Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

The effectiveness of virtual reality-based technology on anatomy teaching: a meta-analysis of randomized controlled studies

The effectiveness of virtual reality-based technology on anatomy teaching: a meta-analysis of... Background: Virtual reality (VR) is an innovation that permits the individual to discover and operate within three- dimensional (3D) environment to gain practical understanding. This research aimed to examine the general efficiency of VR for teaching medical anatomy. Methods: We executed a meta-analysis of randomized controlled studies of the performance of VR anatomy education. We browsed five databases from the year 1990 to 2019. Ultimately, 15 randomized controlled trials with a teaching outcome measure analysis were included. Two authors separately chose studies, extracted information, and examined the risk of bias. The primary outcomes were examination scores of the students. Secondary outcomes were the degrees of satisfaction of the students. Random-effects models were used for the pooled evaluations of scores and satisfaction degrees. Standardized mean difference (SMD) was applied to assess the systematic results. The heterogeneity was determined by I statistics, and then was investigated by meta-regression and subgroup analyses. Results: In this review, we screened and included fifteen randomized controlled researches (816 students). The pooled analysis of primary outcomes showed that VR improves test scores moderately compared with other approaches (standardized mean difference [SMD] = 0.53; 95% Confidence Interval [CI] 0.09–0.97, p < 0.05; I = 87.8%). The high homogeneity indicated that the studies were different from each other. Therefore, we carried out meta- regression as well as subgroup analyses using seven variables (year, country, learners, course, intervention, comparator, and duration). We found that VR improves post-intervention test score of anatomy compared with other types of teaching methods. Conclusions: The finding confirms that VR may act as an efficient way to improve the learners’ level of anatomy knowledge. Future research should assess other factors like degree of satisfaction, cost-effectiveness, and adverse reactions when evaluating the teaching effectiveness of VR in anatomy. Keywords: Augmented and virtual reality, Improving classroom teaching, Teaching/learning strategies * Correspondence: jianghualin1352@stu.xjtu.edu.cn; dingyi.007@163.com Jingjie Zhao and Xin Liang Xu contributed equally to this work. Health Science Center, Xi’an Jiaotong University, Xian, China Department of Pharmacy, Xijing Hospital, Fourth Military Medical University, Xian, China Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Zhao et al. BMC Medical Education (2020) 20:127 Page 2 of 10 Background Three research questions guided this study: Anatomy is a visual science which is thought an important foundation for medical learning [1]. When studying anat- (1) Are the test scores improved using VR education as omy, the learners identify structures and their spatial rela- compared to the other teaching methods? tionships. Nonetheless, medical students often experience (2) Are the satisfaction levels higher in VR education as trouble acquiring adequate understanding of three dimen- compared to the other teaching methods? sional (3D) anatomy from graphic images, such as those (3) Do year of publication, country of study, subject of in textbooks and PowerPoint [2, 3]. So, it has become vital learning, intervention, comparator, and duration to create modern strategies concentrated on efficient as play a moderating role in the distinction? well as high-quality anatomy education and learning. With new learning tools developing, the health and Methods medical education system has started incorporating more Search strategy interactive media and online materials. The utilization of This study adhered to the PRISMA criteria [11]. Search computer-based 3D models in anatomy education has be- terms for OVID MEDLINE was firstly performed and come a favorite over the last years [4]. Notably, VR is a after that adjusted for the others: Embase, Cochrane technology that allows exploring and manipulating Central Register of Controlled Trials, Web of Science computer-generated real or artificial 3D multimedia envi- Core Collection, and clinical trial registries. Terms as ronments in real-time. It allows for a first-person active well as subheadings such as key terms (anatomy) AND learning experience through different levels of immersion. (virtual reality OR virtual learning environment OR The rise of virtual reality technology could be traced back mixed reality OR virtual classrooms OR augmented real- to the 1960’s in the entertainment industry. VR promises ity OR visualization technologies) AND (educat* OR to provide more immersive, engaging experiences, with simulat* OR training). Databases were searched from applications in many domains, including shopping, enter- January 1990 to August 2019. tainment, training, and education [5]. Developers have The search results from various databases were incorpo- created compelling experiences allowing people to travel rated with Endnote software (EndNote X7, Clarivate Ana- through the cells of the body, to explore the Solar System, lytics, Philadelphia), and duplications of included studies and to encounter recreations of ancient battles in history. were eliminated. Two authors (Y.D. and J.J.Z) separately Particularly, virtual reality technologies frequently were screened the search results as well as examined full-text used for flight simulator training and exercises [6]. research studies for inclusion. Any kind of disputes, for Recently, increasing interest has been paid to VR in unclear or missing information were settled via conversa- the medical educational world, particularly for anatomy tion between the authors. teaching and resident surgical training [7, 8]. VR pro- vides students a simulation scene to conceptualize intri- Inclusion and exclusion criteria cate 3D anatomic connections quickly. Some studies We included randomized controlled studies on comparing have compared VR to the other teaching methods for and studying VR intervention with control methods in anatomy such as dissection, lectures, 2D images, and anatomy teaching. In this review, VR methods including blended instruction. For example, in 2019 Maresky et al. types of interactive 3D models, virtual patient or and surgi- tested the effectiveness of a VR simulation of the heart cal simulation could be performed as the single interven- in medical teaching [9]. They found that students (n = tion or blended with others [12]. VR as an intervention for 28) under the VR simulation performed significantly bet- education can be displayed with a variety of tools, including ter than the control group (n = 14) in the final test. In computer or mobile device screens, and VR rooms