Abstract
Background
The aim of this study is to evaluate the effectiveness of virtual reality (VR) for learning ophthalmic anatomy and to determine whether peer-assisted facilitation produces learning outcomes comparable to expert-assisted facilitation among preclinical medical students. A secondary objective was to contextualize learning outcomes by comparing performance of preclinical students with that of senior medical students receiving expert-assisted VR instruction.
Methods
This is a prospective controlled experimental study with repeated-measures assessments. Second-year (preclinical) medical students without prior formal ophthalmic anatomy instruction were recruited voluntarily (peer-assisted n = 31, expert-assisted n = 32). Exclusion criteria included previous VR-based ophthalmic anatomy training, history of severe motion sickness or vertigo, or refusal to provide consent. In addition, fifth-year medical students undertaking their ophthalmology clerkship participated in a parallel prospective pre–post cohort (n = 61 baseline; n = 55 completed the intervention). All participants first attended a conventional didactic lecture on ophthalmic anatomy. Second-year students then completed an immersive VR learning session exploring a stereoscopic 3D model of the globe and adnexa using a head-mounted display. Sessions were facilitated either by an anatomy faculty expert or by trained peer facilitators. Fifth-year students completed the same instructional sequence with expert-assisted VR only. The primary outcome was change in anatomy knowledge measured by a 20-item assessment administered at baseline (T0), after didactic instruction (T1), and after VR (T2).
Results
Among second-year students, knowledge scores improved significantly across timepoints (repeated-measures ANOVA: F(2,124) = 251.0, p<.001, η²ₚ=0.802). Post-VR scores were comparable between peer-assisted and expert-assisted groups (15.1 ± 3.1 vs. 15.2 ± 3.4; t(61 =–0.12, p=.905, d=–0.03). Simulator sickness scores were low and did not differ between groups (4.2 ± 4.0 vs. 3.7 ± 3.7; p=.649), and satisfaction ratings were high in both conditions. Among fifth-year students, knowledge scores also increased significantly across instructional stages (F(2,106) = 33.4, p<.001, η²ₚ=0.387). Despite higher baseline scores among senior students (t(121)=–3.17, p=.002), post-VR performance was comparable between cohorts (t(115) = 1.11, p=.269).
Conclusions
Peer-assisted facilitation produced learning outcomes comparable to expert-assisted instruction while maintaining high learner satisfaction and minimal simulator sickness. These findings suggest that peer facilitation may represent a scalable approach for implementing VR-based anatomy education.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12909-026-09659-w.
Keywords: Virtual reality, Peer-assisted learning, Eye anatomy, Ophthalmology, Medical education
Introduction
Ophthalmic anatomy presents specific learning challenges because the globe and orbital structures are small, deeply situated, and highly three-dimensional. Learning requires strong spatial reasoning, which is an ability consistently linked to anatomy performance [15]. Conventional teaching methods, such as lectures, atlases, and cadaveric dissection, often fail to convey spatial relationships adequately. The delicacy of ophthalmic tissues limits repeated dissection, and substitutes like animal-eye laboratories only partially replicate human anatomy [12]. In addition, undergraduate ophthalmology exposure is brief and inconsistent across curricula [11], highlighting the need for efficient and scalable methods that support 3D understanding.
Immersive technologies such as virtual reality (VR) offer solutions to these spatial and accessibility limitations. A meta-analysis demonstrated that extended-reality tools, including VR, significantly improve anatomy knowledge compared with traditional approaches (standardized mean difference around 0.40) and are perceived as highly beneficial by students [5]. A complementary systematic review documented the growing integration of VR across health-science anatomy curricula and consistently positive learner experiences [1]. These benefits are particularly relevant to anatomy education because VR enables learners to examine anatomical structures as interactive three-dimensional objects rather than static two-dimensional images. Such immersive exploration may strengthen spatial reasoning by allowing students to view structures from multiple angles, appreciate depth and positional relationships, and repeatedly revisit complex regions at their own pace. VR also offers standardized delivery of anatomical content, ensuring that learners encounter the same structures, sequence, and instructional cues across sessions. These features make VR especially suitable for ophthalmic anatomy, where small, deeply located structures and complex spatial relationships are difficult to convey through conventional teaching alone.
While VR can address core spatial challenges in ophthalmic anatomy, its educational impact depends on how learning is facilitated. The medical-education evidence base shows that peer-assisted learning often yields significant academic outcomes, while improving scalability [3]. Within anatomy specifically, near-peer teaching has been recommended as a feasible and educationally valuable approach in laboratory and small-group formats [4]. Its advantages are possibly due to a psychological safe environment and relatability [17]. However, immersive VR also brings specific challenges. A scoping review highlighted cybersickness as a recurring factor influencing learning in VR-based anatomy education [16], and even mild symptoms can reduce engagement. This shows the need to examine whether different facilitation approaches help manage such challenges and support effective learning.
