Effectiveness of VR and traditional training in medical education for mass casualty management: an OSCE-based randomized controlled trial

Abstract

Objective

This study aimed to evaluate the effectiveness of VR-based training compared to traditional lecture-based training for medical trainees in managing MCIs, specifically focusing on road traffic accidents. The primary assessment was performed using an Objective Structured Clinical Examination (OSCE) and a theoretical knowledge test.

Methods

A randomized controlled trial was conducted with 46 medical trainees receiving emergency medicine training. Participants were randomly assigned to either a VR-based Training Group or a Lecture-based Training Group, with each group receiving a 2-hour training session on mass casualty management. The training effectiveness was evaluated through pre- and post-training knowledge tests, OSCE performance, and post-training feedback questionnaires. Statistical analyses were performed to compare the two groups.

Results

Baseline characteristics were well-matched between groups. The VR-based Training Group demonstrated significantly higher post-test scores (83.96 ± 13.11) compared to the Lecture-based Training Group (72.17 ± 20.89, p = 0.03). The learning gain was also significantly greater in the VR-based Training Group (40.26 ± 15.61) compared to the Lecture-based Training Group (28.26 ± 17.04, p = 0.02). OSCE results showed that the VR-based Training Group demonstrated better performance than the Lecture-based Training Group across all stations, with significant improvements in triage, injury assessment, and overall scene management. Additionally, feedback from the post-training questionnaire revealed that the VR-based Training Group reported greater confidence in performing critical tasks.

Conclusion

VR-based training may represent a more effective and engaging approach for teaching mass casualty management relative to traditional lecture-based methods. It may enhance both theoretical knowledge and practical skills, which could help better prepare medical trainees for high-pressure scenarios such as MCIs. As VR technology continues to evolve, its integration into medical education holds considerable potential for improving preparedness and clinical performance in emergency settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12909-026-08759-x.

Introduction

Mass casualty incidents (MCIs) are complex emergencies in which numerous injured individuals need simultaneous medical care under resource constraints [1, 2]. Among these, road traffic accidents are one of the most frequent causes [3], contributing significantly to global injury-related mortality and morbidity. Traffic collisions involving numerous casualties generate substantial challenges for healthcare systems, particularly at the initial triage and management stages, where accurate decisions and effective resource allocation are critical for improving outcomes [4]. Preparing healthcare professionals to respond effectively to such situations requires not only theoretical knowledge but also the ability to make rapid decisions and deliver care under pressure [5].

Traditional medical education methods for teaching mass casualty management have typically relied on didactic lectures and simulated case discussions. While these approaches provide foundational knowledge, they often lack the immersive and realistic environments that help students develop the essential skills needed for real-life MCI scenarios, such as rapid triage decision-making and situation assessment [6]. These limitations have prompted a growing interest in using virtual reality (VR) as an educational tool. By immersing learners in customized, highly interactive environments, VR has the potential to simulate complex emergency situations in a safe and controlled setting [7]. Importantly, VR enables students to experience a fully interactive scene, such as a simulated road traffic accident with multiple casualties, allowing them to practice triage, patient assessment, and resource management as if they were in a real-life emergency scenario [8].

In recent years, VR-based simulations have been increasingly adopted in medical education, particularly in trauma care and surgical training [9]. Previous studies have shown that immersive VR platforms can significantly improve clinical skill acquisition, procedural accuracy, and even decision-making speed under time pressure [10]. For example, VR-based trauma simulations have been associated with enhanced learner engagement, higher confidence in emergency procedures, and improved performance consistency compared to traditional training [11]. However, comparative evidence between VR and conventional instructional approaches remains limited, especially in the field of mass casualty management. Few studies have rigorously evaluated VR-based training for large-scale accident scenarios, particularly those involving road traffic injuries where decision-making processes are complex and time-critical [12].

To adequately assess the practical application of VR in such scenarios, the Objective Structured Clinical Examination (OSCE) serves as an optimal evaluation tool. OSCE provides a structured and objective assessment of clinical decision-making, procedural skills, and critical thinking in a variety of medical contexts [13]. By developing OSCE stations that reflect key components of mass casualty management, such as triage, initial stabilization, and injury assessment, we can effectively evaluate the knowledge and skills gained through VR-based learning relative to traditional approaches. Previous studies support the use of OSCE in assessing skills beyond theoretical knowledge, as it provides performance-based data that is crucial in evaluating readiness for real-world practice [14].

