In-hospital medical emergencies on general wards present critical challenges that require immediate, coordinated actions from multidisciplinary healthcare teams. Effective management of such emergencies demands not only clinical expertise but also seamless collaboration among physicians, nurses, and other healthcare professionals. In high-pressure environments, such as surgical wards, the need for timely interventions and clear communication is paramount to ensuring optimal patient outcomes. However, these emergencies are often unpredictable and can overwhelm staff, underscoring the importance of effective training to improve preparedness, enhance competencies, and reduce the risk of errors (Eisenmann et al., 2018).

Traditional interprofessional education often struggles to foster effective communication and teamwork across disciplines.

From an educational perspective, interprofessional learning aims to enable healthcare professionals to learn with, from, and about each other in order to develop shared mental models, role clarity, and effective communication strategies for collaborative practice. Such competencies are particularly critical in acute care settings, where team performance directly influences patient safety and outcomes (Gilbert et al., 2010; Reeves et al., 2017).

Therefore, it is essential to provide early and targeted opportunities to develop interprofessional collaboration skills. While simulation-based training has shown considerable benefits, it requires substantial logistical and financial resources, particularly when multiple professional groups are involved. Moreover, effective interprofessional collaboration has been demonstrated to reduce in-hospital mortality rates, increase job satisfaction, improve patient safety, and lower overall healthcare costs (Friese and Manojlovich, 2012; Nadzam, 2009; Ajeigbe et al., 2013; Schraeder et al., 2001).

In recent years, virtual reality (VR) training has emerged as a promising approach in medical education, offering immersive and interactive simulations of clinical scenarios, including high-stress emergency situations (Mao et al., 2021). VR enables a safe, controlled environment in which healthcare professionals can train without risking patient harm. This innovative format bridges the gap between theoretical knowledge and clinical practice, fostering critical skills such as teamwork, clinical decision-making, and procedural competence. Additionally, VR provides a scalable and resource-efficient solution, with the added benefits of location-independent learning and the capacity to simulate a wide range of clinical emergencies (Mao et al., 2021; Barteit et al., 2021).

Despite the increasing integration of VR into medical education, particularly in emergency medicine, there is currently no standardized framework that defines quality criteria or learning objectives for interprofessional VR-based emergency training. Although VR has demonstrated benefits in the acquisition of medical skills and knowledge, assessing teamwork within these environments remains a challenge (Bracq et al., 2019; Rosen et al., 2008; Rad et al., 2020). Recent efforts have begun to explore key components of gamification and digital learning environments (Wang et al., 2024), yet a structured consensus on quality standards specifically tailored to VR-based interprofessional emergency training is still lacking. Existing evaluation tools, originally developed for physical simulations, may not fully capture team dynamics in VR due to differences in communication modalities, interaction mechanics, and the altered role of observers (Neher et al., 2025).

Because interprofessional VR-based emergency training involves complex clinical decision-making, teamwork processes, and profession-specific perspectives, the development of shared learning objectives and evaluation criteria requires a structured consensus process integrating multiple professional viewpoints.

To address this gap, structured consensus-building methods such as the Delphi technique offer a systematic approach to informing the development of consensus-based recommendations. The Delphi method describes a structured, iterative process that facilitates expert consensus on specific topics by soliciting the opinions of a panel of knowledgeable individuals. Typically conducted over multiple rounds, this process allows participants to anonymously share their insights, review feedback from others and refine their positions until a shared understanding or agreement is reached.

A modified Delphi procedure is a variation of the traditional Delphi method and typically involves fewer rounds, may include face-to-face discussions, or employ different feedback channels, making the process more flexible and better suited to specific needs or constraints. The goal remains to gather expert insights and identify agreement on a topic (Boulkedid et al., 2011; Shang, 2023).

Delphi studies have long been used in medical education and clinical practice to build expert consensus (for example, on curriculum development or assessment standards) (de Villiers et al., 2005; Adamowski et al., 2008). While other approaches to validating VR-based training programs exist—such as user testing, expert review, or psychometric validation (Chuan et al., 2023; Kim et al., 2023; Tronchot et al., 2021; Mori et al., 2022; Knudsen et al., 2023)—the Delphi method has also been applied, though still infrequently, to the design and evaluation of VR applications, with growing interest in recent years (Kim and Kim, 2023; Ahmad et al., 2024). Some studies have further explored the use of Delphi methods to validate VR in medical contexts, particularly in aligning educational content with clinical relevance (Knight et al., 2018; Palter et al., 2012; Caetano et al., 2025).

