Developing a Nursing Research Education Agent Using Knowledge Graphs and Large Language Models

BACKGROUND: Integration of knowledge graphs (KGs) and large language models (LLMs) holds transformative potential for nursing education, particularly in research methodology and statistical literacy. This proof-of-concept study developed a Nursing Research Education Agent to support learners in understanding research designs and statistical concepts. PURPOSE: To evaluate the agent's feasibility, pedagogical alignment, and AI performance using validated instruments in a nursing education context. METHODS: The agent combined structured KGs of nursing research knowledge with LLMs for interactive, natural-language responses. Ten nursing educators assessed it using the 10-item Pedagogical Fit Evaluation Scale and 10-item AI Performance Evaluation Scale. RESULTS: Educators rated pedagogical fit highly (M = 4.20, SD = 0.63) and AI performance strongly (M = 4.10, SD = 0.56), praising clinical relevance, accuracy, and promotion of critical thinking. Integration into curricula was deemed feasible. CONCLUSIONS: KG-LLM-integrated agents show strong promise for nursing research education. Further development and larger-scale trials are recommended.