This study aims to compare the effectiveness of VR-based training against traditional lecture-based instruction in teaching mass casualty management, with a focus on road traffic accident scenarios. Using OSCE as the primary assessment method, we evaluate medical trainees’ performance in critical tasks such as scene assessment, triage decision-making, and injury evaluation. We hypothesize that students exposed to fully immersive VR simulation will outperform their peers who underwent conventional training in terms of clinical performance and decision-making abilities in a simulated mass casualty setting.

Methods

Study design and setting

This study is a randomized controlled trial (RCT) designed to compare the effectiveness of virtual reality (VR)-based training versus traditional lecture-based training for teaching medical trainees mass casualty management in the context of road traffic accidents. The primary endpoint involves assessing students’ performance in simulated mass casualty scenarios using an Objective Structured Clinical Examination (OSCE) and a written examination to evaluate theoretical knowledge. Additionally, self-reported satisfaction and learning experiences will be collected through post-training questionnaires.

The study will be conducted at Guangxi Zhuang Autonomous Region People’s Hospital, located in Nanning, China, between January 2024 and October 2024. The participants are medical trainees who are receiving emergency medicine training during their hospital-based studies. Each educational session—whether VR-based or lecture-based—lasts for two hours, and all participants will be given the same amount of class time.

Participants are randomly assigned to either the VR-based Training Group or the Lecture-based Training Group, with the allocation ratio being 1:1. Both groups are exposed to the same curriculum content focused on managing mass casualty incidents resulting from road traffic accidents. The key differences lie in the teaching methods: the VR-based Training Group engages with an immersive, dynamic simulation of a mass casualty scenario, while the Lecture group receives conventional didactic instruction on the same topic. We hypothesized that students receiving VR-based training would achieve higher post-training OSCE scores and greater self-reported satisfaction compared with those receiving lecture-based instruction.

Participants

A total of 46 participants were enrolled in the study. All participants were medical trainees—either senior medical trainees in their clinical rotation or junior doctors—who were receiving emergency medicine training in the emergency department. Inclusion criteria required students/doctors to be actively participating in emergency medicine and mass casualty training as part of their ongoing professional development. Participants were excluded if they had previous formal training or significant prior experience in mass casualty management or if they reported conditions (such as severe motion sickness) that could affect their ability to utilize virtual reality equipment.

All eligible participants were randomly assigned to either the VR-based Training Group or the Lecture-based Training Group in an equal 1:1 ratio, with 23 participants in each group. Prior to participation, each individual provided written informed consent after receiving an explanation of the study’s objectives, procedures, risks, and benefits. Participation was voluntary, and participants were free to withdraw at any time without penalty. Care was taken to ensure that the demographic characteristics were balanced between the two groups to reduce potential confounding variables. The participants’ confidentiality and data privacy were maintained in accordance with hospital policies and research ethics guidelines. This study was conducted in accordance with the Declaration of Helsinki. Ethical approval was obtained from the Ethics Committee of the Guangxi Zhuang Autonomous Region People’s Hospital (approval number: KY-SY-2023-014). The study was subsequently registered in the Chinese Clinical Trial Registry (ChiCTR2500096725) with a registration date of February 5, 2025.The random allocation sequence was generated using a computer-based random number generator, and group assignments were concealed in sequentially numbered, opaque, sealed envelopes to ensure allocation concealment. The random sequence was generated by an independent statistician who was not involved in participant enrollment or training. Participants were enrolled by the study coordinator and assigned to intervention groups according to the concealed allocation sequence. A post-hoc power analysis was conducted using GPower version 3.1. Based on the observed difference in OSCE scores between the VR and lecture groups (effect size d = 0.6), with α = 0.05, the achieved statistical power was 0.81. This indicated that the sample size of 46 participants (23 per group) was sufficient to detect a medium effect size. This process ensured complete allocation concealment and minimized the risk of selection bias. The full study protocol is available from the corresponding author upon reasonable request.

Training intervention

Participants in both the VR and Lecture-based Training Groups received the same core curriculum content focusing on mass casualty management due to road traffic injuries. Each training session lasted 2 h. The key distinction between the two groups was the method of content delivery.

VR-based training group

Participants in the VR-based Training Group underwent immersive training using the HTC VIVE virtual reality system. The program—Road Traffic Injury VR Software 1.0—was developed by the Emergency Department of Guangxi Zhuang Autonomous Region People’s Hospital. It was designed to recreate a realistic mass casualty incident (MCI) caused by a multi-vehicle traffic collision.

Each participant trained individually within the VR environment, under supervision from an instructor who ensured proper use of the equipment and guided the learning process. The session was organized into three sequential modules corresponding to key phases of MCI management: (1) scene assessment and initial response, (2) triage and prioritization, and (3) detailed injury evaluation and initial treatment. Each training class was organized in small groups of three to five students, using three HTC VIVE headsets simultaneously. This small-group format was intentionally designed during the early stage of study planning to allow close instructor supervision and timely correction while maintaining individualized immersion.