Despite converging evidence for the value of facilitation and peer-assisted learning, there is a lack of evidence on a direct comparison of expert-assisted versus peer-assisted facilitation within VR-based anatomy. Addressing this gap, particularly in a spatially demanding domain like ophthalmic anatomy, can inform scalable implementation without compromising learning quality. Our study therefore tests whether peer-assisted VR achieves learning gains comparable to expert-assisted VR in Year-2 (preclinical) medical students, while also examining student experience. Moreover, our study compares their performance with Year-5 medical students, who were taught the same anatomy content through expert-assisted VR, to contextualize learning gains across different stages of medical training and to determine whether VR-supported instruction can help narrow baseline knowledge gaps.
Methods
Year-2 students
Study design and setting
This prospective controlled experiment was conducted at Gazi University to evaluate the effectiveness of virtual reality (VR) for teaching ophthalmic anatomy and to compare expert-assisted versus peer-assisted facilitation during VR-based instruction. At Gazi University Faculty of Medicine, ophthalmic anatomy is included in the standard curriculum in Year 2, while clinical ophthalmology is taught during the Year-5 ophthalmology clerkship. There is no additional formal ophthalmology teaching between Years 2 and 5. The VR-enhanced ophthalmic anatomy session evaluated in this study was an additional instructional intervention and was not part of the routine curriculum. Importantly, the Year-2 students participating in the study had not yet received the standard Year-2 ophthalmic anatomy teaching before enrolment. The study targeted pre-clinical medical students in their second year. Clinical trial number is not applicable as it is not a clinical trial.
Participants and eligibility criteria
Participants were second-year medical students who had not yet received formal instruction in ophthalmic anatomy. Recruitment was conducted through class announcements. Exclusion criteria included prior participation in VR-based ophthalmic anatomy training, a self-reported history of severe motion sickness or vertigo, or refusal to provide informed consent. All eligible students who volunteered were included.
Intervention and facilitation conditions
VR platform and content
The instructional intervention utilized a HTC Vive Pro headset running the ophthalmic VR model we developed on RoT Version 2.4.0 (Infotron), Ankara, Turkiye. The application presented an interactive, stereoscopic 3D model of the globe and adnexa, allowing learners to explore surface, sectional, and relational anatomy in immersive space. It was developed by the anatomy department, which is experienced in teaching anatomy and developing VR solutions [2, 8], with the help of ophthalmology department. Faculty members from both departments checked and approved accuracy of the model.
Facilitation conditions
Two facilitation approaches were compared: Expert-Assisted (EA) and Peer-Assisted (PA). EA sessions were led by a faculty member in anatomy with prior experience in VR-based instruction. PA sessions were led by medical students who had previously completed a structured facilitator training. Peer facilitators underwent a 1-hour structured training session under the supervision of the faculty investigator. Training covered VR system operation, session flow, learner engagement techniques, and standardized guidance based on the instructional handout.
Both EA and PA facilitators adhered to the identical session handout, and timing to ensure content standardization and equivalent exposure across groups. Figure 1 includes shots from the VR environment.
Fig. 1.
The VR environment*. *The video is accessible via the QR code and also via this link: https://www.youtube.com/watch?v=vVB5dHgb5yA
Procedures
We followed a structured sequence for all participants. T0 (Pre-test) was for assessing baseline knowledge before any instruction. Then, the didactic session was held and it consisted of a conventional lecture on ophthalmic anatomy delivered to all students. At T1 (Post-didactic test), participants completed the same test instrument as T0 without receiving feedback. This was followed by a VR session in which participants were randomly assigned in a 1:1 ratio to either EA or PA facilitation, and each individual completed a VR session using the same process. This fixed sequence was selected to reflect the intended curricular implementation of VR as an adjunctive learning activity following conventional instruction, rather than as a stand-alone replacement for the lecture. Therefore, the order of didactic and VR instruction was not counterbalanced. Finally, at T2 (Post-VR test), the same assessment instrument was administered again with no feedback provided. Across the second-year cohort, 31 students participated in peer-assisted sessions and 32 in expert-assisted sessions.
Outcome measures and instruments
Primary outcomes
Primary outcomes included the effectiveness of VR (Aim 1), measured as improvement in total test scores across timepoints T0, T1, and T2. The facilitation comparison (Aim 2) examined the difference in post-VR total test scores at T2 between the EA and PA conditions.