of head- 2015, a meta-analysis was conducted to evaluate teach- mounted displays. Studies were excluded with the following ing effect of using 3D visualization approaches in educa- reasons: not randomized controlled study; not in the field tional anatomy [10]. The results showed that 3D of anatomy education, absence of an intervention; absence visualization methods are better teaching tools than 2D of test scores; insufficient data for effect size calculation. Ex- methods in the acquisition of factual anatomy know- clusion was conducted by Y.D. and J.J.Z, and inconformity ledge and spatial anatomy knowledge. However, there is was discussed and resolved. The Kappa score was used to no high level of evidence on how efficient these different calculate the inter-investigator agreement during the inclu- VR approaches are when contrasted to various other sion process for publication-evaluated databases. techniques in randomized controlled studies. Accordingly, the purpose of this meta-analysis was to Data extraction explored the educational effectiveness of VR applied to We extracted data from validity studies according to the anatomy education in comparison with conventional or Cochrane Handbook for Systematic Reviews [11]. In this 2D digital methods in class. review, the main concerned information covered year and Zhao et al. BMC Medical Education (2020) 20:127 Page 3 of 10 region of the publication, details of learners, interventions, duration. Sensitivity analyses was conducted to determine if and duration of the study. Both of authors (J.J.Z and Y.D.) the individual study significantly altered the results of meta- assessed the risk of bias for randomized controlled trials analyses [16]. Publication bias was determined by a funnel by the Cochrane risk of bias tool [13]. plot [17]and Begg’stest[18]. The p value < 0.05 was defined as significant. Data synthesis and heterogeneity assessment All analyses were conducted by Stata 15 (StataCorp, Col- Results lege Station, TX, USA). Comparators included trad- Search results itional education, other forms of digital education, and Overall, 15 studies met the inclusion requirements (Fig. 1 other types of VR. For continuous data of test scores and Table 1). There were 15 randomized controlled and satisfaction levels, we summarized the standardized studies with an overall of 816 learners: 745 were medical mean differences (SMDs) and associated 95% confidence students and 71 were residents. There were seven stud- interval (CI) across studies. We were unable to identify a ies performed in USA, two studies in UK, two studies in clinically meaningful interpretation of SMDs for differ- Canada, and one each in Brazil, Australia and Japan. A ent kinds of VR education interventions. Therefore, the series of VR educational methods were evaluated, in- effect size was determined by the value of SMDs based cluding interactive 3D models, VR or and VR surgical on the Cohen rules: < 0.2 (none), 0.2 to 0.5 (small), 0.5 stimulations. Interventions in the control group ranged to 0.8 (moderate), and > 0.80 (large) [14]. We applied I from traditional learning (lecture, dissection and/or text- statistic to determine heterogeneity. I < 25% (low), 25 to books) to other digital education interventions. The dur- 75% (medium), and > 75% (high) indicate different levels ation of the intervention varied between 10 min to 2 of heterogeneity [15]. The fixed effect model was used to weeks. For all research studies, primary results were de- pool data if there was no heterogeneity (I > 50%); other- termined by evaluation or survey studies at the end. And wise, the random effects model was used (I < 50%). five out of 15 studies assessing satisfaction levels as the Subgroup analysis was conducted when feasible. Seven at- secondary outcome [23, 24, 30, 31, 33]. Table 1 shows tributes of each random were coded as possible moderators: the study characteristics of involved studies. year, region, learners, course, intervention, comparator, and Fig. 1 Flowchart of the search strategy Zhao et al. BMC Medical Education (2020) 20:127 Page 4 of 10 Table 1 Characteristics of included studies First author Participants/Country N (VR/ Course Intervention Comparator Duration control) Anthony, 2011 [19] medical students/UK 12/14 anatomy of the VR dissection and 50 min forearm textbooks Battulga, 2012 [20] medical students/Japan 50/50 shoulder 3D interactive 2D images 60 min models de Faria, 2016 [21] medical students/Brazil 28/28 neuroanatomy 3D interactive 2D images 60 min models Ellington, 2018 [22] residents/UK 16/15 female pelvic VR power point 2 weeks anatomy Hampton, 2010 [23] medical students 3, 21/22 female pelvic 3D interactive dissection and 60 min 4 year /USA anatomy models textbooks Keedy, 2011 [24] medical students 1, 23/23 anatomy of the liver 3D interactive 2D images 1 day 4 year/USA models Khot, 2013 [25] medical students/Canada 20/20 pelvic anatomy VR power point 10 min Kockro, 2015 [26] medical students/Germany 89/80 spatial neuroanatomy 3D interactive power point 20 min models Moro, 2017 [27] medical students/Australia 20/22 skull anatomy VR 3D models 10 min Nicholson, 2004 [28] medical students 29/28 ear anatomy 3D interactive text books 2 day 1 year /USA models Seixas, 2010 [29] surgical trainees/USA 5/5 human anatomy VR 2D images 1 day Solyar, 2008 [30] medical student/USA 7/8 paranasal sinuses VR textbooks 60 min Stepan, 2017 [31] medical students 1,2 year 33/33 neuroanatomy VR text books 1 day /USA Tan, 2012 [32] residents/ Canada 21/19 laryngeal anatomy 3D interactive text books 45 min models Zachary, 2015 [33] medical students/USA 41/32 neuroanatomy 3D interactive 2D images and 3D 65 min models models Inter-investigator agreement Data analysis The inter-investigator agreement (Kappa) was calculated The meta-analysis plots of primary and secondary out- by evaluating the selected titles and abstracts, and then comes are shown in Fig. 3a and b. The effectiveness of obtaining a value for selected articles (kappa = 0.92) pre- intervention on examination scores was reported in all senting a high level of agreement between the reviewers studies. The studies assessed test scores as a primary out- under the Kappa criteria [14]. come with multiple-choice questionnaires. We found that VR significantly increased learners’ examination scores compared with traditional learning in the random-effects Risk of bias assessment model (SMD = 0.53; 95% CI 0.09–0.97, p <0.05; I = The risk of bias in majority of studies involved was un- 87.8%) (Fig. 3a). Nine of the studies (60%) showed that VR clear or high risk as shown in the bias summary (Fig. 2). significantly increased students’ examination scores when Most studies did not have information about allocation compared with traditional learning (lecture, dissection concealment and baseline of learners’ characteristics. and/or textbooks) to other digital 2D methods; and five Due to the nature of the intervention, it is not practical (15%) failed to reveal statistically significant effects be- for blinding of students and teachers during the study. tween the VR and the control groups. Outcomes showed For risk of completeness of data, and selective reporting, that the studies were heterogeneous (p < 0.001) and the most studies were determined low. It was assessed true effects were not consistent among studies. whether the research study was devoid of selective out- A total of five studies assessed satisfaction levels as a come reporting, which checked whether outcomes men- secondary outcome [23, 24, 30, 31, 33]. The pooled results tioned adequately in manuscripts. Five studies were based on the fixed effects model showed that most stu- judged to be of high risk on completeness of data be- dents have a greater interest in learning via VR methods, cause of incomplete or accurate data on outcome stand- rather than conventional or 2D teaching methods (SMD = ard deviation [19, 24, 25, 29]. 