Comparing the effects of GenAI-assisted and facilitator-led debriefing on knowledge and non-technical skills in immersive virtual reality simulation in nursing education: A quasi-experimental study.
BACKGROUND Immersive virtual reality (IVR) simulation is increasingly used in nursing education to support experiential learning and the development of non-technical skills. Debriefing is integral to simulation-based education; however, shortages of qualified facilitators and variability in delivery can constrain scalability and quality. Generative artificial intelligence (GenAI) may offer a standardised and scalable approach to post-simulation debriefing, yet comparative evidence in nursing education remains limited. AIM To compare the effectiveness of GenAI-assisted versus facilitator-led debriefing on nursing students' knowledge acquisition and self-reported non-technical skills following an IVR nursing simulation. DESIGN Quasi-experimental, two-group pretest-posttest study. SETTING A higher education institution in Hong Kong. PARTICIPANTS Undergraduate nursing students with less than four weeks of clinical internship experience. METHODS Participants were allocated by subject enrolment to either GenAI-assisted debriefing plus a 10-minute facilitator summary (experimental group) or facilitator-led debriefing (control group) after an IVR hospital simulation. Knowledge was measured using a multiple-choice questionnaire at baseline (T0) and post-intervention (T1), while self-reported non-technical skills were assessed at T1. RESULTS A total of 178 students participated (experimental n = 102; control n = 76). Baseline knowledge scores were comparable (experimental: M = 10.98, SD = 2.27; control: M = 11.03, SD = 2.56; p = .899). Knowledge increased significantly from T0 to T1 in both groups (experimental: M = 12.71, SD = 2.71, d = 0.79; control: M = 12.68, SD = 2.44, d = 0.72; both p < .001). After adjustment for baseline knowledge, there was no significant between-group difference in post-intervention knowledge (F = 0.03, p = .871). Non-technical skill domain scores did not differ significantly between groups (all p = .272-0.876). CONCLUSIONS GenAI-assisted debriefing produced outcomes comparable to facilitator-led debriefing for knowledge and self-reported non-technical skills. These findings warrant further investigation of GenAI-assisted debriefing as a complement to existing approaches.