Assessment instrument
The 20-item ophthalmic anatomy test was designed to assess foundational knowledge appropriate for Year-2 medical students who had not yet received formal ophthalmic anatomy instruction. The cognitive assessment comprised six single-best-answer multiple-choice questions (1 point each) and fourteen labeled-image identification items (1 point each), for a total of 20 possible points. The assessment covered core structures and spatial relationships included in the didactic and VR sessions. Most items assessed recall and anatomical identification. The intended difficulty level was introductory to moderate. All six multiple-choice questions are provided as Supplementary Material 1. Image-based items were displayed on-screen rather than within the VR environment. Identical test items were used at all three timepoints. All items were developed by an ophthalmology faculty member, reviewed by an additional ophthalmology faculty and an anatomy faculty member to ensure content validity, and further evaluated by a physician with a PhD in medical education. Free-text responses were independently scored by two blinded graders using a predefined answer key. Because the responses were clear and unambiguous, there were no scoring disagreements between graders. To ensure the validity of the assessment instrument, it was piloted with 35 students and reviewed/revised based on this pilot implementation.
Secondary outcomes
Secondary outcomes were assessed using two self-report instruments administered following the VR session. The Simulator Sickness Questionnaire [7] evaluated VR-related discomfort across 16 symptoms, each rated on a 4-point scale ranging from 0 (none) to 3 (severe). In addition, a self-developed VR Satisfaction Questionnaire consisting of eight items was used to evaluate ease of use, engagement, perceived learning benefit, and overall satisfaction on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Items were informed by commonly evaluated domains in technology-enhanced learning and VR acceptability studies, but the questionnaire was not adapted from a single previously validated instrument and was not validated on an independent sample prior to use. Therefore, satisfaction data were treated as exploratory secondary acceptability outcomes. All surveys were administered in printed form immediately following the VR session.
Assignment
Participants were assigned to either the expert-assisted or peer-assisted condition using manual allocation by considering arrival time at the experiment site. For the comparison between expert-assisted and peer-assisted VR facilitation, a sample size calculation was performed. Assuming equal allocation, a two-sided α = 0.05, 80% power, and an effect size of d = 0.70, the required sample size was 35 students per arm (total N = 70). The assumed effect size was selected pragmatically as a moderate-to-large, educationally meaningful between-group difference that the study should be able to detect. This estimate was not informed by a prior directly comparable VR ophthalmic anatomy facilitation trial or by pilot data, as such data were unavailable. Then, the plausibility of this assumption was supported by the broader peer-assisted learning literature in medical education. In a systematic review and meta-analysis, Brierley et al., reported an overall positive effect of peer-assisted learning on academic performance (SMD = 0.52), with a larger effect for practical-skills teaching (SMD = 0.69) [3]. Therefore, d = 0.70 was considered an appropriate threshold for detecting a practically important disadvantage of peer-assisted facilitation, should such a difference exist.
Year-5 students
A prospective pre–post experimental design was implemented with fifth-year medical students during their ophthalmology clerkship at Gazi University. This cohort served as a secondary comparator to contextualize learning outcomes across different stages of medical training. Because Year-5 students were differed from Year-2 students in training level, this comparison was observational and should be interpreted as contextual rather than as a between-cohort comparison. Data were collected at three timepoints: T0 (baseline, at the very beginning of the clerkship), T1 (after a didactic lecture delivered by an ophthalmology faculty expert), and T2 (after an expert-assisted VR session).
Eligible participants were fifth-year medical students enrolled in the ophthalmology clerkship who consented to participate. Exclusion criteria were participation in VR-based ophthalmic anatomy instruction, a self-reported history of severe motion sickness or vertigo, or refusal to provide informed consent. A total of 61 students completed the pre-test (T0), 55 completed the post-didactic test (T1), and 55 completed the post-VR test (T2). Following baseline assessment (T0), all students attended a conventional didactic session on ophthalmic anatomy delivered by an expert faculty member. Subsequently, participants completed the same knowledge test to evaluate immediate post-lecture learning (T1). Each student then undertook a VR session facilitated by an expert faculty member. The VR models were identical to those used in the Year-2 study. Year-5 students used Oculus Quest 3 as a headset.
Knowledge was assessed using the identical 20-item test administered to the Year-2 students, comprising six single-best-answer multiple-choice questions and fourteen labeled-image identification items. In addition, Year-5 participants completed an expanded 50-item version that included 30 additional MCQs specifically tailored to their advanced Year-5 level. These 30 additional MCQs were developed by ophthalmology experts (faculty members) with the help of an anatomy expert and a medical doctor with PhD in medical education. The same scoring procedures and review steps were applied. Primary outcomes were within-group improvement in knowledge across the three timepoints (T0-T1-T2). No secondary measures (simulator sickness or satisfaction) were administered in the Year-5 students, because this cohort participated during the routine ophthalmology clerkship, where available time for additional post-session questionnaires was limited. Consequently, the Year-5 arm was designed to evaluate knowledge gain only, whereas simulator sickness and satisfaction were assessed in the Year-2 experimental comparison of peer- and expert-assisted facilitation.