0.77; 95% CI 0.47–1.07, p <0.05; I = 20.5%). However, Zhao et al. BMC Medical Education (2020) 20:127 Page 5 of 10 the funnel plot was found to be symmetrical. Meanwhile, the result of Begg’s test show a non-significant asym- metry (p = 0.54) [34]. Thus, there was no significant pub- lication bias indicated in this review. Subgroup analyses A random-effects model was used for the subgroup ana- lysis due to each subgroup being heterogeneous according to the results of tests (Table 2)[35]. As indicated in Table 2, the categorical variables were as follows: region (USA or others), learners (medical students or residents), course (skeletal anatomy or neuroanatomy or others), interven- tion (3D interactive models or VR simulations), compara- tor (traditional methods or other digital methods) and duration (< 1 day or ≥ 1 day). Other potential moderators could not be analyzed because they were reported inad- equately to do a subgroup analysis. The differences in the subgroups for Q statistics are non-significant (I >75%). Interestingly, the moderator analysis revealed significant benefits of VR in the subgroup of medical students (SMD = 0.51; 95% CI 0.02–1.01, p =0.04), whereas VR have no significant influence on residents (SMD = 0.67; 95% CI -0.45–1.01, p = 0.24). Also, moderator analysis of control type showed that test scores of the VR group was not significantly better than using other 2D digital methods (SMD = 0.35; 95% CI -0.25–0.95, p =0.25), while there was a significant improvement when compared with the traditional intervention group (SMD = 0.81; 95% CI 0.15–1.47, p = 0.02). For the duration analysis, VR inter- ventions for at least 1 day had moderately-to-large effects on scores (SMD = 0.71; 95% CI 0.42–1.10, p <0.001), whereas those which were < 1 day had only a small effect (SMD = 0.35; 95% CI 0.18–0.52, p <0.001). Meta-regression analyses To determine whether there were any moderation effects on primary outcomes, meta-regression analyses were con- ducted. We regressed effect sizes on 7 potential moderators: year, country, learners, course, intervention, comparator, and duration. As shown in Table 3,noneofthe moderators were significant at a level of p <0.05. Sensitivity analyses Due to the significant heterogeneity (> 75%), a sensitivity analysis was used to verify the reliability of the result. Fig. 2 Risk of bias assessment of included studies When any research was removed from the model, the significant results of the VR effect on examination scores only one study mentioned the adverse effects that some were unchanged in the models (SMD = 0.53, 95% CI: participants using VR displayed, including headaches, diz- 0.01–1.07) (Fig. 5). Thus, the results indicated that the ziness, or blurred vision [27]. findings for examination scores were robust. Publication bias Discussion For the primary analyses, funnel-plots were made to This meta-analysis of randomized controlled studies was check for risk of publication bias (Fig. 4). The shape of conducted to examine the effectiveness of VR-based Zhao et al. BMC Medical Education (2020) 20:127 Page 6 of 10 Fig. 3 Forest plots for examination scores (a) and satisfaction outcomes (b) technology in anatomy teaching. We found that VR inter- satisfaction scores as a secondary outcome with a result ventions have a moderate enhancement (SMD = 0.53) in that most of students more interested in using VR to learn test scores of learners in comparation with conventional anatomy. Naturally, the fact that no included randomized or other 2D digital methods (p < 0.01). As has been previ- controlled studies were found in databases before 2004 ously found, more interactive interventions could moder- suggested that VR was an emerging academic method ately improve medical learners’ academic scores in [37], attracting increasing interest from the world of edu- anatomy [36]. Among 15 studies, only five studies assessed cation. In general, the risk of bias for most studies was Zhao et al. BMC Medical Education (2020) 20:127 Page 7 of 10 Fig. 4 Funnel plot analysis for examination scores unclear for a lack of description or data. Potentially high learners and interventions in researches in this review, in- risk of incomplete reporting bias was identified in some consistent methodological method makes it difficult to studies. However, results of sensitivity and subgroup ana- draw accurate conclusions. lyses were nonsignificant for variables (year, country, In the subgroup analysis for levels of learners, the learners, course, intervention, comparator, and duration) source of high heterogeneity could be diverse phases of on the outcome variables. Since the different types of participants’ medical education among included studies. Learners are first-year medical students in two studies [28, 31], while learners in another two studies are forth- Table 2 Summary statistics for moderators related to year medical students [23, 24]. Of course, the longer examination scores learners acquired more knowledge of anatomy, which Subgroup n SMD 95% CI p value I leads to comparing results complex or paradoxical. As region Hattie et al. had concluded in 2015, the different degrees USA 7 1.14 0.56, 1.72 0.00 79.8% of expertise of learners are remarkable in education [38]. others 8 0.03 −0.57, 0.63 0.92 89.6% Therefore, medical students could be more easily moti- vated and effective in front of the fictitious scenarios of learners VR because they have fewer clinical experiences com- medical students 12 0.51 0.02, 1.01 0.04 89.6% pared to residents. In addition, various organs or body residents 3 0.67 −0.45, 1.79 0.24 77.8% parts learned present different levels of complexity, lead- course ing to the heterogeneity in results. For example, learning skeletal anatomy 6 −0.07 −0.95, 0.81 0.88 91.4% the anatomy of the brain was demonstrated harder than neuroanatomy 4 0.52 −0.04, 1.10 0.07 84.9% learning skeletal parts [31, 33]. In terms of duration, the