Leveraging PubMed to Create a Specialty-Based Sense Inventory for Spanish Acronym Resolution.

Area of Science:

• Clinical Natural Language Processing
• Biomedical Informatics
• Computational Linguistics

Background:

• Acronyms are prevalent in clinical texts, posing challenges for identification and disambiguation.
• Accurate acronym resolution is crucial for effective clinical natural language processing (NLP).

Purpose of the Study:

• To develop a comprehensive resource for Spanish clinical acronym resolution.
• To organize acronyms by clinical specialty with expansions and corpus-driven features.

Main Methods:

• Creation of a Spanish acronym resource covering 51 clinical specialties.
• Identification of language-independent and language-dependent acronym expansions.
• Analysis of acronym sense inventory across different medical fields.

Main Results:

• The resource contains 3,603 acronyms across 51 specialties.
• Identified 228 language-independent acronyms and 391 language-dependent expansions.
• Novel insights into Spanish biomedical acronym usage were uncovered.

Conclusions:

• The developed resource significantly contributes to Spanish clinical NLP.
• Facilitates improved accuracy in identifying and resolving clinical acronyms.
• Enhances the analysis of biomedical Spanish texts through better acronym understanding.