Statistical analysis
Data were analyzed using Jamovi (version 2.6), an R-based open-source software [10]. For within-group comparisons across timepoints (T0-T1-T2), one-way repeated-measures ANOVA was performed, with Mauchly’s test used to evaluate sphericity and Greenhouse–Geisser correction applied when violated. Between-group comparisons (expert-assisted vs. peer-assisted facilitation; Year-2 vs. Year-5 cohorts) were examined using independent-samples T-tests or mixed-design ANOVA as appropriate. For each repeated-measures analysis involving three timepoints (T0, T1, and T2), three pairwise comparisons were conducted: T0-T1, T1-T2, and T0-T2. These post hoc comparisons were adjusted using Tukey’s method to control the family-wise error rate. The family-wise significance level was set at α = 0.05; therefore, Tukey-adjusted p-values < 0.05 were considered statistically significant. No additional Bonferroni-type per-comparison alpha threshold was applied, because significance decisions were based on Tukey-adjusted p-values. Effect sizes were reported as partial eta-squared (η²ₚ) for ANOVA and Cohen’s d for T-tests. Statistical significance was set at p < .05 for all analyses.
Ethical considerations
This study was approved by the Gazi University Institutional Review Board (Approval No: 2025 − 1492). All participants provided informed consent before participation. Data were anonymized prior to analysis, and participation was voluntary. The scores in this study did not affect their faculty grades and were independent of academic evaluation.
Results
Year-2 students
Effect of training on knowledge scores
A one-way repeated-measures ANOVA examined changes in total anatomy test scores across three timepoints: pre-test (T0), post-didactic (T1), and post-VR (T2). There was a significant main effect of time, F(2, 124) = 251.0, p < .001, η²ₚ = 0.802, indicating statistically significant improvement in performance across sessions. Mauchly’s test of sphericity was not significant (p = .184), confirming that the assumption of sphericity was met and no correction was necessary.
Post hoc pairwise comparisons (Tukey-adjusted) revealed significant gains from T0 to T1 (p < .001), T1 to T2 (p < .001), and T0 to T2 (p < .001). These results demonstrate robust and progressive learning, with additional improvement following the VR session beyond the didactic component.
Comparison of peer- and expert-assisted facilitation
To examine whether facilitation type influenced learning outcomes, independent-samples T-tests were conducted at each timepoint. Before instruction (T0), expert-assisted students scored higher (M = 7.4 ± 4.2) than peer-assisted students (M = 5.4 ± 2.7), t(61) = − 2.21, p = .031, d = − 0.56. However, no significant group differences were observed at T1 (t(61) = − 0.06, p = .949, d = − 0.02) or T2 (t(61) = − 0.12, p = .905, d = − 0.03). At the post-VR assessment (T2), mean scores were comparable between the peer-assisted (M = 15.1 ± 3.1) and expert-assisted (M = 15.2 ± 3.4) groups, indicating that both facilitation approaches yielded equivalent learning outcomes by the end of the intervention. Table 1 shows the descriptive statistics.
Table 1.
Descriptive statistics based on groups in terms of 20-item anatomy test scores
| Group | T0 | T1 | T2 | |
|---|---|---|---|---|
| N | Peer-Assisted | 32 | 32 | 32 |
| Expert-Assisted | 31 | 31 | 31 | |
| Mean | Peer-Assisted | 5.4 | 12.1 | 15.1 |
| Expert-Assisted | 7.4 | 12.2 | 15.2 | |
| Standard deviation | Peer-Assisted | 2.7 | 4.2 | 3.1 |
| Expert-Assisted | 4.2 | 4.2 | 3.4 | |
| Minimum | Peer-Assisted | 1 | 5 | 7 |
| Expert-Assisted | 2 | 4 | 9 | |
| Maximum | Peer-Assisted | 13 | 18 | 20 |
| Expert-Assisted | 19 | 19 | 20 |
Secondary outcomes
Self-reported measures of sickness and satisfaction were compared between facilitation conditions using independent-samples T-tests. No significant group differences were found for any secondary outcome.
Simulator sickness scores were low overall (peer-assisted: M = 4.2 ± 4.0; expert-assisted: M = 3.7 ± 3.7), with no significant difference between groups, t(61) = 0.46, p = .649, d = 0.12. Table 2 shows the detailed statistics.
Table 2.
