A total of 119 final-year medical students were recruited through in-class announcements. During lecture sessions, researchers provided an overview of the study’s objectives, procedures, risks, and benefits, allowing students to ask questions before deciding to enroll. Participation was voluntary, with no financial or material compensation provided. Interested students signed informed consent forms, which outlined confidentiality procedures and withdrawal rights. Recruitment was completed within 1 month. Participants ranged in age from 22 to 45 (mean 24.03, SD 3.4) years, with 78.4% (93/119) identifying as female. The sample was representative of the student body demographics at the institution. Prior VR experience was assessed as a binary variable (yes/no)—participants reported whether they had previously used VR for any purpose. Those with prior VR exposure (31/119, 26.1%) engaged with VR-based applications (eg, gaming, training, and immersive experiences), whereas the remaining 73.9% (88/119) had no prior VR experience.

Eligible participants were final-year medical students aged 22‐45 years without uncorrected visual impairments. Students using corrective eyewear (glasses or contact lenses) were eligible. To ensure the validity of HRV measurements and usability assessments, the following exclusion criteria were applied: (1) cardiac medication use (eg, β-blockers) due to potential effects on autonomic regulation; (2) excessive caffeine consumption (>400 mg/day, equivalent to ~4 cups of coffee), as it can alter physiological stress responses; (3) use of sympathomimetic substances (eg, decongestants and stimulants) due to potential HRV interference; (4) history of severe motion sickness, which could bias usability and simulator sickness assessments; and (5) incomplete data or poor-quality HRV recordings, ensuring data integrity. These criteria minimized confounding factors and ensured a controlled evaluation of usability and physiological responses.

The VR pericardiocentesis simulation was developed in Unity 2021.3.24f1 (Unity Technologies) with the Oculus Integration SDK. The environment included a virtual hospital room, a patient bed, vital signs monitor, and interactive procedural tools.

On top of the bed lies a male patient, and besides him are 2 easily accessible tables with the following instruments: three electrodes, a pressure cuff, a pulse oximeter, an oxygen mask, a box of gloves, a Betadine bottle, forceps with a clean gauze, a surgical cloth, a tray with a syringe, a straight endovascular device, and a bent endovascular device. All these items are interactive and should be used to perform the procedure in the proper order and places.

The vital signs monitor updates when the pressure cuff, the electrodes, the pulse oximeter, and the oxygen masks are placed on the patient, and when the procedure finishes. The data shown on the monitor are accompanied by a beeping sound, recorded from a real device. The complete setup can be seen in Figure 2.

The user does not see their own body but can see 2 virtual hands (Figure 3), which also show the surgical gloves, if appropriate. All the interactions inside the simulation were diegetic: using virtual elements present in the scenario that replicate real items, without any menu, teleportation mechanism, heads-up display, or other immersion-breaking controls (Figure 4). The user can displace themselves only by direct movement, without any teleportation of stick-controlled movement mechanisms.

The VR simulation was conducted using the Oculus Rift S headset, which was calibrated before each session to ensure accurate tracking and optimal user experience. The calibration process included adjusting the interpupillary distance, ensuring proper headset fit, and setting the default room-scale boundary to prevent tracking errors. Participants were instructed on how to adjust the headset for comfort, ensuring that the lenses were correctly aligned to minimize motion blur and discomfort. Environmental conditions were controlled to ensure consistency across all sessions. The simulation took place in a dedicated laboratory with dim lighting to reduce glare and reflections and a quiet setting to minimize distractions. Participants were allowed limited standing movement within a 3×3 m play area, which was predefined using the Guardian System of the Oculus Rift S to prevent accidental collisions. The temperature was maintained at a comfortable level to avoid physiological stress unrelated to the simulation. Each session lasted approximately 30 minutes, including headset fitting, familiarization with the controls, and the execution of the pericardiocentesis procedure. To ensure consistency across participants, all interactions were standardized, and the experimenter provided identical procedural instructions before each session.

Participants completed the VR pericardiocentesis simulation in three phases:

The study was approved by the University Hospital research ethics committee (code: 18.12.1339.GHM), following the Declaration of Helsinki. Participants provided written informed consent after a full briefing on study objectives, risks, and withdrawal rights. No financial or material incentives were provided, and all participant data were anonymized.

