Intention to Use Large Language Models Among Clinical Nurses in China With Prior Familiarity With or Experience Using...

To explore the intention to use large language models (LLMs) among clinical nurses with prior familiarity with or experience using LLMs, and to examine how relevant facilitators, constraints, perceived risks, and professional boundary considerations were reflected in such intention.

Guided by the Unified Theory of Acceptance and Use of Technology (UTAUT), this descriptive qualitative study employed semistructured interviews with 17 clinical nurses from different hospital levels, specialties, and roles. Data were analyzed using directed content analysis.

Six themes were identified: performance expectancy, effort expectancy, social influence, facilitating conditions, perceived risk, and professional boundaries and identity. Participants expressed a positive but conditional intention to use LLMs. Performance expectancy, effort expectancy, social influence, and facilitating conditions were reflected in participants' accounts as perceived enabling conditions, whereas perceived risk and professional boundaries and identity were described as important sources of caution. Participants acknowledged the potential supportive value of LLMs in standardized and relatively low-risk tasks, such as documentation, knowledge support, and teaching-related work, but emphasized that LLMs should not replace nurses' clinical judgment, accountability, or humanistic communication.

Among clinical nurses with prior familiarity with or experience using LLMs, intention to use these tools was reflected in participants' accounts of perceived value, usability, organizational conditions, risk appraisal, and professional boundaries. Conditional intention should be understood as an interpretive finding rather than as a validated new construct or a formal extension of UTAUT. These findings provide context-specific insights for cautious pilot testing, governance, and implementation of LLM-supported nursing applications.

Nurse managers may consider cautious pilot testing of LLM-supported nursing applications in low-risk and standardized tasks. Priorities include defining appropriate-use boundaries, data security, and accountability requirements, ensuring human review, providing tiered training, and evaluating usability, workflow fit, safety concerns, and nurses' feedback before wider implementation.