Despite its suitability, there is a lack of published research applying Delphi validation specifically to interprofessional emergency scenarios in VR and empirical data in this area remain rare (Neher et al., 2025; Lehmann et al., 2025; Soltan and Kim, 2016).

Despite increasing interest in VR-based training for medical education, there is a lack of structured, consensus-based frameworks that define interprofessional learning objectives, critical actions, and evaluation criteria for complex ward-based emergency scenarios. To address this gap, the present study applied a modified interprofessional Delphi approach to systematically integrate expert perspectives from medicine and nursing and to develop a shared evaluation-oriented framework for VR-based interprofessional emergency training. By focusing on both clinical management and teamwork processes, this study aims to support the structured development, evaluation, and future refinement of VR-based training scenarios in interprofessional medical-surgical emergency care.

Sociodemographic characteristics of the Delphi panelists are summarized in Table 1. The final Delphi panel consisted of 17 experts, including nursing educators (n = 9, 52.9%) and physicians (n = 8, 47.1%), reflecting an interprofessional composition. The mean age of participants was 42.4 years (IQR 36.5–48), with the majority of panelists aged between 31 and 50 years. Most participants were male (76.5%). Overall, the panel represented experienced healthcare professionals from both nursing and medical backgrounds relevant to ward-based emergency care.

During Round 1, participants identified key themes relevant to interprofessional emergency training, including the importance of effective communication, timely decision-making, and clear role delineation between physicians and nurses. Experts emphasized the need for realistic scenarios that reflect real-life ward emergencies and highlighted the educational value of interactive elements such as role-playing and collaborative decision-making.

In Round 2, following hands-on engagement with the VR training simulation, participants further refined the proposed learning objectives and evaluation criteria. Particular emphasis was placed on time-critical management, accurate execution of medical procedures under pressure, and the definition of role-specific responsibilities within interprofessional emergency scenarios.

In Round 3, participants rated all 157 items using a five-point Likert scale. The resulting consensus levels for the general statements are summarized in Table 2, demonstrating a broad range of agreement. The majority of items achieved high or very high consensus, including several statements with unanimous or near-unanimous endorsement. A smaller number of items showed moderate or low levels of consensus, indicating areas of differing expert perspectives. The complete list of all rated statements is provided in Supplementary Table A1 in the appendix to ensure transparency and traceability of the Delphi process.

Consensus outcomes varied across item categories and are summarized in Tables 2–7. Items assigned to Learning Objectives showed the strongest agreement, with no statement rated as low consensus and most achieving very high consensus, including multiple items with 100% agreement (Table 7). Similarly, items related to the Evaluation of Key Actions and Evaluation of Timeline predominantly reached high or very high consensus, reflecting strong expert alignment on essential clinical actions and time-critical performance criteria (Tables 5, 6).

In contrast, items categorized under Role/Team and Process demonstrated a broader distribution of consensus levels, with a higher proportion of statements rated as moderate or low consensus (Tables 3, 4). These categories primarily addressed more context-dependent aspects of interprofessional collaboration, task allocation, and workflow organization.

Across all categories, a total of 76 items were classified as having low (n = 35) or moderate (n = 41) consensus, defined by an interquartile range (IQR) > 1, a percentage of agreement <80%, or a median <4. The distribution of overall consensus levels after Round 3 is illustrated in Figure 6.

Exploratory subgroup analyses comparing ratings between physicians and nursing educators were conducted to identify potential profession-specific patterns of agreement or disagreement. Given the limited panel size and the item-level nature of these comparisons, results should be interpreted cautiously and are not intended to support confirmatory conclusions.

As most item response distributions did not meet assumptions of normality, non-parametric Mann–Whitney U tests were applied consistently across subgroup comparisons. Items showing a difference in median ≥1, or a difference in median = 1 combined with an IQR > 1 in either professional group, were selected for further analysis. Results of these analyses after Round 3 are presented in Table 8. For eight items, exploratory subgroup analyses indicated statistically significant differences in ratings between professional groups (Table 8).

Non-consensus items identified after Round 3 were discussed in Round 4 and re-evaluated in Round 5. Revised statements are indicated by the addition of “0.1” to the item number. Following this final rating round, new consensus was achieved for 50 additional items, with 25 reaching high and 25 reaching very high consensus, including 12 items with unanimous agreement (Table 9).

