Virtual reality enhanced training for trauma-informed care: A pre–post evaluation in residential and child mental health education. (Preprint)

Background: Trauma-informed care (TIC) is an organisational and clinical framework that embeds safety, trust, choice, collaboration, and empowerment into routine practice. Although training improves TIC attitudes, implementation in European residential child and adolescent settings is uneven. Immersive technologies such as virtual reality (VR) may accelerate experiential learning but remain under-evaluated for TIC. Objective: To evaluate whether a VR-enhanced training programme improves attitudes toward TIC among staff and trainees in residential child and adolescent care (including a child psychiatric ward) across multiple European sites, and to examine site-level heterogeneity and participant factors associated with change. Methods: We conducted a multi-site pre–post evaluation within the EU-co-funded Safe4Child project. Participants completed a standardised online TIC module (4 ECTS) followed by a mentor-facilitated VR simulation (1 ECTS). Attitudes were measured immediately pre- and post-intervention using the ARTIC-10 (7-point scale; higher scores indicate stronger TIC alignment). Analyses included descriptive statistics, Wilcoxon signed-rank tests, Cohen’s d, centre-specific comparisons, multivariable linear/logistic regression, and Bayesian models with sceptical priors. Internal consistency (Cronbach’s α) was assessed. Only matched cases were analysed (n = 79) from Bulgaria, Finland, and Germany. Results: Cronbach’s α increased overall from 0.839 (95% CI 0.792–0.886) pre- to 0.894 (0.863–0.926) post-intervention. Mean ARTIC-10 rose from 5.33 (SD 1.05) to 5.57 (SD 1.20); the change was significant (p < 0.001) with a small effect (d = 0.22). Centre-specific effects differed: Germany showed a significant improvement (p < 0.001; d = 0.56), whereas Bulgaria (p = 0.796) and Finland (p = 0.186) did not. The proportion with increased scores was 80% in Germany (95% CI 61.4–92.3%), 60% in Bulgaria (38.7–78.9%), and 54.2% in Finland (32.8–74.4%). In multivariable models, the German centre remained associated with greater gain (β ≈ 0.46). Bayesian estimation yielded a mean paired change Δ = 0.24 (95% CrI 0.11–0.37); P(Δ > 0) ≈ 1.00 and P(Δ > 0.20) = 0.73. Prior trauma-related training showed a probable but uncertain additional benefit. Conclusion: A VR-enhanced TIC programme produced a small overall improvement in TIC attitudes with a moderate, context-dependent effect in Germany, suggesting that immersive training can strengthen TIC learning when aligned with local educational and practice contexts. Results highlight site heterogeneity and the likely value of foundational training before VR. Limitations include single-group pre–post design, immediate post-assessment, reliance on self-report, and modest, predominantly female sample. Larger, longitudinal, and implementation-focused studies are warranted to test durability and translate attitudinal gains into practice change.