Background: Nursing students experience substantial academic and clinical demands that may impair well-being. However, longitudinal evidence on the roles of perceived stress and resilience remains limited.Purpose: To examine the longitudinal associations of resilience and perceived stress with well-being among nursing students.Methods: A longitudinal study of 181 nursing students in southern Taiwan assessed well-being, resilience, and perceived stress annually using validated instruments. Multiple linear regression and generalized estimating equations were used for analysis.Results: Perceived stress was consistently negatively associated with well-being, whereas resilience was positively associated with well-being, particularly in later years. Longitudinal analyses indicated that perceived stress remained a significant negative predictor, while resilience showed stronger protective effects over time. Well-being declined significantly in years 4 and 5 compared with year 2.Conclusions: Perceived stress is a stable risk factor for reduced well-being, whereas resilience plays an increasingly protective role among nursing students.
Background: Empathic communication is a core competency in nursing practice, yet providing nursing students with realistic opportunities to develop this competency remains challenging. Recent advances in artificial intelligence (AI) have enabled the integration of behavioral and physiological data into simulation-based education. Objective: To evaluate the effectiveness of a multimodal AI-supported simulation in improving empathy, communication self-efficacy, and learning engagement among undergraduate nursing students. Methods: A quasi-experimental pretest-posttest study was conducted with 134 undergraduate nursing students (68 experimental and 66 control). The experimental group received multimodal AI-supported simulation integrating conversational AI, facial emotion recognition, speech analysis, and physiological monitoring, whereas the control group received conventional role-play–based simulation with faculty-led debriefing. Results: After adjustment for baseline scores, the experimental group showed significantly greater improvements in empathy, communication self-efficacy, and emotional, behavioral, and cognitive engagement than the control group (all P < .001). Heart rate variability and galvanic skin response showed significant group-by-time differences, indicating differential physiological responses during the simulation. Qualitative findings indicated enhanced empathy awareness, communication confidence, and reflective learning. Conclusions: Multimodal AI-supported simulation improved empathy, communication self-efficacy, and learning engagement by providing integrated behavioral and physiological feedback. This approach may serve as a complementary educational strategy to support empathic communication training in undergraduate nursing education.
Background: Generative artificial intelligence (GenAI) is increasingly used in nursing education. Its educational effects in nursing remain uncertain. Purpose: To evaluate the effects of using GenAI in teaching nursing. Methods: PubMed, Embase, Web of Science, and the Cochrane Library were searched from inception to July 2026. Standardized mean differences and 95% confidence intervals (95% CI) were pooled using random-effects models. Heterogeneity, sensitivity analyses, meta-regression, and small-study effects were explored. Results: Seventeen studies involving 1456 participants were included. GenAI-integrated teaching was associated with higher knowledge scores (standardized mean differences 0.53, 95% CI 0.03-1.04), practice-related scores (SMD 1.25, 95% CI 0.23-2.28), and teaching satisfaction (SMD 0.58, 95% CI 0.02-1.13). Attitudes, competency, and self-efficacy did not differ significantly. Meta-regression findings were exploratory. Conclusions: GenAI-integrated teaching may improve knowledge, practice-related performance, and satisfaction in nursing education. However, extreme heterogeneity and sensitivity to individual studies limit confidence in the magnitude and reproducibility of these effects. GenAI should be used as a supervised pedagogical adjunct pending high-quality trials.
