Developing a Nursing Research Education Agent Using Knowledge Graphs and Large Language Models: A Proof-of-Concept...

Area of Science:

• Nursing Education
• Artificial Intelligence in Healthcare
• Educational Technology

Background:

• Knowledge graphs (KGs) and large language models (LLMs) offer significant potential for advancing nursing education, particularly in research methodology and statistical literacy.
• A proof-of-concept study developed a specialized agent to aid nursing learners in grasping research designs and statistical concepts.

Purpose of the Study:

• To assess the feasibility and pedagogical suitability of the developed agent within a nursing education setting.
• To evaluate the artificial intelligence (AI) performance of the agent using established assessment tools.

Main Methods:

• The agent integrated structured KGs of nursing research knowledge with LLMs to provide interactive, natural language responses.
• Ten nursing educators evaluated the agent using the Pedagogical Fit Evaluation Scale and the AI Performance Evaluation Scale.

Main Results:

• Educators reported high ratings for pedagogical fit (M=4.20, SD=0.63) and AI performance (M=4.10, SD=0.56).
• Positive feedback highlighted the agent's clinical relevance, accuracy, and its ability to foster critical thinking skills among students.
• The integration of the agent into nursing curricula was considered feasible.

Conclusions:

• Agents that integrate KGs and LLMs demonstrate considerable potential for enhancing nursing research education.
• Further research and large-scale trials are recommended to fully explore the capabilities and impact of these integrated agents.