Operationalizing machine-assisted translation in healthcare

Area of Science:

• Health Informatics
• Natural Language Processing
• Healthcare Management

Background:

• Over 25 million U.S. patients have limited English proficiency, facing risks due to untranslated medical information.
• Current translation services are often too slow, compromising patient safety and care quality.
• Lack of implementation guidance hinders the adoption of advanced translation technologies in healthcare.

Purpose of the Study:

• To address the critical need for timely translation of healthcare materials for non-English speaking patients.
• To provide a practical implementation roadmap for integrating large language model (LLM) machine-assisted translation in healthcare settings.
• To guide healthcare leaders and policymakers in leveraging LLMs to improve health equity.

Main Methods:

• Utilized the Consolidated Framework for Implementation Research (CFIR) to identify key implementation factors.
• Analyzed considerations across five CFIR domains: innovation, individuals, inner setting, implementation process, and outer setting.
• Developed a strategic framework for integrating LLM-assisted translation.

Main Results:

• Identified critical implementation factors for successful LLM translation integration.
• Outlined specific considerations for innovation, stakeholder engagement, organizational readiness, process management, and external policies.
• Provided a structured approach for healthcare systems to adopt machine-assisted translation.

Conclusions:

• LLM-assisted translation offers a scalable solution to bridge the language gap in healthcare.
• A systematic, framework-guided approach is essential for effective implementation of translation technologies.
• Integrating LLMs can significantly enhance patient safety and reduce health disparities for diverse patient populations.