The VR environment presented multiple virtual patients with diverse injuries (e.g., head trauma, hemorrhage, fractures, respiratory distress). Using VR motion controllers, trainees performed actions such as airway management, bleeding control, and resource allocation under time constraints.

The program included interactive instructional prompts and immediate feedback: when a participant made an incorrect triage decision or procedural error, the system provided real-time cues and short text/audio explanations to reinforce correct decision-making. This design encouraged learning through guided experiential practice, combining trial-and-error with stepwise guidance.

Upon completing the simulation, a structured debriefing session (10–15 min) was conducted by an instructor. The debrief focused on key learning points, common errors, and optimal triage strategies, encouraging reflective discussion. The total duration of each session, including the debriefing, was approximately two hours. No adverse effects such as nausea, dizziness, or visual discomfort were reported during or after VR use, indicating the session length was well tolerated. Theoretical concepts covered during the VR session were identical to those presented in the lecture-based group, ensuring content equivalence across both formats.

Lecture-based training group

The Lecture-based Training Group received instruction through standard classroom-based, didactic lectures, which covered the same learning objectives as the VR-based Training Group, namely, triage, injury evaluation, and mass casualty management principles applied to road traffic incidents. The 2-hour session involved PowerPoint presentations, instructor-led discussions, and case studies. However, no immersive or interactive elements were available to the participants in this group, and their engagement with the content was largely passive, relying on the conceptual understanding provided through the instructor.

Both groups received equal time for exposure to the curriculum direction and were evaluated through the same methods of assessment post-training.

Evaluation methods

Three OSCE stations were selected to comprehensively represent the key stages of mass casualty management (scene evaluation, triage, and injury assessment) while ensuring feasibility within the training schedule. To assess the effectiveness of both training methods, students’ knowledge and skills were evaluated using three assessment tools: a written knowledge test, an Objective Structured Clinical Examination (OSCE), and a post-training feedback questionnaire.

Written knowledge test

Students in both groups were administered a written multiple-choice exam both before and after the training as a measure of their theoretical knowledge of mass casualty management. The test consisted of 20 questions designed to assess competency in areas such as triage protocols, injury identification, and emergency management according to standardized mass casualty guidelines. The time allocated for completing each test was 30 min. The maximum possible score was 100 points.

OSCE stations

A structured clinical skills examination (OSCE) was conducted after the training to evaluate students’ practical abilities in managing mass casualty scenarios. The OSCE comprised three stations, each simulating a different key component of mass casualty management based on the road traffic accident scenario and focusing on different clinical tasks. Each OSCE station was scored out of 30 points, with a total possible score of 90 points across all three stations.

Station 1: scene evaluation and initial management

Students were introduced to a simulated mass casualty event with multiple victims. They were expected to conduct an initial scene evaluation, identifying the most critically injured patients and deciding on immediate life-saving measures such as airway management, hemorrhage control, or vascular access. Emphasis was placed on their ability to correctly assess the situation and prioritize treatments within the context of limited resources.

Station 2: triage

This station focused entirely on triage, where students had to classify simulated victims using a standardized system (such as START triage) based on their severity of injury. Students were required to justify their categorizations (e.g., red—immediate, yellow—delayed, green—ambulatory, black—deceased) within a time-limited scenario. Resource allocation and victim prioritization were components of the assessment.

Station 3: detailed injury evaluation

In this station, students had to carry out a thorough assessment of an individual simulated patient with severe injuries (e.g., traumatic brain injury, chest trauma), determine a diagnosis, and outline the advanced management steps necessary to stabilize the patient. Specific aspects such as monitoring vital signs, intravenous access, and creating a treatment plan were evaluated.

Post-training feedback questionnaire

Following the completion of both the lecture-based and VR-based training, students were asked to complete a Likert-scale questionnaire evaluating their experience. The questionnaire was developed specifically for this study and is provided as a supplementary file. Questions assessed their satisfaction with the teaching method, their perceived competence in managing mass casualty scenarios, and their preferences for future training. The questionnaire consisted of eight Likert-scale questions, with scores ranging from 1 (strongly disagree) to 5 (strongly agree), and one open-ended question allowing for qualitative feedback.