Descriptive statistics based on groups in terms of sickness symptoms*
| Items | Group | N | Mean | SD | p | Effect Size (Cohen’s d) |
|---|---|---|---|---|---|---|
| General discomfort | Peer-Assisted | 32 | 0.44 | 0.76 | 0.774 | 0.07 |
| Expert-Assisted | 31 | 0.39 | 0.62 | |||
| Fatigue | Peer-Assisted | 32 | 0.63 | 0.91 | 0.097 | 0.42 |
| Expert-Assisted | 31 | 0.29 | 0.64 | |||
| Headache | Peer-Assisted | 32 | 0.34 | 0.75 | 0.149 | 0.37 |
| Expert-Assisted | 31 | 0.13 | 0.34 | |||
| Eye strain | Peer-Assisted | 32 | 0.34 | 0.55 | 0.308 | -0.26 |
| Expert-Assisted | 31 | 0.51 | 0.77 | |||
| Difficulty focusing | Peer-Assisted | 32 | 0.34 | 0.60 | 0.400 | 0.21 |
| Expert-Assisted | 31 | 0.23 | 0.50 | |||
| Increased salivation | Peer-Assisted | 32 | 0.28 | 0.58 | 0.242 | 0.30 |
| Expert-Assisted | 31 | 0.13 | 0.43 | |||
| Sweating | Peer-Assisted | 32 | 0.34 | 0.55 | 0.723 | 0.09 |
| Expert-Assisted | 31 | 0.29 | 0.64 | |||
| Nausea | Peer-Assisted | 32 | 0.25 | 0.62 | 0.250 | 0.29 |
| Expert-Assisted | 31 | 0.10 | 0.40 | |||
| Difficulty concentrating | Peer-Assisted | 32 | 0.16 | 0.37 | 0.216 | -0.31 |
| Expert-Assisted | 31 | 0.32 | 0.65 | |||
| Fullness of head | Peer-Assisted | 32 | 0.53 | 0.72 | 0.432 | -0.20 |
| Expert-Assisted | 31 | 0.68 | 0.75 | |||
| Blurred vision | Peer-Assisted | 32 | 0.28 | 0.46 | 0.620 | 0.12 |
| Expert-Assisted | 31 | 0.23 | 0.43 | |||
| Dizziness (eyes open) | Peer-Assisted | 32 | 0.03 | 0.18 | 0.397 | -0.21 |
| Expert-Assisted | 31 | 0.10 | 0.40 | |||
| Dizziness (eyes closed) | Peer-Assisted | 32 | 0.09 | 0.30 | 0.265 | -0.28 |
| Expert-Assisted | 31 | 0.19 | 0.40 | |||
| Vertigo | Peer-Assisted | 32 | 0.00 | 0.00 | 0.149 | -0.37 |
| Expert-Assisted | 31 | 0.06 | 0.25 | |||
| Stomach Awareness | Peer-Assisted | 32 | 0.13 | 0.42 | 0.761 | 0.08 |
| Expert-Assisted | 31 | 0.10 | 0.30 | |||
| Burping | Peer-Assisted | 32 | 0.00 | 0.00 | NaN** | NaN** |
| Expert-Assisted | 31 | 0.00 | 0.00 |
* 0 = None, 1 = Slight, 2 = Moderate, 3 = Severe
** For the “Burping” item, all participants in both groups reported 0 symptoms; therefore, no between-group test or Cohen’s d was estimated because of zero variance in both groups
Satisfaction scores were uniformly high (peer-assisted: M = 36.4 ± 2.4; expert-assisted: M = 36.9 ± 2.6), with no statistically significant difference, t(61) = − 0.84, p = .404, d = − 0.21. Table 3 shows the detailed statistics.
Table 3.
Descriptive statistics based on groups in terms of satisfaction*
| Items | Group | N | Mean | SD | p | Effect size (Cohen’s d) |
|---|---|---|---|---|---|---|
| Using the VR system was easy for me. | Peer-Assisted | 32 | 4.5 | 0.80 | 0.942 | -0.02 |
| Expert-Assisted | 31 | 4.5 | 0.72 | |||
| I encountered no technical difficulties using the VR hardware. | Peer-Assisted | 32 | 4.2 | 0.79 | 0.050 | 0.50 |
| Expert-Assisted | 31 | 3.7 | 1.30 | |||
| The VR training was engaging. | Peer-Assisted | 32 | 4.8 | 0.37 | 0.957 | 0.01 |
| Expert-Assisted | 31 | 4.8 | 0.37 | |||
| After using the VR, I felt more confident in identifying anatomical structures. | Peer-Assisted | 32 | 4.5 | 0.57 | 0.378 | -0.22 |
| Expert-Assisted | 30 | 4.6 | 0.62 | |||
| My spatial awareness of eye anatomy increased thanks to the VR training. | Peer-Assisted | 32 | 4.7 | 0.46 | 0.422 | -0.20 |
| Expert-Assisted | 31 | 4.8 | 0.40 | |||
| Compared to the traditional method, the VR training was more effective. | Peer-Assisted | 32 | 4.7 | 0.59 | 0.075 | -0.46 |
| Expert-Assisted | 31 | 4.9 | 0.30 | |||
| The VR training overall met my expectations. | Peer-Assisted | 32 | 4.5 | 0.62 | 0.013 | -0.64 |
| Expert-Assisted | 31 | 4.8 | 0.40 | |||
| I would like to participate in similar VR training in the future. | Peer-Assisted | 31 | 4.7 | 0.54 | 0.006 | -0.72 |
| Expert-Assisted | 31 | 5.0 | 0.18 |
* 1 = Strongly Disagree, 2 = Disagree, 3 = Neutral, 4 = Agree, 5 = Strongly Agree
Year-5 students
A one-way repeated-measures ANOVA examined changes in Year-5 students’ performance on the shared 20-item ophthalmic anatomy test across three timepoints: pre-test (T0), post-didactic (T1), and post-VR (T2). The number of participants in T0, T1, and T2 was 60, 60, and 54, respectively. There was a significant main effect of time, F(2, 106) = 33.4, p < .001, η²ₚ = 0.387, indicating progressive improvement in knowledge scores throughout the instructional sequence. Mauchly’s test of sphericity was significant (p = .018); however, the Greenhouse–Geisser correction (ε = 0.875) produced an equivalent result. Mean scores increased from 8.78 ± 4.79 at T0 to 10.70 ± 4.54 at T1 and 14.4 ± 3.97 at T2.