results of this review showed that a course for 1 day or others 5 1.34 0.52, 2.14 0.00 87.8% longer had a larger effect size than a course for several intervention hours (0.71 vs 0.35). Thus, the learning duration has in- 3D interactive models 8 0.64 0.47, 0.81 0.00 82.5% fluenced the educational efficiency of VR methods, VR 7 −0.09 −0.37, 0.18 0.50 89.2% which should be considered and adjusted in practice. comparator Types of comparator is another source of variation. traditional methords 5 0.81 0.15, 1.47 0.02 82.6% Only five of 15 studies were found where this technology was compared to traditional methods such as lectures, dis- other digital methods 10 0.35 −0.25, 0.95 0.25 90.2% section or textbooks. However, it would be more mean- duration ingful to conduct evaluations of studies that compare the < 1 day 10 0.35 0.18, 0.52 0.00 89.4% different features of digital-based methods rather than ≥1 day 5 0.71 0.42, 1.10 0.00 84.4% those which compare digital-based to traditional methods Zhao et al. BMC Medical Education (2020) 20:127 Page 8 of 10 Table 3 Meta-regression analysis for exploration of the sources third of participants found the VR methods disorienting of heterogeneity factors and frustrating [27]. Using virtual reality could result in Factors Coefficient Standard error 95% CI p value cybersickness, such as nausea, disorientation and head- ache [41]. Thus, more studies should focus on the ad- year −0.12 0.20 −3.06, 0.67 0.21 verse effects such as blurred-vision and disorientation country −1.19 0.95 −2.99, 0.54 0.17 caused by VR. learners 1.08 1.24 −1.35, 3.52 0.38 As a fast-moving technology, the cost of VR will be a course −0.26 0.89 −2.01, 1.49 0.77 critical aspect when considering to apply it into education intervention −0.33 0.79 −0.53, 0.27 0.67 especially for low-income settings. In this review, only one comparator 0.29 0.86 −1.40, 1.99 0.73 study is from a lower income setting [21], which reduces the applicability of innovative educational methods to de- duration 0.09 0.95 −1.77, 1.97 0.91 veloping regions. Unfortunately, no randomized con- trolled studies reported on cost-effectiveness of VR [19]. Dissection is regarded as the standard teaching compared with other teaching methods. method for anatomy. In this review, only two of 15 studies compared VR with dissection for anatomy teaching [19, Strengths and limitations 23]. In fact, VR could be used as an adjunct to dissection VR is currently a new visualization technique, so there in class with fewer lab hours or resources. For example, in was no high-quality evidence on the effectiveness of VR- 2006 SN Biasutto et al. demonstrated that the best possi- based technology. It is hard to offer an overall conclu- bility in teaching anatomy is the correct association of ca- sion of the efficacy of these strategies. The strengths of daver dissections and computerized resources based on this meta-analysis included detailed search on random- their studies [39]. ized controlled studies, and the data was drawn out by For satisfaction scores, the pooled results of the com- two of authors independently. Because of the variability parison of VR versus others was significantly in favor of in studies, we also assessed the risk of bias, sensitivity VR, which could be due in part to the novelty of the analyses and meta-regression analyses on outcomes from method. Most of the participants in the studies reported articles. Results of sensitivity and subgroup analyses that the VR methods were easier and more enjoyable to were nonsignificant, indicating that the findings were use. In 2011, researchers had revealed that there was a robust. significant positive correlation between motivation and This review also has several limitations. First, the in- academic record of students [40]. However, due to the cluded researches mainly reported post-intervention infor- complicated anatomical configuration, in one study, one mation, so we did not compute pre-to post-intervention Fig. 5 Sensitivity analysis assessing the influence of each study on the pooled analysis Zhao et al. BMC Medical Education (2020) 20:127 Page 9 of 10 modification. The validity of the different assessments Received: 9 January 2020 Accepted: 4 March 2020 used in the included studies might constitute a bias. Gen- der information was not easy to obtain in the current meta-analysis, but it is an important factor influencing References 1. McLachlan JC, Patten D. Anatomy teaching: ghosts of the past, present and teaching effect [42]. In future studies, information on gen- future. Med Educ. 2006;40(3):243–53. der ratio for treatment and control group may be collected 2. Wainman B, Wolak L, Pukas G, Zheng E, Norman GR. The superiority of for analysis. Another limitation was that, none of the stud- three-dimensional physical models to two-dimensional computer presentations in anatomy learning. Med Educ. 2018;52(11):1138–46. ies assessed the cost of setup and maintenance of the VR- 3. Yammine K, Violato C. The effectiveness of physical models in teaching based intervention. Further research should evaluate the anatomy: a meta-analysis of comparative studies. Adv Health Sci Educ. 2016; effectiveness of VR in a variety of settings and evaluate 21(4):883–95. outcomes such as attitude, adverse effects, and cost- 4. Sugand K, Abrahams P, Khurana A. The anatomy of anatomy: a review for its modernization. Anat Sci Educ. 2010;3(2):83–93. effectiveness. 5. Hu-Au E, Lee JJ. Virtual reality in education: a tool for learning in the experience age. Int J Innov Educ. 2017;4(4):215–26. 6. Hawkins DG. Virtual reality and passive simulators: the future of fun. In Conclusions Biocca F., & Levy MR (Eds.), Communication in the age of virtual reality. As an emerging and new technology, VR has the potential Hillsdale: Lawrence Erlbaum; 1995. p. 159–89. in transforming medical teaching. In this meta-analysis, 7. Pilot A. Virtual reality and cardiac anatomy: exploring; 2018. 8. Erolin C, Reid L, McDougall S. Using virtual reality to complement and results showed that when compared with traditional or enhance anatomy education. J Vis Commun Med. 2019;42(3):1–9. 2D digital methods, VR can potentially improve teaching 9. Maresky H, Oikonomou A, Ali I, Ditkofsky N, Pakkal M, Ballyk B. Virtual reality effectiveness of anatomy. However, our results are not cer- and cardiac anatomy: exploring immersive three-dimensional cardiac imaging, a pilot study in undergraduate medical anatomy education. Clin tain for lack of standardized measures and high hetero- Anat. 2019;32(2):238–43. geneity among studies, and the appropriate mode of 10. Yammine K, Violato C. A meta-analysis of the educational effectiveness of integrating VR into class needs to be further explored. To three-dimensional visualization technologies in teaching anatomy. Anat Sci enhance the teaching quality, VR as an implement could Educ. 2015;8(6):525–38. 11. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for be considered on the medical teaching situations by uni- systematic reviews and meta-analyses: the PRISMA statement. Ann Intern versities and hospitals. Med. 2009;151(4):264–9. 12. Fernández-Palacios BJ, Morabito D, Remondino F. Access to complex reality- Abbreviations based 3D models using virtual reality solutions. J Cult Herit. 2017;23:40–8. CI: Confidence Interval; VR: Virtual reality; SMD: Standardized mean difference; 13. Higgins JP: Cochrane handbook for systematic reviews of interventions 3D: three-dimensional version 5.0. 1. The Cochrane Collaboration. 2008. http://www.cochrane- handbook.org. Acknowledgments 14. Landis JR, Koch GG. The measurement of observer agreement for This work was funded by research fund program of Xijing university categorical data. Biometrics. 1977;33(1):159–74. (XJ180203). The authors also thank the editor William Brooks, and reviewers 15. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in Caroline Erolin and Melissa Taylor for their detailed and constructive meta-analyses. BMJ. 2003;327(7414):557–60. comments that greatly improved the manuscript. 16. Kong Y, Seo YS, Zhai L. Comparison of reading performance on screen and on paper: a meta-analysis. Comput Educ. 2018;123:138–49. Authors’ contributions 17. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of HLJ and XLX involved in designing the study, collecting the data, analyzing testing and adjusting for publication bias in meta-analysis. Biometrics. 2000; the data, and writing the manuscript. YD and JJZ contributed to the study 56(2):455–63. methodology and corrected the final version of the manuscript. The 18. Orwin RG. A fail-safe N for effect size in meta-analysis. J Educ Stat. 1983;8(2): author(s) read and approved the final manuscript. 157–9. 19. Codd AM, Choudhury B. Virtual reality anatomy: is it comparable with Funding traditional methods in the teaching of human forearm musculoskeletal The authors declare that they have no source of funding for the research. anatomy? Anat Sci Educ. 2011;4(3):119–25. 20. Battulga B, Konishi T, Tamura Y, Moriguchi H. The Effectiveness of an Availability of data and materials Interactive 3-Dimensional Computer Graphics Model for Medical Education. The datasets used and/or analysed during the current study are available Interact J Med Res. 2012;1(2):e2. from the corresponding author on reasonable request. 21. de Faria JWV, Teixeira MJ, Júnior LD, Otoch JP, Figueiredo EG. Virtual and stereoscopic anatomy: when virtual reality meets medical education. J Ethics approval and consent to participate Neurosurg. 2016;125(5):1105–11. Not applicable. 22. Ellington DR, Shum PC, Dennis EA, Willis HL, Szychowski JM, Richter HE. Female Pelvic Floor Immersive Simulation: A Randomized Trial to Test the Consent for publication Effectiveness of a Virtual Reality Anatomic Model on Resident Knowledge of Not applicable. Female Pelvic Anatomy. J Minim Invasive Gynecol. 2019;26(5):897–901. 23. Hampton BS, Sung VW. Improving medical student knowledge of female Competing interests pelvic floor dysfunction and anatomy: a randomized trial. Am J Obstet The authors declare that they have no competing interests. Gynecol. 2010;202(6):601.e601–8. 24. Keedy AW, Durack JC, Sandhu P, Chen EM, O'Sullivan PS, Breiman RS. Author details Comparison of traditional methods with 3D computer models in the 1 2 School of Media, Xijing University, Xian, China. Department of Traumatic instruction of hepatobiliary anatomy. Anat Sci Educ. 2011;4(2):84–91. Surgery, Jining No.1 Peoples Hospital, Jining, China. Health Science Center, 25. Khot Z, Quinlan K, Norman GR, Wainman B. The relative effectiveness of Xi’an Jiaotong University, Xian, China. Department of Pharmacy, Xijing computer-based and traditional resources for education in anatomy. Anat Hospital, Fourth Military Medical University, Xian, China. Sci Educ. 2013;6(4):211–5. Zhao et al. BMC Medical Education (2020) 20:127 Page 10 of 10 26. Kockro RA, Amaxopoulou C, Killeen T, Wagner W, Reisch R, Schwandt E, Gutenberg A, Giese A, Stofft E, Stadie AT. Stereoscopic neuroanatomy lectures using a three-dimensional virtual reality environment. Ann Anat. 2015;201:91–8. 27. Moro C, Stromberga Z, Raikos A, Stirling A. The effectiveness of virtual and augmented reality in health sciences and medical anatomy. Anat Sci Educ. 2017;10(6):549–59. 28. Nicholson DT, Chalk C, Funnell WRJ, Daniel SJ. Can virtual reality improve anatomy education? A randomised controlled study of a computer- generated three-dimensional anatomical ear model. Med Educ. 2006;40(11): 1081–7. 29. Seixas-Mikelus SA, Adal A, Kesavadas T, Baheti A, Srimathveeravalli G, Hussain A, Chandrasekhar R, Wilding GE, Guru KA. Can image-based virtual reality help teach anatomy? J Endourol. 2010;24(4):629–34. 30. Solyar A, Cuellar H, Sadoughi B, Olson TR, Fried MP. Endoscopic sinus surgery simulator as a teaching tool for anatomy education. Am J Surg. 2008;196(1):120–4. 31. Stepan K, Zeiger J, Hanchuk S, Del Signore A, Shrivastava R, Govindaraj S, Iloreta A. Immersive virtual reality as a teaching tool for neuroanatomy. Int Forum Allergy Rhinol. 2017;7(10):1006–13. 32. Tan S, Hu A, Wilson T, Ladak H, Haase P, Fung K. Role of a computer- generated three-dimensional laryngeal model in anatomy teaching for advanced learners. J Laryngol Otol. 2012;126(4):395–401. 33. Drapkin ZA, Lindgren KA, Lopez MJ, Stabio ME. Development and assessment of a new 3D neuroanatomy teaching tool for MRI training. Anat Sci Educ. 2015;8(6):502–9. 34. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50:1088–101. 35. Harris RJ, Deeks JJ, Altman DG, Bradburn MJ, Harbord RM, Sterne JA. Metan: fixed-and random-effects meta-analysis. Stata J. 2008;8(1):3–28. 36. Shiozawa T, Butz B, Herlan S, Kramer A, Hirt B. Interactive anatomical and surgical live stream lectures improve students' academic performance in applied clinical anatomy. Anat Sci Educ. 2017;10(1):46–52. 37. Liu D, Bhagat KK, Gao Y, Chang T-W, Huang R. The potentials and trends of virtual reality in education. In: Virtual, augmented, and mixed realities in education. Singapore: Springer; 2017. p. 105–30. 38. Hattie J. The applicability of visible learning to higher education. Scholarsh Teach Learn Psychol. 2015;1(1):79. 39. Biasutto SN, Caussa LI, del Río LEC. Teaching anatomy: cadavers vs. computers? Ann Anat. 2006;188(2):187–90. 40. Kim K-J, Frick TW. Changes in student motivation during online learning. J Educ Comput Res. 2011;44(1):1–23. 41. Rebenitsch L, Owen C. Review on cybersickness in applications and visual displays. Virtual Reality. 2016;20(2):101–25. 42. Bleakley A. Gender matters in medical education. In: Patient-Centred Medicine in Transition: Springer, Cham; 2014. p. 111–26. Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Medical Education Springer Journals

The effectiveness of virtual reality-based technology on anatomy teaching: a meta-analysis of randomized controlled studies

Loading next page...