(Supplementary File: Post-Training Feedback Questionnaire)

Each OSCE station was evaluated using a structured checklist developed by three senior emergency medicine educators. Each task item was rated on a 0–6 scale: full performance received 6 points, partial or incomplete performance received 3–5 points depending on accuracy and completeness, and no performance received 0 points. Two independent raters scored all participants’ performances. Both raters were blinded to group allocation, and discrepancies greater than two points were reviewed through discussion to reach consensus. Inter-rater reliability was assessed using Cronbach’s alpha to verify the internal consistency of the scoring system.

To evaluate performance change while minimizing potential practice effects, the OSCE was administered twice—once before and once after the training intervention. Both assessments consisted of the same three station structures (scene evaluation, triage, and injury assessment), but the specific case details and simulated patient scenarios were varied between the two sessions. This ensured comparable task complexity while preventing recall bias.

All OSCE assessments were conducted under standardized conditions by the same pair of blinded raters, who were unaware of whether the performance being evaluated was from the pre- or post-training session. This procedure was designed to ensure objectivity and to reduce measurement bias.

Statistical analysis

All data from the tests and assessments were collected and anonymized for analysis. Descriptive statistics were used to summarize baseline characteristics and test scores. For continuous variables, normality of the data distribution was first assessed using the Shapiro-Wilk test. Paired t-tests were used for normally distributed data, while Mann-Whitney U tests were applied to non-normally distributed data. Changes in performance across the three OSCE stations were analyzed using ANCOVA, with post-training scores as the dependent variable and pre-training scores (baseline written knowledge) as covariates to adjust for initial differences between groups. Effect sizes (Cohen’s d for continuous outcomes or partial η² for ANCOVA) and 95% confidence intervals were reported alongside p-values to provide a more comprehensive assessment of group differences.

Chi-square tests or Fisher’s exact tests were used to compare categorical responses from post-training questionnaires, such as preferences for future training methods. Reliability of the instruments was evaluated using Cronbach’s alpha for both the written knowledge test and the post-training feedback questionnaire, with α > 0.7 considered acceptable. All statistical analyses were conducted using SPSS Version 25.0 and R Version 4.3.1, under guidance from an independent statistician to ensure the appropriateness of the methods. Statistical significance was set at P < 0.05.

Result

Baseline characteristics

The flow of participants through the study is shown in Fig. 1. A total of 46 participants were assessed for eligibility, all met inclusion criteria, and were randomized equally into the VR-based Training Group (n = 23) and the Lecture-based Training Group (n = 23). All participants completed the intervention and were included in the final analysis.

Fig. 1.

CONSORT flow diagram of participant enrollment, randomization, and analysis

The baseline characteristics of the participants in both groups were comparable (Table 1). The proportion of male participants was similar between the VR-based Training Group (52.17%) and the Lecture-based Training Group (56.52%). The mean age of participants in the VR-based Training Group was 26.30 ± 1.73 years, while in the Lecture-based Training Group it was 26.26 ± 2.01 years, with no significant difference (p = 0.94). Additionally, the average work experience was 2.82 ± 1.05 years in the VR-based Training Group and 2.87 ± 1.19 years in the Lecture-based Training Group, with no statistically significant difference observed (p = 0.89). These results indicate that the two groups were well-matched in terms of gender, age, and work experience prior to the intervention (Fig. 2).

Table 1.

Baseline characteristics of participants

Characteristics	VR-based Training Group (n = 23)	Lecture-based Training Group (n = 23)	P-value
Gender, male (%)	12(52.17%)	13(56.52%)
Age (years, mean ± SD)	26.30 ± 1.73	26.26 ± 2.01	0.94
Work Experience (years)	2.82 ± 1.05	2.87 ± 1.19	0.89

Fig. 2.

Detailed views of VR training and OSCE assessments. A Simulation process within the VR training module. B User interaction and operation screen during VR exercises. C Desktop simulation for triage categorization. D Examination setup for injury assessment evaluation

Theoretical knowledge test results

The pre-training knowledge test scores were comparable between the VR-based Training Group (43.66 ± 10.24) and the Lecture-based Training Group (43.91 ± 9.99), with no significant difference observed (p = 0.94). However, after the training, the VR-based Training Group demonstrated significantly higher post-test scores (83.96 ± 13.11) compared to the Lecture-based Training Group (72.17 ± 20.89), with a p-value of 0.03. Additionally, the learning gain, calculated as the difference between pre- and post-test scores, was significantly greater in the VR-based Training Group (40.26 ± 15.61) than in the Lecture-based Training Group (28.26 ± 17.04), with p = 0.02. The effect size for the difference in post-test scores was medium (Cohen’s d = 0.67, 95% CI: 0.12–1.20), and for learning gain was similar (d = 0.71, 95% CI: 0.16–1.25), indicating that VR-based training was associated with a meaningful improvement in theoretical knowledge compared with traditional lectures (Table 2). Both groups showed significant improvement in knowledge scores after training (p < 0.001). The reliability of the knowledge test was acceptable (Cronbach’s α = 0.81).

