Methodological Quality of Reviews and Effectiveness of Training Modalities on Resuscitation Education in Healthcare

Background: High-quality resuscitation training is essential for patient safety, yet determining the most effective instructional approaches remains challenging. Resuscitation requires integrated cognitive, psychomotor, and affective competence under pressure. Virtual reality (VR) has emerged as a potential training modality, but its effectiveness among healthcare professionals, particularly nurses, is uncertain. In parallel, although systematic reviews of resuscitation training have proliferated, the comparative effectiveness of instructional designs and the methodological quality of existing reviews remain unclear. Objectives: This dissertation aimed to (1) evaluate the effectiveness of VR-based training across learning domains; (2) appraise the methodological quality of systematic reviews on resuscitation training; and (3) compare and rank resuscitation training instructional designs. Methods: A three-phase, sequential evidence synthesis approach was employed. First, a pairwise meta-analysis evaluated the effects of VR-based education among nursing staff using randomized and quasi-experimental studies identified from eight databases; standardized mean differences were estimated with random-effects models. Second, an umbrella review appraised the methodological quality of systematic reviews on resuscitation training by searching eight databases through August 2025, with quality assessed using AMSTAR 2 and findings synthesized narratively following the SWiM framework. Third, a network meta-analysis compared and ranked structured resuscitation training methods using randomized controlled trials identified from twelve databases, trial registries, and grey literature through May 2025, with analyses conducted using random-effects models. Results: In Phase 1, virtual reality (VR)–based training was associated with improvements in cognitive (SMD = 1.48; 95% CI, 0.33–2.63), affective (SMD = 0.59; 95% CI, 0.34–0.86), and psychomotor outcomes (SMD = 0.90; 95% CI, 0.49–1.31), as well as learner satisfaction (SMD = 0.47; 95% CI, 0.17–0.77), although heterogeneity was substantial. Phase 2 identified a steady increase in resuscitation education reviews (5 per year), yet none were rated as high quality; most were critically low (70.4%), with the remainder rated low (14.8%) or moderate (14.8%). In Phase 3, interactive learning, simulation-based training, VR, and high-fidelity manikins improved performance acquisition compared with traditional training (SMD range, 0.57–2.52), while simulation-based, self-directed, and booster training improved performance retention (SMD range, 0.79–1.19). Conclusion: Virtual reality (VR) shows potential as a training modality for enhancing nursing competence; however, larger, well-designed randomized controlled trials across diverse clinical settings are needed to clarify its effectiveness. The methodological quality of existing systematic reviews in this field remains predominantly low or critically low, underscoring the need for stronger adherence to established standards to ensure that training policies are informed by reliable evidence. In contrast, instructional designs that emphasize active practice and structured feedback consistently outperform traditional approaches in improving resuscitation competence. Periodic refresher training and simulation-based practice appear important for sustaining performance over time.