Subsequent exploratory subgroup analyses after Round 5, using the same statistical approach, identified statistically significant differences for four items (items 56.1, 63, 64, and 65), as illustrated in Figure 7 and Table 10.

In total, after completion of all five Delphi rounds, overall consensus was achieved for 131 out of 157 items (84%). No consensus was reached for 26 items (16%), including 8 items with low and 18 items with moderate levels of agreement. Notably, 17 of these non-consensus items demonstrated an improvement in agreement levels between Rounds 3 and 5. The final distribution of consensus levels across all items is shown in Figure 8.

Virtual reality (VR) is increasingly recognized as a valuable tool in medical education, particularly for training in complex, high-stakes scenarios. Interprofessional VR-based training programs offer the potential to realistically simulate collaborative emergency care settings, providing both medical and nursing students with the opportunity to train essential teamwork and clinical decision-making skills. However, the validation of such programs remains a methodological challenge, especially when both medical accuracy and interprofessional learning objectives must be evaluated in tandem.

Several alternative approaches to validating VR-based training have been described in the literature. Chuan et al. validated a VR-based simulation for ultrasound-guided regional anesthesia through performance scores and comparisons with traditional simulation and real-life benchmarks (Chuan et al., 2023). Kim et al. examined a robotic surgery VR curriculum using user satisfaction surveys as the primary validation metric (Kim et al., 2023). In contrast, Tronchot et al. relied on experienced surgeons informally interacting with a VR arthroscopy simulation to establish a de facto benchmark for realism and usability (Tronchot et al., 2021). Similarly, Mori et al. validated a VR laparoscopic surgery program by comparing performance between expert and novice users (Mori et al., 2022). Knudsen et al. assessed perceived cognitive workload in a VR emergency medicine setting using the NASA Task Load Index (Knudsen et al., 2023).

While these studies demonstrate the diversity of available validation methods, they also highlight significant limitations. Many rely heavily on subjective measures (e.g., satisfaction ratings or self-assessed workload), informal expert opinion, or simple performance comparisons without a clear framework. These approaches often lack transparency, reproducibility, and generalizability. In contrast, structured Delphi methods offer a systematic, consensus-based method involving diverse expert panels, making them particularly well-suited to validating complex educational interventions like interprofessional VR training.

The Delphi method has a well-established role in medical education. Adamowski et al. used it to design psychiatry teaching content (Adamowski et al., 2008), while De Villiers et al. applied the technique to identify core competencies for medical practice in the African context (de Villiers et al., 2005). These examples illustrate the method’s utility in defining educational priorities across diverse domains.

In the field of VR-based training, Delphi studies are increasingly being used to define the scope and structure of programs. Basri et al. conducted a Delphi study to identify meaningful social skills for autism training in VR (Ahmad et al., 2024), while Kim et al. used it to determine content for an English language learning simulation (Kim et al., 2022).

More specifically, the use of Delphi procedures in VR-based medical education has been demonstrated in several surgical contexts. Knight et al. applied a Delphi approach to define the essential procedural steps to be included in a VR simulation for total laparoscopic hysterectomy (Knight et al., 2018). Palter et al. used a similar method for colorectal surgery training, involving expert consensus on the tasks and structure of the VR simulation (Palter et al., 2012). Caetano et al. employed Delphi methodology to structure a VR-based cognitive training program for individuals with substance abuse disorders (Caetano et al., 2025).

While these studies underline the growing relevance of Delphi methodology in the development of VR-based medical training, only a few have addressed interprofessional scenarios. Neher et al. developed an interprofessional training program involving medical and nursing students that included a VR component; however, validation relied primarily on pre/post questionnaire results rather than a structured expert process (Neher et al., 2025; Neher et al., 2024). The work by Lehmann et al. comes closest to the present study, describing the adaptation and validation of the Team Performance Observation Tool using expert consensus via a modified Delphi method, along with reliability and validity testing in recorded VR sessions (Lehmann et al., 2025). However, no detailed results on the Delphi procedure or the resulting consensus have been published to date.