Table 2.

Results from the pre- and post-training knowledge tests

Metric	VR-based Training Group (Mean ± SD)	Lecture-based Training Group (Mean ± SD)	P-value
Pre-Test Score	43.66 ± 10.24	43.91 ± 9.99	0.94
Post-Test Score	83.96 ± 13.11	72.17 ± 20.89	0.03
Learning Gain	40.26 ± 15.61	28.26 ± 17.04	0.02

OSCE performance results

Table 3 presents the scoring components and structure for each of the three OSCE stations, each with a maximum score of 30 points. These stations were designed to assess students’ practical skills and decision-making abilities in managing simulated mass casualty scenarios. OSCE performance was evaluated before and after training in both the VR-based Training Group and the Lecture-based Training Group. The total possible score across all stations was 90 points, as summarized in Table 4.

Table 3.

OSCE station content and scoring details

OSCE Station	Scoring Criteria (6 points each)	Maximum Points
Station 1: Scene Evaluation	Personal Protective Equipment (PPE)	6
	Assessment of Scene Safety	6
	Communication with Traffic and Fire Departments	6
	Information Gathering and Reporting	6
	Mechanism of Injury Analysis	6
Station 2: Triage	Triage Process Order	6
	START Method Selection	6
	Time	6
	Accuracy	6
	Decisiveness	6
Station 3: Initial Injury Assessment	Airway	6
	Breathing	6
	Circulation	6
	Nervous system	6
	Environment and exposure	6
Total score		90

Table 4.

OSCE scores pre- and post-training for two groups

Category	VR-based Training Group (n = 23)	Lecture-based Training Group (n = 23)	P-value
Station 1
Pre-Test Score (Mean ± SD)	10.70 ± 1.26	11.04 ± 1.39	0.39
Post-Test Score (Mean ± SD)	22.83 ± 1.74	20.57 ± 2.18	< 0.01
Station 2
Pre-Test Score (Mean ± SD)	9.26 ± 1.36	9.13 ± 1.68	0.78
Post-Test Score (Mean ± SD)	23.04 ± 1.71	20.22 ± 2.50	< 0.01
Station 3
Pre-Test Score (Mean ± SD)	10.70 ± 2.01	10.61 ± 1.58	0.87
Post-Test Score (Mean ± SD)	20.83 ± 1.63	17.83 ± 1.95	< 0.01
Total
Pre-Test Score (Mean ± SD)	30.65 ± 2.69	30.78 ± 3.29	0.89
Post-Test Score (Mean ± SD)	66.70 ± 2.45	58.61 ± 3.73	< 0.01

Before training, there were no significant differences between the two groups across all stations (all P > 0.05), indicating comparable baseline performance. After training, both groups showed significant improvement; however, the VR-based Training Group consistently achieved higher post-test scores than the Lecture-based Training Group. Specifically, post-training scores were significantly higher in the VR group for scene evaluation (22.83 ± 1.74 vs. 20.57 ± 2.18, P < 0.01), triage (23.04 ± 1.71 vs. 20.22 ± 2.50, P < 0.01), and injury assessment (20.83 ± 1.63 vs. 17.83 ± 1.95, P < 0.01).

The overall post-training OSCE total score was also markedly higher in the VR-based Training Group (66.70 ± 2.45) than in the Lecture-based Training Group (58.61 ± 3.73, P < 0.01). These findings indicate that VR-based training produced greater improvements in both operational competence and decision-making performance compared with traditional lecture-based methods (Fig. 3). The largest gains were observed in tasks requiring real-time judgment and prioritization, particularly during triage, suggesting that immersive simulation environments enhance situational awareness and cognitive processing under pressure. Inter-rater reliability for OSCE scoring was high (Cronbach’s α = 0.89), indicating excellent internal consistency between raters.

Fig. 3.

