Development of an integrated AI-VR simulation system for preoperative nursing competency: A pilot study

AI-driven virtual standardized patients (VSPs) integrated within virtual reality (VR) environments offer structured opportunities for safe communication practice in nursing education; however, empirical evidence regarding their preliminary feasibility and educational effects remains limited, particularly in domain-specific clinical contexts such as preoperative nursing care.

This mixed-methods pilot study pursued two objectives: (1) to develop a VR simulation integrating an AI-driven VSP (AI-VSP) for preoperative nursing communication, and (2) to pilot-test its preliminary feasibility and exploratory pre-post effects. Ten third-year nursing students completed a single 50-minute preoperative care simulation using the VIRTI platform (Meta Quest 3). Quantitative data on satisfaction, self-confidence, simulation design perceptions, and immersion were collected using four validated instruments and analysed using paired-samples t-tests with effect sizes and 95% confidence intervals. Qualitative data were collected via semi-structured interviews and analysed using inductive thematic analysis.

All four outcome variables showed statistically significant pre-post differences (all p ≤ .004), with large effect sizes (Cohen's d = 1.10-1.75). Qualitative themes indicated realistic conversational learning experiences and perceived clinical relevance, while identifying technical limitations in nonverbal feedback and scenario-level construct alignment as areas for refinement.

These preliminary findings suggest that AI-VSP-based VR simulation may offer a feasible and structured pre-clinical learning environment for preoperative nursing communication competency development. Future work should prioritise multimodal feedback enhancement and controlled comparative designs to evaluate educational effectiveness more rigorously.