JMIR Publications

ABSTRACT

Background:

Workplace violence (WPV)—including physical assaults, threats, and verbal/psychological aggression—disproportionately affects healthcare workers, particularly in high-acuity and behavioral health settings. Although many organizations provide WPV prevention education, training is often delivered through scalable didactic or online modules that may be limited in preparing staff for high-stress, interpersonal crisis encounters. Immersive virtual reality (VR) offers a safe-to-fail, experiential modality that can evoke realistic emotional responses and support repeated practice of de-escalation skills. However, evidence is limited regarding how debriefing format influences learning from VR-based WPV training.

Objective:

To evaluate the usability of DEFUSE, an immersive VR-based WPV prevention and behavioral management training, and to compare debriefing formats (individualized, in-headset VR debrief vs educator-led, in-person group debrief) with respect to changes in aggression attribution beliefs.

Methods:

Frontline clinical and nonclinical staff were recruited from three major healthcare institutions in Central Ohio. Using a quasi-experimental design with convenience allocation, participants completed a single DEFUSE training session consisting of two VR modules and one of two debriefing formats: (1) individualized, in-headset VR debriefing following each module (n = 44), or (2) educator-led group debriefing following completion of both modules (n = 47). Usability was assessed immediately after post-exposure. Beliefs toward causes of patient aggression and management were assessed using the Management of Aggression and Violence Attitude Scale (MAVAS) at pre-post exposure, and 4–6-week follow-up. Participants were also invited to semi-structured focus groups; interviews were audio-recorded and thematically analyzed. Quantitative outcomes were examined using ANCOVA and multilevel modeling.

Results:

Ninety-one participants completed the study, with 44 (52%) completing follow-up at 4–6 weeks. Overall VR usability was high (SUS-equivalent M = 78.44) and did not differ between debriefing conditions (p = .457). Baseline-adjusted ANCOVA indicated a significant advantage for the VR-based debrief on MAVAS environmental attributions at immediate post-exposure (p < .05), whereas other between-condition differences were nonsignificant. Multilevel models showed significant time effects for situational (B = 0.08, SE = 0.022, p = .001) and environmental (B = 0.10, SE = 0.029, p = .001) attributions across three assessments, while biological attributions showed no significant change over time. Condition effects were not retained in final longitudinal models. Qualitative findings indicated that emotional realism and behavioral fidelity supported experiential learning, and that individualized, emotionally paced debriefing with timely, action-linked feedback facilitated reflection and skill consolidation. Participants emphasized that scalability depends on minimizing interface friction for novice users and embedding VR training within onboarding and shift-based workflows.

Conclusions:

DEFUSE demonstrated strong usability and supported changes in staff beliefs toward more contextual understandings of patient aggression over time. The VR-based, in-headset debrief produced outcomes largely comparable to educator-led debriefing, with indications of added benefit for environmental attributions.