Post-hoc comparisons (Tukey-adjusted) confirmed significant gains from T0 to T1 (t(53) = − 2.58, p = .033), T1 to T2 (t(53) = − 6.58, p < .001), and T0 to T2 (t(53) = − 7.54, p < .001). These results demonstrate robust, stepwise learning among Year-5 students, with additional substantial improvement following the expert-assisted VR session beyond the effect of the didactic lecture.
Another repeated-measures ANOVA was performed on the 30 advanced multiple-choice questions administered to Year-5 students across three timepoints. The analysis revealed a significant main effect of time, F(2, 106) = 57.3, p < .001, η²ₚ = 0.519, indicating substantial improvement over the course of instruction. Mauchly’s test of sphericity was not significant (p = .103). Mean scores rose progressively from 9.92 ± 4.35 at T0 to 13.90 ± 5.23 at T1 and 18.0 ± 5.10 at T2. Post-hoc pairwise comparisons confirmed significant gains from T0 to T1 (t(53) = − 5.95, p < .001), T1 to T2 (t(53) = − 5.61, p < .001), and T0 to T2 (t(53) = − 9.47, p < .001). These results demonstrate cumulative learning among Year-5 students, with marked additional advancement following the expert-assisted VR session beyond the effect of didactic instruction alone.
Performance comparison between Year-2 and Year-5 students
A two-way mixed repeated-measures ANOVA assessed the effects of time (T0, T1, T2) and cohort (Year-2 vs. Year-5) on the 20-item ophthalmic anatomy test. There was a significant main effect of time, F(2, 230) = 171.4, p < .001, η²ₚ = 0.599, reflecting substantial knowledge gains across instructional sessions. The main effect of cohort was not significant, F(1, 115) = 0.006, p = .939, η²ₚ < 0.001, indicating comparable overall performance between groups. A significant Time × Cohort interaction emerged, F(2, 230) = 13.3, p < .001, η²ₚ = 0.104, demonstrating that learning trajectories differed between cohorts. Both Year-2 and Year-5 students improved significantly after the didactic and VR sessions, but Year-2 students exhibited larger relative gains from baseline, whereas Year-5 students began with higher initial scores and showed smaller incremental increases (Fig. 2).
Fig. 2.
Pre-test (T0), post-didactic (T1), and post-VR (T2) scores of Year-2 and Year-5 medical students on an identical 20-item anatomy test
Independent-samples T-tests were conducted to compare Year-2 and Year-5 students’ knowledge scores at each timepoint on the shared 20-item test. At baseline (T0), Year-5 students scored significantly higher than Year-2 students, t(121) = − 3.17, p = .002, d = 0.57, 95% CI [–0.93, − 0.21], indicating a moderate effect size favoring Year-5. Following the didactic session (T1), the difference was no longer significant, t(121) = 1.84, p = .069, d = 0.33. At the post-VR assessment (T2), mean scores remained statistically comparable between cohorts, t(115) = 1.11, p = .269, d = 0.21.
Discussion
This study examined the effectiveness of VR for teaching ophthalmic anatomy and evaluated whether peer-assisted facilitation yields learning outcomes comparable to expert-assisted VR. Across both Year-2 and Year-5 cohorts, VR produced significant gains in anatomy knowledge beyond those achieved through a conventional didactic lecture. Notably, among pre-clinical Year-2 students, peer-assisted VR was equivalent to expert-assisted facilitation in terms of post-VR performance, self-reported satisfaction, and levels of simulator sickness. Furthermore, despite the substantial baseline performance differences between Year-2 and Year-5 students on the shared anatomy assessment, post-VR scores were comparable. These findings collectively indicate that VR can support learning across stages of medical training and that carefully prepared peer facilitators can effectively guide VR-based anatomy instruction without compromising learning quality.
