Reconciling users' needs and formal requirements: issues in developing a reusable ontology for medicine

Area of Science:

• Medical Informatics
• Computational Linguistics
• Knowledge Representation

Background:

• Establishing a standardized medical terminology is crucial for individual clinical systems and integrating distributed healthcare applications.
• Formal representations using description logics are increasingly adopted for medical terminologies.

Purpose of the Study:

• To address challenges in making description logic-based ontologies accessible to clinicians and developers.
• To identify essential features for description logic ontologies to achieve aims in medical terminology.

Main Methods:

• Proposing four contentions based on GALEN's experience with description logic ontologies.
• Discussing the necessity of natural language generation, intermediate representations, orthogonal taxonomies, and transitive relation support.

Main Results:

• Natural language generation is vital for ontology accessibility.
• Intermediate representations and user-specific perspectives enhance usability.
• Orthogonal taxonomies support ontology independence and reuse.
• Description logic must support transitive relations, with acceptable computational costs.

Conclusions:

• These features are necessary for developing large, reusable medical ontologies.
• Balancing expressivity with computational cost is key for practical application.
• User-centered design and modularity are essential for effective medical terminology systems.