Spearman correlation examined relationships between HRV metrics, usability, and simulator sickness. Mann-Whitney U tests compared usability and stress responses between participants with and with no VR experience, while Friedman tests compared overall usability scores across scales. All data were securely stored and anonymized following institutional data protection guidelines. Statistical analyses were performed using IBM SPSS Statistics (version 25.0; IBM Corp), and physiological data were processed using OpenSignals software to ensure accuracy and reliability in HRV analysis.

The usability of the VR simulation was assessed using the SUS, PSSUQ, PQ, and SSQ. The internal reliability of these scales was evaluated using Cronbach α (Table 1). High reliability (α>0.8) was found for SUS (α=0.810) and PSSUQ (α=0.980), confirming their consistency in assessing usability and user satisfaction. Some subscales, such as Nausea Symptoms (SSQ; α=0.472) and Perceived Naturalness (PQ; α=0.608), exhibited lower reliability, likely due to individual variability in motion sickness susceptibility and expectations regarding VR realism.

Despite these variations, the overall reliability of the usability scales remains robust, supporting the validity of the findings. Future studies could refine these subscales or integrate biometric indicators (eg, galvanic skin response and pupil dilation) to improve measurement consistency.

Correlational analysis examined physiological stress (HRV) and simulator sickness symptoms (Figure 5):

HRV analysis revealed that higher parasympathetic activity was associated with lower simulator sickness symptoms. However, HRV metrics did not significantly correlate with usability or presence scores, suggesting that physiological stress responses were more related to physical discomfort than to system usability. No significant correlations were found between HRV (rMSSD) and usability scores, as shown in Figure 6.

Participants with prior VR experience reported lower simulator sickness and higher usability perceptions (Table 2):

Prior VR experience was linked to lower levels of simulator sickness but did not significantly affect HRV stress indicators, suggesting that the system was user-friendly across different experience levels.

The usability assessments revealed that participants rated the VR system as highly usable, with a mean SUS score of 75 (SD 6.41, 95% CI 73.42-76.58; P=.002), exceeding the commonly used benchmark of 68. This indicates that the system was well designed and accessible, allowing users to interact with it effectively. Both novice and experienced users provided comparable usability ratings, suggesting that the reliance on direct movement and the absence of nondiegetic interactions eliminated the need for prior VR navigation experience. These findings align with previous research emphasizing the importance of designing VR systems that are intuitive and inclusive, ensuring accessibility across users with varying levels of expertise [15,16]. Moreover, usability-focused approaches tailored to novice users have been recommended, highlighting the importance of gradual adaptation strategies to facilitate immersion and reduce initial cognitive load [17].

Although usability scores were high, simulator sickness remained a challenge, particularly for novice users, who reported significantly greater nausea and oculomotor discomfort than experienced users. These results are consistent with prior studies showing that previous VR exposure reduces discomfort, likely due to increased adaptation to immersive environments [18]. Adaptation mechanisms such as shorter initial exposure sessions or gradual immersion through simplified VR experiences could facilitate acclimatization for first-time users [19]. The 10-minute session duration may have contributed to minimizing simulator sickness symptoms, as shorter exposure times are recommended to optimize user comfort.

Although simulator sickness was more pronounced among novices, this study did not track symptom progression over time, limiting the ability to assess whether participants acclimated during the session. Research suggests that VR-induced discomfort often declines with repeated exposure as users adapt to the sensorimotor feedback of the environment [20]. Motion sickness symptoms typically peak within the first few minutes of VR exposure and gradually subside, particularly with repeated use. However, because symptoms were measured only at the end of the session, it remains unclear whether adaptation occurred during the simulation itself.

The high SUS score of 75 suggests that participants found the VR system not only intuitive but also effective for training. Previous research evaluating medical VR applications has reported lower initial usability ratings, which tend to improve with repeated exposure. In contrast, the high usability scores observed in this study suggest that the system was well optimized for first-time users, reducing potential barriers to interaction. The Perceived Ease of Use subscale further supports this, indicating that while users found the system easy to navigate, some effort was required to learn the procedural steps. The PSSUQ results, reflecting moderate user satisfaction, are consistent with prior studies evaluating VR-based procedural training, reinforcing the system’s effectiveness for medical education.

