Undergraduate nursing students' perceptions, needs, and expectations regarding large language model-based virtual...

Area of Science:

• Nursing Education
• Artificial Intelligence in Healthcare
• Communication Skills Training

Background:

• Traditional nursing communication training faces limitations in authenticity and practice opportunities.
• Large language model-based virtual patients (LLM-VPs) present a novel approach to address these limitations.
• Limited research exists on student perceptions of LLM-VPs in nursing education.

Purpose of the Study:

• To explore undergraduate nursing students' anticipated perceptions of LLM-VPs for communication training.
• To identify functional needs, acceptance, and potential risks associated with LLM-VP integration.
• To gather preliminary evidence for incorporating LLM-VPs into nursing education.

Main Methods:

• Descriptive qualitative design utilizing semi-structured interviews with 15 undergraduate nursing students.
• Data collection and analysis were performed concurrently until theoretical saturation.
• Thematic analysis, aided by NVivo 14 software, was employed for data interpretation.

Main Results:

• Four themes emerged: educational gaps, system design requirements, trust mechanisms, and perceived risks.
• Students expressed cautiously optimistic expectations for LLM-VPs.
• Diverse functional expectations were articulated to align LLM-VPs with nursing-specific learning needs.

Conclusions:

• LLM-VPs show practical appeal and potential utility for nursing communication education from a student perspective.
• The educational value of LLM-VPs is conditional and requires validation through real-world studies.
• Students view LLM-VPs as supplementary tools for practice and feedback, not replacements for clinical encounters.