OSCE station scores before and after training for VR and Lecture-based Training Groups, highlighting significant improvements post-training. A Scores for each OSCE station—scene evaluation, triage, and injury assessment—before and after training in the VR and Lecture-based Training Groups. The full score for each station is 30 points. B Post-training Total OSCE scores of both groups, with a maximum total of 90 points. **P < 0.01 indicates significant differences in post-training scores between the VR-based training group and the lecture-based training group

Post-training feedback questionnaire results

The questionnaire results (Table 5) indicated that the VR-based Training Group showed significantly greater improvements compared to the Lecture-based Training Group in several key areas(Fig. 4). For example, 95.65% of participants in the VR-based Training Group reported enhanced ability to perform injury assessments during mass casualty incidents, compared to 82.61% in the Lecture-based Training Group (P < 0.05). Understanding of triage principles was also higher in the VR-based Training Group (86.96% vs. 47.83%, P < 0.01). Additionally, more participants in the VR-based Training Group felt confident in performing scene evaluations (78.26% vs. 39.13%, P < 0.01) and managing specific traumatic conditions (78.26% vs. 52.18%, P < 0.05). The VR-based Training Group exhibited greater improvement in decision-making under resource constraints (65.21% vs. 34.79%, P < 0.01). While no significant difference was found regarding the structure and timing of the training sessions (P = 0.59), confidence in applying gained knowledge in real-world scenarios approached statistical significance in the VR-based Training Group (100% vs. 86.96%, P < 0.05). The overall questionnaire effect size was moderate (Cohen’s d = 0.64, 95% CI: 0.21–1.06), supporting the greater perceived benefit of the VR-based training approach. The questionnaire demonstrated good internal consistency (Cronbach’s α = 0.86).

Table 5.

Post-training questionnaire results: VR vs. traditional training

	Strongly disagree	Disagree	Neutral	Agree	Strongly agree	P-value
1. The training effectively improved my ability to perform injury assessments during mass casualty incidents.
VR-based Training Group	0(0.00%)	0(0.00%)	1(4.35%)	9(39.13%)	13(56.52%)	< 0.05
Lecture-based Training Group	0(0.00%)	1(4.35%)	3(13.04%)	13(56.52%)	6(26.09%)
2. I gained a thorough understanding of the principles and application of triage in mass casualty events as a result of the training.
VR-based Training Group	0(0.00%)	1(4.35%)	2(8.70%)	10(43.48%)	10(43.48%)	< 0.01
Lecture-based Training Group	0(0.00%)	3(13.04%)	9(39.13%)	9(39.13%)	2(8.70%)
3. Following the training program, I feel confident in performing an initial scene evaluation of mass casualty incidents.
VR-based Training Group	0(0.00%)	1(4.35%)	4(17.39%)	11(47.83%)	7(30.43%)	< 0.01
Lecture-based Training Group	0(0.00%)	3(13.04%)	11(47.83%)	7(30.43%)	2(8.70%)
4. The training enhanced my ability to manage specific traumatic conditions, such as traumatic brain injuries, hemorrhages, and fractures, in a mass casualty setting.
VR-based Training Group	0(0.00%)	2(8.70%)	3(13.04%)	8(34.78%)	10(43.48%)	< 0.05
Lecture-based Training Group	0(0.00%)	3(13.04%)	8(34.78%)	10(43.48%)	2(8.70%)
5. I feel the training improved my decision-making abilities regarding resource allocation and prioritization in a constrained mass casualty scenario.
VR-based Training Group	0(0.00%)	2(8.70%)	6(26.09%)	8(34.78%)	7(30.43%)	< 0.01
Lecture-based Training Group	2(8.70%)	5(21.74%)	8(34.78%)	6(26.09%)	2(8.70%)
6. The structure and timing of the training sessions were appropriate for covering the essential principles of mass casualty management.
VR-based Training Group	0(0.00%)	0(0.00%)	0(0.00%)	10(43.48%)	13(56.52%)	0.59
Lecture-based Training Group	0(0.00%)	0(0.00%)	1(4.35%)	10(43.48%)	12(52.17%)
7. The training content provided me with a deeper understanding of the overall management strategies required in mass casualty incidents.
VR-based Training Group	1(4.35%)	3(13.04%)	7(30.43%)	8(34.78%)	4(17.39%)	0.26
Lecture-based Training Group	1(4.35%)	4(17.39%)	11(47.83%)	5(21.74%)	2(8.70%)
8. I feel confident in applying the knowledge and skills acquired from this training to real-world mass casualty scenarios.
VR-based Training Group	0(0.00%)	0(0.00%)	0(0.00%)	11(47.83%)	12(52.17%)	< 0.05
Lecture-based Training Group	0(0.00%)	0(0.00%)	3(13.04%)	13(56.52%)	7(30.43%)

Fig. 4.

Answers to questions 1–8 (Q1-Q8) as a stacked bar chart, displayed for two groups separately. Distribution of responses to questions 1–8 (Q1–Q8) presented as stacked bar charts for the VR-based and Lecture-based Training Groups

No adverse events or discomfort were reported in either group during the training or assessment process.

