Conceptual analysis of Chatbots in nursing education applications

Chatbots are receiving growing attention in nursing education, yet the term is often used imprecisely and is sometimes conflated with current large language model (LLM)-based systems. This lack of conceptual clarity weakens cumulative research, hinders comparison across studies, and complicates implementation in educational practice.

To clarify the concept of chatbots in nursing education using Walker and Avant's concept analysis method.

Walker and Avant's eight-step concept analysis method was applied. To examine both ordinary and disciplinary uses of the concept, dictionary sources and literature from PubMed, Scopus, Embase, CINAHL, Web of Science, ProQuest, ScienceDirect, China National Knowledge Infrastructure, Wanfang, and VIP were reviewed through June 2025. A structured extraction matrix captured how the term was used, the chatbot's interaction form, response architecture, educational purpose, and conceptual boundaries. Primary empirical studies were prioritized when deriving defining attributes, whereas reviews and discussion or conceptual papers were used to map broader terminology, technological heterogeneity, and contested boundary claims.

Twenty-four papers were included: 11 primary empirical studies, 7 reviews, and 6 discussion or conceptual sources. Comparison across the extraction matrix showed that three kinds of description recurred when authors clearly referred to a nursing education chatbot: iterative dialogue, automated system response, and explicit educational alignment with nursing. These features were retained because they were recurrent, necessary to recognize the concept, and distinguishing from adjacent technologies. By contrast, LLM-based design, personalized learning support, and clinical simulation were excluded from the defining set because they appeared mainly as technical variants, reported benefits, or role-specific applications rather than universal properties. Antecedents were retained only when they described pre-existing conditions that enabled development or adoption. Consequences were treated as reported or anticipated downstream effects rather than as evidence of effectiveness. The model, borderline, and contrary cases were used to test the conceptual boundary by illustrating full presence, partial presence, and absence of the defining attributes. Empirical referents were specified as observable indicators for each attribute rather than as a single perception instrument.

A nursing education chatbot is best understood as a conversational digital agent intentionally designed or deployed for nursing education and able to provide automated responses through rule-based, retrieval-based, or AI-based mechanisms. This refined definition separates the core concept from any single generation of technology and provides a clearer conceptual basis for future reporting, instrument development, and outcome evaluation without implying that chatbots inherently improve educational effectiveness.