 
/lp/springer-journals/the-effectiveness-of-virtual-reality-based-technology-on-anatomy-d7NqDR1M00

References (49)

Publisher
Springer Journals
Copyright
Copyright © The Author(s). 2020
eISSN
1472-6920
DOI
10.1186/s12909-020-1994-z
Publisher site
See Article on Publisher Site

Abstract

Background: Virtual reality (VR) is an innovation that permits the individual to discover and operate within three- dimensional (3D) environment to gain practical understanding. This research aimed to examine the general efficiency of VR for teaching medical anatomy. Methods: We executed a meta-analysis of randomized controlled studies of the performance of VR anatomy education. We browsed five databases from the year 1990 to 2019. Ultimately, 15 randomized controlled trials with a teaching outcome measure analysis were included. Two authors separately chose studies, extracted information, and examined the risk of bias. The primary outcomes were examination scores of the students. Secondary outcomes were the degrees of satisfaction of the students. Random-effects models were used for the pooled evaluations of scores and satisfaction degrees. Standardized mean difference (SMD) was applied to assess the systematic results. The heterogeneity was determined by I statistics, and then was investigated by meta-regression and subgroup analyses. Results: In this review, we screened and included fifteen randomized controlled researches (816 students). The pooled analysis of primary outcomes showed that VR improves test scores moderately compared with other approaches (standardized mean difference [SMD] = 0.53; 95% Confidence Interval [CI] 0.09–0.97, p < 0.05; I = 87.8%). The high homogeneity indicated that the studies were different from each other. Therefore, we carried out meta- regression as well as subgroup analyses using seven variables (year, country, learners, course, intervention, comparator, and duration). We found that VR improves post-intervention test score of anatomy compared with other types of teaching methods. Conclusions: The finding confirms that VR may act as an efficient way to improve the learners’ level of anatomy knowledge. Future research should assess other factors like degree of satisfaction, cost-effectiveness, and adverse reactions when evaluating the teaching effectiveness of VR in anatomy. Keywords: Augmented and virtual reality, Improving classroom teaching, Teaching/learning strategies * Correspondence: jianghualin1352@stu.xjtu.edu.cn; dingyi.007@163.com Jingjie Zhao and Xin Liang Xu contributed equally to this work. Health Science Center, Xi’an Jiaotong University, Xian, China Department of Pharmacy, Xijing Hospital, Fourth Military Medical University, Xian, China Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Zhao et al. BMC Medical Education (2020) 20:127 Page 2 of 10 Background Three research questions guided this study: Anatomy is a visual science which is thought an important foundation for medical learning [1]. When studying anat- (1) Are the test scores improved using VR education as omy, the learners identify structures and their spatial rela- compared to the other teaching methods? tionships. Nonetheless, medical students often experience (2) Are the satisfaction levels higher in VR education as trouble acquiring adequate understanding of three dimen- compared to the other teaching methods? sional (3D) anatomy from graphic images, such as those (3) Do year of publication, country of study, subject of in textbooks and PowerPoint [2, 3]. So, it has become vital learning, intervention, comparator, and duration to create modern strategies concentrated on efficient as play a moderating role in the distinction? well as high-quality anatomy education and learning. With new learning tools developing, the health and Methods medical education system has started incorporating more Search strategy interactive media and online materials. The utilization of This study adhered to the PRISMA criteria [11]. Search computer-based 3D models in anatomy education has be- terms for OVID MEDLINE was firstly performed and come a favorite over the last years [4]. Notably, VR is a after that adjusted for the others: Embase, Cochrane technology that allows exploring and manipulating Central Register of Controlled Trials, Web of Science computer-generated real or artificial 3D multimedia envi- Core Collection, and clinical trial registries. Terms as ronments in real-time. It allows for a first-person active well as subheadings such as key terms (anatomy) AND learning experience through different levels of immersion. (virtual reality OR virtual learning environment OR The rise of virtual reality technology could be traced back mixed reality OR virtual classrooms OR augmented real- to the 1960’s in the entertainment industry. VR promises ity OR visualization technologies) AND (educat* OR to provide more immersive, engaging experiences, with simulat* OR training). Databases were searched from applications in many domains, including shopping, enter- January 1990 to August 2019. tainment, training, and education [5]. Developers have The search results from various databases were incorpo- created compelling experiences allowing people to travel rated with Endnote software (EndNote X7, Clarivate Ana- through the cells of the body, to explore the Solar System, lytics, Philadelphia), and duplications of included studies and to encounter recreations of ancient battles in history. were eliminated. Two authors (Y.D. and J.J.Z) separately Particularly, virtual reality technologies frequently were screened the search results as well as examined full-text used for flight simulator training and exercises [6]. research studies for inclusion. Any kind of disputes, for Recently, increasing interest has been paid to VR in unclear or missing information were settled via conversa- the medical educational world, particularly for anatomy tion between the authors. teaching and resident surgical training [7, 8]. VR pro- vides students a simulation scene to conceptualize intri- Inclusion and exclusion criteria cate 3D anatomic connections quickly. Some studies We included randomized controlled studies on comparing have compared VR to the other teaching methods for and studying VR intervention with control methods in anatomy such as