Discussion

This study suggests that VR-based training may offer advantages over than traditional lecture-based methods in preparing medical trainees for managing mass casualty incidents, based on the observed OSCE and knowledge test results. The VR-based Training Group’s superior performance was observed not only in written knowledge tests but also in practical skills assessments evaluated through an Objective Structured Clinical Examination (OSCE). The OSCE specifically highlighted the VR-based Training Group’s enhanced capabilities in critical real-world tasks, such as scene management, triage decision-making, and injury assessment under stress. By offering structured, objective stations designed to simulate mass casualty scenarios, the OSCE was able to capture the nuanced improvements in clinical judgment and decision-making enabled by the immersive VR training. Additionally, participants in the VR-based Training Group reported greater confidence in their ability to handle mass casualty scenarios, further supporting the potential effectiveness of VR in a highly interactive and realistic environment. These findings support our hypothesis that integrating VR simulations with OSCE-based evaluations provides a robust platform for assessing both the acquisition of theoretical knowledge and the development of practical, hands-on skills essential in emergency settings.

The findings of this study align with existing research on the benefits of VR in medical education, particularly in fields such as trauma care and Cardiopulmonary Resuscitation training, where VR has been shown to enhance procedural accuracy and decision-making under pressure [15, 16]. However, a key distinction of this study is the integration of OSCE station to evaluate practical skills in mass casualty management—an area that has received limited attention in VR-based training research. Unlike written examinations, which primarily assess theoretical knowledge, the OSCE provides a structured and standardized framework for evaluating hands-on skills and real-time decision-making in simulated emergency scenarios. Previous studies [17], have highlighted the utility of OSCEs in emergency medicine, but most relied on static, mannequin-based simulations. In contrast, our incorporation of a fully immersive VR platform adds a dynamic layer of interactivity that closely replicates real-world mass casualty events, such as multi-victim road traffic accidents. This combination of VR training with OSCE-based assessment offers a more comprehensive evaluation by not only capturing improvements in theoretical knowledge but also demonstrating superior clinical performance in high-pressure, resource-limited environments. These findings suggest that VR, when paired with OSCE, may provide a promising framework for assessing both cognitive and psychomotor skills in medical trainees.

The relatively better performance of the VR-based Training Group in this study may be partially attributed to several key factors inherent to the immersive nature of VR simulations. Unlike traditional lecture-based instruction, VR provides an interactive environment that closely mimics real-world scenarios, enabling learners to engage in critical decision-making while being exposed to realistic stressors, such as time pressure and chaotic environments [18]. This type of experiential learning fosters not only knowledge acquisition but also the development of situational awareness and practical problem-solving skills, which are crucial in mass casualty management. The immersive VR platform used in this study allowed participants to practice complex tasks—such as rapid triage and resource allocation—in a controlled yet highly dynamic setting. This interactive engagement promotes active learning, leading to better retention of information and faster reflexes when faced with decision-making in emergency contexts. Importantly, VR also offers the advantage of repeated practice without the risks associated with live patients, thus allowing students to hone their skills in a safe yet challenging environment [19]. These factors, combined with the objective assessment capabilities of the OSCE, explain the superior clinical outcomes observed in the VR-based Training Group, highlighting the potential of VR as a highly effective tool for bridging the gap between theoretical knowledge and real-world clinical application.

One of the major strengths of this study is its randomized controlled trial (RCT) design, which minimizes potential biases and ensures a balanced comparison between the VR and Lecture-based Training Groups. The random allocation of participants and the 1:1 ratio between the groups helped maintain homogeneity in baseline characteristics, such as age, gender, and work experience, thereby reducing confounding variables. Another key strength is the integration of Objective Structured Clinical Examination (OSCE) as the primary method of practical skill evaluation. The use of OSCE is widely accepted as a gold standard for objective assessment of clinical competence, and its inclusion ensures that the skills being measured—particularly in areas such as triage, scene management, and injury evaluation—accurately reflect real-world clinical demands in mass casualty settings [20]. Additionally, combining OSCE with VR for skill assessment offers a unique advantage, as it allowed participants to engage in repetitive, immersive practice while being objectively evaluated in real-time. This dual approach strengthens the validity of the study’s findings by providing both subjective self-assessment feedback and quantifiable performance metrics [21, 22]. Finally, the focus on mass casualty management, a critical yet underexplored area in medical education, enhances the study’s relevance, as healthcare systems around the world increasingly face the challenges of emergency response in complex, multi-victim scenarios.