The reliability analysis of usability and experience assessment scales confirmed that SUS and PSSUQ demonstrated strong internal consistency. However, some subscales, such as the Nausea Symptoms subscale of the SSQ and the Perceived Naturalness subscale of the PQ, had lower-than-expected reliability. Several factors may account for this variability. Simulator sickness symptoms can vary widely across individuals, influenced by differences in motion sensitivity, vestibular adaptation, and prior VR exposure. Similarly, expectations regarding VR realism may have affected the Perceived Naturalness subscale, particularly among experienced users, who may have had higher expectations for environmental fidelity. While these lower reliability values indicate variability in participant responses, they do not necessarily invalidate the subscales.

The PQ scores indicated a strong sense of immersion among participants, with no significant differences between novice and experienced users, except for a borderline significant difference in the Perceived Naturalness subscale. This suggests that while experienced users may perceive VR environments as more natural, both groups experienced a high level of presence [21]. Maintaining a strong sense of presence is essential in medical simulations, as immersive learning enhances procedural training and stress management skills. Prior studies suggest that enhancing interactivity, improving haptic feedback, and refining visual realism could further increase presence scores in VR-based medical training.

The analysis of HRV data revealed no significant correlations between physiological stress and usability or presence scores. This suggests that participants’ perceptions of usability were primarily cognitive, influenced by their interaction with the interface rather than underlying physiological stress responses. These findings are consistent with previous literature indicating that usability is predominantly shaped by system design and functionality rather than physiological arousal [22]. However, the absence of correlation between HRV and usability or presence does not preclude interactions between cognitive and physiological factors. Several potential mediators or confounding variables could explain this finding. Individual stress tolerance, task difficulty, cognitive load, and personal coping strategies may have influenced HRV responses in ways not fully captured by usability assessments [23]. For instance, some participants may have experienced heightened physiological stress without perceiving usability issues, while others may have found the interface challenging but did not exhibit strong physiological responses. Prior research indicates that task complexity and prior clinical experience can modulate physiological stress responses in simulation-based training [24].

HRV metrics are also influenced by various physiological and psychological factors beyond stress, including baseline autonomic function, fatigue, and attentional demands. Given that VR simulations inherently require motor adaptation and spatial awareness adjustments, some users may have exhibited physiological responses unrelated to usability perceptions. These findings highlight the importance of multidimensional stress assessment approaches that incorporate both HRV and subjective stress measures (eg, NASA-TLX, perceived workload scales) to better capture cognitive and emotional load during VR training.

Interestingly, HRV parameters did not significantly correlate with presence, suggesting that stress responses do not strongly influence immersion in virtual environments. Instead, presence appears to be more dependent on design quality, realism, and interactivity than physiological arousal [25]. These findings align with prior research indicating that immersion in VR is primarily driven by environmental realism and user engagement rather than stress levels [26]. However, subtle interactions between physiological responses and presence may still exist, potentially influenced by individual factors such as trait anxiety or cognitive flexibility. Future research should explore these relationships further.

Recent studies have demonstrated that VR-based stress management interventions, including adaptive exposure and biofeedback integration, can enhance stress resilience and user engagement [27]. The use of real-time physiological feedback, such as HRV-guided VR adjustments, has been proposed as a strategy to optimize stress regulation in immersive environments. Given that stress responses in medical training simulations are complex and influenced by multiple factors, incorporating similar adaptive VR mechanisms in future iterations of medical VR training tools could improve user experience, reduce discomfort, and enhance long-term skill retention [4].

The results demonstrated significant differences in simulator sickness symptoms between participants with and with no prior VR experience. Novice users reported significantly higher levels of nausea and oculomotor discomfort, reinforcing the idea that familiarity with immersive environments helps mitigate the physical discomfort associated with VR [28]. These findings align with prior studies suggesting that repeated VR exposure can lead to adaptation, gradually reducing the severity of simulator sickness symptoms over time [29]. This adaptation effect highlights the need for tailored protocols for novice users, particularly in medical and emergency training contexts, to minimize discomfort and optimize the learning experience.

Despite differences in simulator sickness, there were no significant differences between novice and experienced users in PSSUQ or PQ scores. This suggests that the system was designed to be accessible and intuitive for users regardless of prior VR experience, reinforcing its inclusiveness. The lack of differences in usability scores further supports the idea that first-time users can effectively engage with VR simulations without a steep learning curve [30].