The central contribution of this study is the direct comparison of expert-assisted and peer-assisted facilitation within VR-based anatomy teaching, an area where empirical evidence is scarce. While peer-assisted learning, together with near-peer teaching [4], has a long history in anatomy education [9, 13], demonstrating benefits such as improved scalability [6], improved academic performance in gross anatomy [14], its applicability within VR settings has not been well explored. Our finding that peer facilitators performed as effectively as faculty experts suggests that VR may reduce the instructional demands traditionally placed on faculty expert facilitators. Additionally, structured facilitator training and a standardized session handout likely minimized variability between conditions. These results support the feasibility of peer-assisted facilitation as a scalable model for integrating VR into anatomy curricula, particularly in resource-constrained settings where faculty availability is limited.
Comparison between pre-clinical and clinical students offers further insight into how VR supports learning across training stages. Year-5 students, as expected, entered with higher baseline scores due to their broader exposure to anatomy and clinical ophthalmology. However, the magnitude of their gains following the VR session was substantial and comparable to the incremental improvements seen in Year-2 students. More importantly, the two cohorts achieved similar performance levels on the shared assessment after VR. This pattern suggests that VR may help bridge early-phase knowledge gaps by providing novices with a more intuitive understanding of three-dimensional spatial anatomy, while still offering clinically meaningful refinement for advanced learners. Such findings justify the use of VR both early in the curriculum and as a reinforcement tool during clinical training. Student experience metrics further support the acceptability and feasibility of VR-based instruction.
This study has several limitations. First, it was conducted at a single institution using specific VR hardware and software, which may limit generalizability. Second, knowledge was assessed using identical test items across timepoints, raising the possibility of testing effects despite the absence of feedback between assessments. Third, the study focused on short-term learning; long-term retention and transfer to clinical performance were not assessed. Fourth, the comparison between Year-2 and Year-5 cohorts was observational rather than randomized, and the absence of a traditional-only control group limits its comparability. Fifth, Year-2 and Year-5 students used different VR headsets, which may have influenced user experience or learning outcomes independently of cohort differences; therefore, the cross-cohort comparison should be interpreted cautiously. In addition, all Year-2 students received VR instruction after the didactic lecture. Although this sequence reflected our intended curricular use of VR as an additive component, it limits the ability to isolate the independent pedagogical contribution of VR from prior didactic exposure. A counterbalanced randomized design would provide stronger evidence regarding the unique and additive effects of VR. Moreover, the VR Satisfaction Questionnaire was developed specifically for this study and was not validated on an independent sample before implementation. Finally, peer facilitators in this study completed a structured training session under faculty supervision, and outcomes may differ in settings where peer preparation is less controlled. Future research should examine long-term retention of VR-acquired knowledge. Additionally, investigations into how facilitation styles interact with different types of VR content, ranging from highly interactive simulations to more passive visualizations, may clarify when expert guidance is most essential.
Supplementary Information
Authors’ contributions
Conceptualization: AS, AYÜ, YSK, KB, TVP; Methodology: AS, AYÜ, RA, GAT, İA, EÖ, HTA, BSU, BT, YSK, KB, TVP; Data Collection: AS, AYÜ, RA, GAT, İA, EÖ, HTA, BSU, BT, KB, TVP; Statistical Analysis: AS, YSK, RA; Writing – Original Draft Preparation: AS, AYÜ, YSK; Writing – Review & Editing: AS, AYÜ, RA, GAT, İA, EÖ, HTA, BSU, BT, YSK, KB, TVP.
Funding
Gazi University Scientific Research Projects Unit (BAP) supported this study (TCD-2025-10104).
Data availability
The dataset supporting the findings of this study is available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
This study was approved by the Gazi University Institutional Review Board (Approval No: 2025-1492), and conducted in accordance with the Declaration of Helsinki.