Despite its strengths, this study has several limitations that should be acknowledged. First, the relatively small sample size (n = 46) may limit the generalizability of the findings to broader populations of medical trainees or healthcare professionals. While the results are encouraging, larger-scale studies are needed to confirm these findings and explore whether they are consistent across different training environments and demographic groups. Second, although no obvious VR-related discomfort was observed during the training sessions, VR-induced sickness was not assessed using a standardized instrument such as the Simulator Sickness Questionnaire. As a result, mild or transient symptoms may have been underreported and should be considered when interpreting participant experience. In addition, while VR offers a highly immersive experience, it cannot fully replicate all the complexities and stressors of real-world mass casualty events [23]. Factors such as emotional stress, unpredictable resource availability, and the physical demands of working in chaotic environments are not entirely captured in virtual simulations, which may affect how trainees apply these skills in actual emergency situations [24]. Furthermore, the training duration was limited to two hours, and study outcomes were assessed immediately after the intervention. This design does not allow conclusions regarding long-term knowledge retention or transfer of skills to real clinical practice. Future research should therefore include extended training exposure and longitudinal follow-up to evaluate skill durability over time. Lastly, although the VR intervention was carefully organized for this study, the feasibility of implementing two-hour immersive VR sessions for larger cohorts may vary across institutions. Factors such as cost, equipment availability, faculty support, and scheduling constraints may influence scalability, particularly in resource-limited settings, and should be considered when interpreting the findings [25].

The findings of this study emphasize the potential of VR-based training in enhancing medical education focused on emergency medicine and disaster response, particularly in the context of road traffic injuries and mass casualty incidents. Future research should explore the long-term retention of skills acquired through VR training in these high-pressure scenarios, as well as the transferability of these skills to real-world emergency settings. Longitudinal studies following graduates into clinical practice could help determine how well VR-trained students perform under actual mass casualty conditions, where factors such as emotional stress, resource scarcity, and unexpected environmental variables come into play. Moreover, expanding VR-based simulations to cover a broader variety of disaster scenarios, such as natural disasters and large-scale public emergencies, could further strengthen the preparedness of healthcare professionals in managing diverse mass casualty incidents. These simulations could incorporate more complex triage algorithms, refined injury assessment techniques, and advanced decision-making under critical resource constraints, providing trainees with comprehensive exposure to different emergency contexts [26].

For practical implementation, the combination of VR with structured evaluation methods like the OSCE offers a powerful tool for assessing both cognitive knowledge and clinical performance in disaster scenarios. Training programs in emergency medicine should consider integrating VR simulations into their core curricula, particularly for students and professionals preparing to face the growing demands of disaster response and mass casualty management. Ultimately, incorporating VR in the educational pipeline for emergency and disaster medicine could significantly improve clinical competency and decision-making capabilities, enhancing the overall responsiveness and effectiveness of healthcare systems in crisis situations.

Conclusion

This study suggests that combining VR-based training with the OSCE framework may offer an effective approach to teaching mass casualty management, particularly in areas such as triage, injury evaluation, and decision-making. The immersive nature of VR may contribute to improved practical skill acquisition, while OSCE provides a structured and objective means to evaluate both theoretical knowledge and clinical performance. Together, these tools represent a promising advance in emergency and disaster management training, offering a relatively more realistic and comprehensive learning experience compared to traditional methods.

Abbreviations

MCI

Mass casualty incidents

OSCE

Objective Structured Clinical Examination

RCT

Randomized controlled trial

SD

Standard Deviation

Authors’ contributions

Z.L. (Zhe Li) designed the study, performed data analysis, and wrote the main manuscript text. L.L. (Liwen Lyu), as the corresponding author, supervised the study, critically revised the manuscript, and ensured the integrity of the work. W.C. (Wan Chen) and G.Q. (Guozheng Qiu) contributed to data collection and interpretation. L.S. (Lei Shi) and Y.T. (Yutao Tang) assisted in interpreting the results and revising the manuscript. X.X. (Xibin Xu) and S.Z. (Sanshan Zhu) provided valuable feedback on methodology and helped refine the final manuscript. All authors read and approved the final manuscript.

Funding

This research was supported by the Guangxi Medical and Health Appropriate Technology Development and Promotion Application Project [grant number: S2023013] and the Guangxi Natural Science Foundation [grant number: 2024GXNSFAA010071]. The funding provided resources for experimental materials, data analysis software, and research assistant stipends.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki. Ethical approval was obtained from the Ethics Committee of the Guangxi Zhuang Autonomous Region People’s Hospital (approval number: KY-SY-2023-014), and the study was registered in the Chinese Clinical Trial Registry (ChiCTR2500096725).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.