Against this backdrop, the present study addresses an identified gap by applying a structured, interprofessional modified Delphi study to expert-informed recommendations and learning objectives for a VR-based emergency training scenario involving medical and nursing students. This process included five iterative rounds: an initial playthrough of the VR scenario by medical and nursing education experts, followed by anonymous rating of predefined statements, interprofessional expert discussion, and subsequent re-evaluation through another round of anonymous rating. This combination of experiential immersion, structured consensus-building, and iterative feedback ensured a comprehensive validation process.

Furthermore, the planned application of the VR training program incorporates user experience and perceived workload feedback from students—capturing subjective validation metrics comparable to those used in other studies (e.g., NASA Task Load Index, satisfaction scores). Thus, this study not only introduces a rigorous Delphi-based validation approach but also integrates all major elements of previously described validation methods, including expert benchmarking, comparative assessment, experiential feedback, and structured consensus. In doing so, it proposes a structured and multi-perspective validation approach that integrates expert consensus with experiential and comparative elements.

Of the 157 predefined statements evaluated during the Delphi study, 113 achieved high or very high consensus, while 44 items showed moderate or low levels of agreement, indicating areas of divergent expert perspectives. Items that achieved a very high or high level of agreement were primarily related to key learning objectives, such as the consistent application of the ABCDE approach by both professional groups, the realism and pedagogical value of the scenario design (e.g., handover structure, time-sensitive diagnostics and interventions), and the recognition of essential clinical actions and decision-making criteria under emergency conditions. Additionally, high agreement was found for statements describing fundamental principles of interprofessional teamwork, including shared leadership responsibilities, respectful communication, closed-loop handovers, and structured coordination under time pressure.

Statements that received only moderate or low agreement tended to address more nuanced or context-dependent aspects of the simulation, such as interpretations of leadership hierarchy (e.g., whether physicians should generally lead), subtleties of communication style (e.g., the importance of tone or verbal contribution), and indirect indicators of performance (e.g., whether fast completion time or prior acquaintance among team members correlates with good teamwork). This variation in consensus likely reflects differences in educational philosophy, clinical practice norms, and perceived realism among experts from diverse professional backgrounds. Nevertheless, the overall pattern demonstrates a strong alignment on core interprofessional competencies and clinical performance indicators considered relevant for VR-based emergency training. Statements that received very high or high consensus primarily addressed the educational design, scenario structure, and the applicability of the xABCDE (eXsanguination, Airway, Breathing, Circulation, Disability, Exposure) approach (Soltan and Kim, 2016), as well as the expected clinical and team-based behaviors in emergency situations. In contrast, items with lower agreement often dealt with nuances of leadership roles, subtle communication styles, and edge-case interpretations of simulation outcomes. This distribution suggests a strong expert alignment on core learning objectives and essential actions, while revealing some variability in attitudes toward leadership dynamics and communication formality within interprofessional teams.

The lack of disagreement among participants in most items suggests that the identified learning objectives and recommendations were broadly supported by the participating experts. This consensus-based approach provides a solid foundation for the development of a comprehensive VR training program that is both medically accurate and effective in enhancing interprofessional teamwork and emergency response skills.

The findings of this study should be interpreted in light of the specific context in which the Delphi process was conducted. The expert panel was recruited from a single academic medical center and the consensus process was based on one defined VR-based emergency scenario. Accordingly, the resulting framework reflects expert perspectives shaped by this institutional and scenario-specific context.

The resulting case report form (CRF) comprises a comparatively large number of items. This breadth reflects a deliberate design choice to comprehensively capture interprofessional emergency performance across multiple dimensions, including clinical assessment, diagnostic and therapeutic actions, teamwork processes, communication, time management, and learning objectives.

Importantly, the CRF is not intended to be applied as a rigid checklist in routine educational settings. Rather, it should be understood as a modular reference framework. For day-to-day VR training and formative teaching, a reduced subset of core items, focusing on key actions, fundamental teamwork principles, and critical decision points, can be selected, while the full CRF remains valuable for research purposes, curriculum development, and in-depth performance analysis. To facilitate practical interpretation and implementation, the final CRF items can be conceptually grouped into overarching competency domains. Table 11 provides a concise overview of these core domains derived from the interprofessional Delphi process.

Apparent redundancies within the CRF, such as repeated references to the ABCDE approach, were intentionally retained. These items address the same core competency from different educational perspectives, including learning objectives, observable clinical behavior, and assessment criteria, thereby enabling vertical alignment between teaching goals, scenario design, and evaluation.