dissection, lectures, 2D images, and anatomy teaching. In this review, VR methods including blended instruction. For example, in 2019 Maresky et al. types of interactive 3D models, virtual patient or and surgi- tested the effectiveness of a VR simulation of the heart cal simulation could be performed as the single interven- in medical teaching [9]. They found that students (n = tion or blended with others [12]. VR as an intervention for 28) under the VR simulation performed significantly bet- education can be displayed with a variety of tools, including ter than the control group (n = 14) in the final test. In computer or mobile device screens, and VR rooms of head- 2015, a meta-analysis was conducted to evaluate teach- mounted displays. Studies were excluded with the following ing effect of using 3D visualization approaches in educa- reasons: not randomized controlled study; not in the field tional anatomy [10]. The results showed that 3D of anatomy education, absence of an intervention; absence visualization methods are better teaching tools than 2D of test scores; insufficient data for effect size calculation. Ex- methods in the acquisition of factual anatomy know- clusion was conducted by Y.D. and J.J.Z, and inconformity ledge and spatial anatomy knowledge. However, there is was discussed and resolved. The Kappa score was used to no high level of evidence on how efficient these different calculate the inter-investigator agreement during the inclu- VR approaches are when contrasted to various other sion process for publication-evaluated databases. techniques in randomized controlled studies. Accordingly, the purpose of this meta-analysis was to Data extraction explored the educational effectiveness of VR applied to We extracted data from validity studies according to the anatomy education in comparison with conventional or Cochrane Handbook for Systematic Reviews [11]. In this 2D digital methods in class. review, the main concerned information covered year and Zhao et al. BMC Medical Education (2020) 20:127 Page 3 of 10 region of the publication, details of learners, interventions, duration. Sensitivity analyses was conducted to determine if and duration of the study. Both of authors (J.J.Z and Y.D.) the individual study significantly altered the results of meta- assessed the risk of bias for randomized controlled trials analyses [16]. Publication bias was determined by a funnel by the Cochrane risk of bias tool [13]. plot [17]and Begg’stest[18]. The p value < 0.05 was defined as significant. Data synthesis and heterogeneity assessment All analyses were conducted by Stata 15 (StataCorp, Col- Results lege Station, TX, USA). Comparators included trad- Search results itional education, other forms of digital education, and Overall, 15 studies met the inclusion requirements (Fig. 1 other types of VR. For continuous data of test scores and Table 1). There were 15 randomized controlled and satisfaction levels, we summarized the standardized studies with an overall of 816 learners: 745 were medical mean differences (SMDs) and associated 95% confidence students and 71 were residents. There were seven stud- interval (CI) across studies. We were unable to identify a ies performed in USA, two studies in UK, two studies in clinically meaningful interpretation of SMDs for differ- Canada, and one each in Brazil, Australia and Japan. A ent kinds of VR education interventions. Therefore, the series of VR educational methods were evaluated, in- effect size was determined by the value of SMDs based cluding interactive 3D models, VR or and VR surgical on the Cohen rules: < 0.2 (none), 0.2 to 0.5 (small), 0.5 stimulations. Interventions in the control group ranged to 0.8 (moderate), and > 0.80 (large) [14]. We applied I from traditional learning (lecture, dissection and/or text- statistic to determine heterogeneity. I < 25% (low), 25 to books) to other digital education interventions. The dur- 75% (medium), and > 75% (high) indicate different levels ation of the intervention varied between 10 min to 2 of heterogeneity [15]. The fixed effect model was used to weeks. For all research studies, primary results were de- pool data if there was no heterogeneity (I > 50%); other- termined by evaluation or survey studies at the end. And wise, the random effects model was used (I < 50%). five out of 15 studies assessing satisfaction levels as the Subgroup analysis was conducted when feasible. Seven at- secondary outcome [23, 24, 30, 31, 33]. Table 1 shows tributes of each random were coded as possible moderators: the study characteristics of involved studies. year, region, learners, course, intervention, comparator, and Fig. 1 Flowchart of the search strategy Zhao et al. BMC Medical Education (2020) 20:127 Page 4 of 10 Table 1 Characteristics of included studies First author Participants/Country N (VR/ Course Intervention Comparator Duration control) Anthony, 2011 [19] medical students/UK 12/14 anatomy of the VR dissection and 50 min forearm textbooks Battulga, 2012 [20] medical students/Japan 50/50 shoulder 3D interactive 2D images 60 min models de Faria, 2016 [21] medical students/Brazil 28/28 neuroanatomy 3D interactive 2D images 60 min models Ellington, 2018 [22] residents/UK 16/15 female pelvic VR power point 2 weeks anatomy Hampton, 2010 [23] medical students 3, 21/22 female pelvic 3D interactive dissection and 60 min 4 year /USA anatomy models textbooks Keedy, 2011 [24] medical students 1, 23/23 anatomy of the liver 3D interactive 2D images 1 day 4 year/USA models Khot, 2013 [25] medical students/Canada 20/20 pelvic anatomy VR power point 10 min Kockro, 2015 [26] medical students/Germany 89/80 spatial neuroanatomy 3D interactive power point 20 min models Moro, 2017 [27] medical students/Australia 20/22 skull anatomy VR 3D models 10 min Nicholson, 2004 [28] medical students 29/28 ear anatomy 3D interactive text books 2 day 1 year /USA models Seixas, 2010 [29] surgical trainees/USA 5/5 human anatomy VR 2D images 1 day Solyar, 2008 [30] medical student/USA 7/8 paranasal sinuses VR textbooks 60 min Stepan, 2017 [31] medical students 1,2 year 33/33 neuroanatomy VR text books 1 day /USA Tan, 2012 [32] residents/ Canada 21/19 laryngeal anatomy 3D interactive text books 45 min models Zachary, 2015 [33] medical students/USA 41/32 neuroanatomy 3D