The borderline significance observed in the Perceived Naturalness subscale of the PQ suggests that experienced users may be more attuned to subtle environmental details, contributing to a heightened perception of realism. For novice users, improving environmental naturalness could enhance immersion. While the current simulation featured low-polygon, cartoon-like graphics, incorporating more photorealistic elements may improve engagement, provided that the “uncanny valley” effect is avoided.

This study has several limitations. First, the sample consisted solely of final-year medical students from a single institution, limiting the generalizability of findings to practice clinicians or professionals with different levels of experience. Since clinical expertise influences usability perceptions and physiological stress responses in VR environments, future studies should include more diverse populations across various career stages and medical specialties. Second, usability (SUS and PSSUQ) and simulator sickness (SSQ) were assessed using self-reported measures, which may introduce response biases. External factors such as fatigue or prior VR exposure could have influenced participants’ perceptions. Future research should complement self-reported data with objective behavioral metrics, such as task completion times or interaction errors, to enhance measurement validity. Third, while HRV provided an objective measure of autonomic stress responses, it did not capture cognitive load or psychological stress. The lack of subjective stress assessments (eg, NASA-TLX) and neurocognitive performance measures (eg, reaction times) limits the ability to fully understand the interaction between stress and usability in VR-based procedural training. Fourth, the study’s short VR session (10 minutes) restricted the assessment of long-term adaptation to simulator sickness and stress. Research suggests that symptoms and stress responses diminish with repeated exposure, but this study only measured outcomes postsession. Future studies should track symptom progression at multiple time points to assess adaptation over time. Fifth, the absence of real-world consequences for procedural errors may have reduced perceived urgency and stress. In clinical settings, mistakes during pericardiocentesis can have severe outcomes, whereas errors in VR had no direct impact. Introducing performance-based feedback or simulated consequences could enhance realism and improve training effectiveness. In addition, the low-polygon VR environment, while minimizing cognitive load, may have affected immersion, particularly for users with prior VR experience. Higher-fidelity graphics could enhance realism, although careful design is needed to avoid cognitive overload or the “uncanny valley” effect. Furthermore, a limitation of this study is the binary classification of prior VR experience (yes/no), which does not capture potential differences in familiarity with VR applications, such as gaming, medical simulations, or other interactive experiences. A more granular classification—differentiating casual VR users, gaming-oriented users, and those with professional training in VR environments—could provide deeper insights into how prior exposure influences usability and simulator sickness outcomes.

To address these limitations, future research should include broader participant samples, incorporate both objective and subjective stress measures, evaluate longer VR sessions, and refine simulation realism. Exploring different VR platforms and interaction methods could also inform the best practices for improving usability and immersion in medical training simulations.

The findings of this study provide practical insights into optimizing VR-based medical training. Given the high usability of the system (SUS=75), future improvements should prioritize reducing simulator sickness rather than addressing usability concerns. Implementing adaptive exposure strategies—such as shorter initial sessions and gradual complexity increases—could help novice users acclimate, minimizing nausea and oculomotor discomfort. The negative correlation between HRV parameters and simulator sickness symptoms suggests that HRV could be integrated as a real-time monitoring tool to identify users experiencing discomfort. This would allow proactive adjustments, such as pausing sessions or modifying simulation parameters, to enhance user comfort. To maintain usability and realism, VR training should continue using diegetic interactions that mirror real-world clinical procedures, minimizing artificial UI elements that could disrupt immersion. In addition, incorporating performance-based feedback—such as error tracking or consequence-driven scenarios—could reinforce learning and better simulate real-world clinical pressures. By refining exposure strategies, integrating real-time stress monitoring, and enhancing realism, VR-based simulations can provide a more immersive, effective, and user-friendly training experience for medical learners.

This study suggests that VR-based simulations for high-stress medical procedures, such as pericardiocentesis, offer high usability (SUS=75) and strong immersion, making them effective training tools for users with varying VR experience levels. Physiological stress responses (HRV) correlated more with simulator sickness than with usability or presence, indicating that physical discomfort plays a key role in stress modulation within VR environments.

Prior VR experience was associated with lower simulator sickness symptoms, reinforcing the potential benefits of gradual exposure strategies for first-time users. These findings highlight the complex interaction between usability, stress responses, and prior experience in VR-based medical training.

While these results contribute to understanding VR usability and stress responses, further research is needed to validate these findings in larger and more diverse samples. Future studies should also assess longitudinal effects to determine whether adaptation to simulator sickness occurs over repeated exposures.