From an implementation perspective, the comprehensive structure of the CRF also enables future automated or semi-automated evaluation approaches. Ongoing development work aims to link CRF items to algorithmic or AI-supported analysis of user actions within the VR environment, allowing systematic performance evaluation and the generation of structured, individualized feedback without the need for manual item-by-item scoring.

Given the interprofessional nature of the training scenario, it was essential to include both medical and nursing perspectives in the Delphi method to ensure a balanced and comprehensive definition of learning objectives. To explore potential differences in opinion between the professional groups, we analyzed the ratings after Rounds three and five separately for nursing and medical educators. After Round three, statistically significant differences between groups (p < 0.05) were found for eight items. By Round five—following an interprofessional discussion—this number was reduced to four, suggesting that the consensus-building discussion effectively aligned perspectives.

Interestingly, for all items showing significant differences in both rounds, nursing educators consistently rated the statements higher than their medical counterparts. These items included aspects such as recognizing non-physician leadership, responding to incomplete handovers, and the consequences of incorrect assumptions in clinical reasoning. This trend may reflect differing educational priorities or clinical experiences between the professions and underlines the importance of addressing interprofessional disagreement explicitly within simulation-based training. In clinical practice, such discrepancies can influence task delegation, perceived responsibility, and ultimately patient outcomes.

While the overall number of statistically significant disagreements was small and decreased over time, the presence of such differences highlights the value of a structured, dialogue-based Delphi approach in surfacing and reconciling divergent perspectives. However, the small sample size (n = 9 nursing educators, n = 8 physicians) limits the generalizability of these findings. Future research with larger expert panels and across additional institutions will be necessary to confirm these trends and further refine interprofessional learning objectives in VR-based training contexts.

Several limitations of this study should be acknowledged and considered when interpreting the scope and applicability of the presented findings. First, the number of participants in the Delphi panel was relatively small, although comparable to other expert-based Delphi studies in medical education. Second, the study was conducted at a single academic center, which may limit the transferability of the findings to other institutional contexts. Third, the modified Delphi design included face-to-face discussion rounds. While such interactions are well accepted within modified Delphi methodologies, they may introduce group dynamics that differ from fully anonymous procedures and could have influenced individual opinions. Taken together, these limitations indicate that the derived framework should be understood as context-specific and exploratory, providing expert-informed guidance rather than definitive or universally applicable standards.

This study successfully utilized an interprofessional Delphi procedure to establish consensus on key learning objectives, best practices, and common mistakes in VR-based interprofessional training for surgical emergency situations. The high level of agreement among physicians and nursing educators underscores the importance of interprofessional collaboration in developing effective and practice-relevant training programs. By including both professions, the study ensured a comprehensive perspective that captures the full range of knowledge and skills required in real-world clinical emergencies.

Key components identified through the consensus process include clearly defined roles, effective communication, and structured time management—all of which are critical for coordinated responses in acute care settings. At the same time, the study helped highlight common pitfalls, such as miscommunication, leadership ambiguity, and delays in task execution, that training curricula must address to improve clinical preparedness and reduce the risk of error.

To date, no standardized framework exists for designing and evaluating VR-based interprofessional training in emergency medicine. While previous studies have introduced validation approaches and emphasized the need for interprofessional education, none have defined the essential quality criteria and learning objectives in a structured, consensus-based manner.

This study applies a modified Delphi approach to the development of a VR curriculum for in-hospital surgical emergencies involving interprofessional teams, an area that has been sparsely addressed in previous research. The iterative, feedback-driven procedure with medical and nursing educators resulted in a structured catalog of key actions and competencies, establishing a quality-assured foundation for future training interventions.

This catalog may serve as a reference framework to inform curriculum development, implementation, and evaluation in comparable educational contexts. Furthermore, the study illustrates the value of structured Delphi methods in facilitating meaningful interprofessional dialogue and aligning educational content with real-world practice requirements.

Future studies should focus on the implementation and evaluation of the developed training program in clinical education settings to assess its impact on team dynamics, communication, and learning outcomes. In addition, applying this framework in different institutional settings, with diverse expert panels and alternative VR scenarios, will be important to further explore its transferability and robustness.

In conclusion, this study demonstrates the feasibility of using a structured, interprofessional Delphi approach to derive consensus-based learning objectives and evaluation criteria for VR-based emergency training. By incorporating medical and nursing expertise, the resulting framework provides a transparent and reproducible reference that may support the future development and evaluation of interprofessional VR training scenarios.