarchical relationships.