All participants gave informed consent before enrolling. Participation was voluntary, unpaid, and did not affect their academic standing.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Adnan S, Benson AC, Xiao J. How virtual reality is being adopted in anatomy education in health sciences and allied health: A systematic review. Anat Sci Educ. 2025;18(5):496–525. 10.1002/ase.70027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Alim E, Coşkun Ö, Peker TV. Comparison and Evaluation of the Effectiveness of Traditional Neuroanatomy Teaching in Medical Education with Virtual-Reality Application Based On 3D Virtual. Gazi Med J. 2024;35(4):407–15. 10.12996/gmj.2024.4191. [DOI] [Google Scholar]
- 3.Brierley C, Ellis L, Reid ER. Peer-assisted learning in medical education: A systematic review and meta‐analysis. Med Educ. 2022;56(4):365–73. 10.1111/medu.14672. [DOI] [PubMed] [Google Scholar]
- 4.Evans DJR, Cuffe T. Near-peer teaching in anatomy: An approach for deeper learning. Anat Sci Educ. 2009;2(5):227–33. 10.1002/ase.110. [DOI] [PubMed] [Google Scholar]
- 5.García-Robles P, Cortés‐Pérez I, Nieto‐Escámez FA, García‐López H, Obrero‐Gaitán E, Osuna‐Pérez MC. Immersive virtual reality and augmented reality in anatomy education: A systematic review and meta‐analysis. Anat Sci Educ. 2024;17(3):514–28. 10.1002/ase.2397. [DOI] [PubMed] [Google Scholar]
- 6.Johansson E, Holmin TE, Johansson BR, Braide M. Improving near-peer teaching quality in anatomy by educating teaching assistants: An example from Sweden. Anat Sci Educ. 2018;11(4):403–9. 10.1002/ase.1775. [DOI] [PubMed] [Google Scholar]
- 7.Kennedy RS, Lane NE, Berbaum KS, Lilienthal MG. Simulator Sickness Questionnaire: An Enhanced Method for Quantifying Simulator Sickness. Int J Aviat Psychol. 1993;3(3):203–20. 10.1207/s15327108ijap0303_3. [DOI] [Google Scholar]
- 8.Kılıç MF, Yurtsever AZ, Açıkgöz F, Başgut B, Mavi B, Ertuç E, Sevim S, Oruk T, Kıyak YS, Peker T. A new classmate in anatomy education: 3D anatomical modeling medical students’ engagement on learning through self-prepared anatomical models. Anat Sci Educ. 2025;18(7):727–37. 10.1002/ase.70070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Krych AJ, March CN, Bryan RE, Peake BJ, Pawlina W, Carmichael SW. Reciprocal peer teaching: Students teaching students in the gross anatomy laboratory. Clin Anat. 2005;18(4):296–301. 10.1002/ca.20090. [DOI] [PubMed] [Google Scholar]
- 10.Leppink J. Data analysis: More expensive does not imply better. Revista Española de Educación Médica. 2022;3(2):12–3. 10.6018/edumed.516731. [DOI] [Google Scholar]
- 11.Liao J, Wright RR, Vora GK. The Decline of Basic Ophthalmology in General Medical Education: A Scoping Review and Recommended Potential Solutions. J Med Educ Curric Dev. 2024;11:23821205241245635. 10.1177/23821205241245635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Nakashizuka H, Wakatsuki Y, Machida Y, Okubo Y, Shinojima A, Hattori T, Shimada H, Yuzawa M. Wet laboratory training using porcine eyes with eyelids. Can J Ophthalmol. 2017;52(4):398–402. 10.1016/j.jcjo.2017.01.007. [DOI] [PubMed] [Google Scholar]
- 13.Nnodim JO. A controlled trial of peer-teaching in practical gross anatomy. Clin Anat. 1997;10(2):112–7. 10.1002/(SICI)1098-2353(1997)10:2<112::AID-CA7>3.0.CO;2-X. [DOI] [PubMed]
- 14.Pessoa KJDN, Da Silva CKC, De Lima AC, Coswig VS, Neto B, De Assis Teles Santos O, De Lira D, C. A. B., Viana B, R. Near-peer teaching in a gross human anatomy course: Students’ academic performance and perceptions. J Biol Educ. 2025;1–25. 10.1080/00219266.2025.2546792. [DOI]
- 15.Roach VA, Mi M, Mussell J, Van Nuland SE, Lufler RS, DeVeau KM, Dunham SM, Husmann P, Herriott HL, Edwards DN, Doubleday AF, Wilson BM, Wilson AB. Correlating Spatial Ability With Anatomy Assessment Performance: A Meta-Analysis. Anat Sci Educ. 2021;14(3):317–29. 10.1002/ase.2029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sinha S, DeYoung V, Nehru A, Brewer-Deluce D, Wainman BC. Determinants of Learning Anatomy in an Immersive Virtual Reality Environment—A Scoping Review. Med Sci Educ. 2022;33(1):287–97. 10.1007/s40670-022-01701-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zheng B, Maciuba JM, Dong T, Durning SJ. Fellows and medical students as tutors: Facilitating beginning medical students’ clinical reasoning abilities. Med Teach. 2025;47(11):1825–35. 10.1080/0142159X.2025.2503379. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The dataset supporting the findings of this study is available from the corresponding author upon reasonable request.