On the Design of a Sign Language Corpus of Medical Terms for Automatic Translation Systems: Mixed Methods Approach

Area of Science:

• Medical Informatics
• Health Communication
• Assistive Technology

Background:

• Hearing loss presents global communication barriers in healthcare access.
• Ineffective communication can lead to patient complications and iatrogenic events.
• Assistive technologies are crucial for enhancing communication efficiency in medical settings.

Purpose of the Study:

• To develop a medical term corpus for the Captar-Libras project.
• To improve communication between healthcare professionals and deaf patients.
• To create a bidirectional sign language system for healthcare.

Main Methods:

• Utilized the Delphi method for expert consensus on key medical terms.
• Developed and distributed questionnaires on emergency questions to healthcare professionals.
• Adapted terms to Brazilian Sign Language (Libras) with expert input and simulated scenarios for validation.

Main Results:

• Physicians (87.5%) with emergency care experience participated.
• Questions on initial symptoms and pain intensity received high importance scores.
• Modifications improved clarity on smoking/alcohol use, enhancing medical interview structure.

Conclusions:

• This study provides a methodological framework for developing a Libras medical corpus.
• Involved deaf sign language experts to ensure authentic usage and articulation in Libras.
• The structured, participatory approach with interdisciplinary validation enhances sign language recognition in medical settings.