interactive 2D images and 3D 65 min models models Inter-investigator agreement Data analysis The inter-investigator agreement (Kappa) was calculated The meta-analysis plots of primary and secondary out- by evaluating the selected titles and abstracts, and then comes are shown in Fig. 3a and b. The effectiveness of obtaining a value for selected articles (kappa = 0.92) pre- intervention on examination scores was reported in all senting a high level of agreement between the reviewers studies. The studies assessed test scores as a primary out- under the Kappa criteria [14]. come with multiple-choice questionnaires. We found that VR significantly increased learners’ examination scores compared with traditional learning in the random-effects Risk of bias assessment model (SMD = 0.53; 95% CI 0.09–0.97, p <0.05; I = The risk of bias in majority of studies involved was un- 87.8%) (Fig. 3a). Nine of the studies (60%) showed that VR clear or high risk as shown in the bias summary (Fig. 2). significantly increased students’ examination scores when Most studies did not have information about allocation compared with traditional learning (lecture, dissection concealment and baseline of learners’ characteristics. and/or textbooks) to other digital 2D methods; and five Due to the nature of the intervention, it is not practical (15%) failed to reveal statistically significant effects be- for blinding of students and teachers during the study. tween the VR and the control groups. Outcomes showed For risk of completeness of data, and selective reporting, that the studies were heterogeneous (p < 0.001) and the most studies were determined low. It was assessed true effects were not consistent among studies. whether the research study was devoid of selective out- A total of five studies assessed satisfaction levels as a come reporting, which checked whether outcomes men- secondary outcome [23, 24, 30, 31, 33]. The pooled results tioned adequately in manuscripts. Five studies were based on the fixed effects model showed that most stu- judged to be of high risk on completeness of data be- dents have a greater interest in learning via VR methods, cause of incomplete or accurate data on outcome stand- rather than conventional or 2D teaching methods (SMD = ard deviation [19, 24, 25, 29]. 0.77; 95% CI 0.47–1.07, p <0.05; I = 20.5%). However, Zhao et al. BMC Medical Education (2020) 20:127 Page 5 of 10 the funnel plot was found to be symmetrical. Meanwhile, the result of Begg’s test show a non-significant asym- metry (p = 0.54) [34]. Thus, there was no significant pub- lication bias indicated in this review. Subgroup analyses A random-effects model was used for the subgroup ana- lysis due to each subgroup being heterogeneous according to the results of tests (Table 2)[35]. As indicated in Table 2, the categorical variables were as follows: region (USA or others), learners (medical students or residents), course (skeletal anatomy or neuroanatomy or others), interven- tion (3D interactive models or VR simulations), compara- tor (traditional methods or other digital methods) and duration (< 1 day or ≥ 1 day). Other potential moderators could not be analyzed because they were reported inad- equately to do a subgroup analysis. The differences in the subgroups for Q statistics are non-significant (I >75%). Interestingly, the moderator analysis revealed significant benefits of VR in the subgroup of medical students (SMD = 0.51; 95% CI 0.02–1.01, p =0.04), whereas VR have no significant influence on residents (SMD = 0.67; 95% CI -0.45–1.01, p = 0.24). Also, moderator analysis of control type showed that test scores of the VR group was not significantly better than using other 2D digital methods (SMD = 0.35; 95% CI -0.25–0.95, p =0.25), while there was a significant improvement when compared with the traditional intervention group (SMD = 0.81; 95% CI 0.15–1.47, p = 0.02). For the duration analysis, VR inter- ventions for at least 1 day had moderately-to-large effects on scores (SMD = 0.71; 95% CI 0.42–1.10, p <0.001), whereas those which were < 1 day had only a small effect (SMD = 0.35; 95% CI 0.18–0.52, p <0.001). Meta-regression analyses To determine whether there were any moderation effects on primary outcomes, meta-regression analyses were con- ducted. We regressed effect sizes on 7 potential moderators: year, country, learners, course, intervention, comparator, and duration. As shown in Table 3,noneofthe moderators were significant at a level of p <0.05. Sensitivity analyses Due to the significant heterogeneity (> 75%), a sensitivity analysis was used to verify the reliability of the result. Fig. 2 Risk of bias assessment of included studies When any research was removed from the model, the significant results of the VR effect on examination scores only one study mentioned the adverse effects that some were unchanged in the models (SMD = 0.53, 95% CI: participants using VR displayed, including headaches, diz- 0.01–1.07) (Fig. 5). Thus, the results indicated that the ziness, or blurred vision [27]. findings for examination scores were robust. Publication bias Discussion For the primary analyses, funnel-plots were made to This meta-analysis of randomized controlled studies was check for risk of publication bias (Fig. 4). The shape of conducted to examine the effectiveness of VR-based Zhao et al. BMC Medical Education (2020) 20:127 Page 6 of 10 Fig. 3 Forest plots for examination scores (a) and satisfaction outcomes (b) technology in anatomy teaching. We found that VR inter- satisfaction scores as a secondary outcome with a result ventions have a moderate enhancement (SMD = 0.53) in that most of students more interested in using VR to learn test scores of learners in comparation with conventional anatomy. Naturally, the fact that no included randomized or other 2D digital methods (p < 0.01). As has been previ- controlled studies were found in databases before 2004 ously found, more interactive interventions could moder- suggested that VR was an emerging academic method ately improve medical learners’ academic scores in [37], attracting increasing interest from the world of edu- anatomy [36]. Among 15 studies, only five studies assessed cation. In general, the risk of bias for most studies was Zhao et al. BMC Medical Education (2020) 20:127 Page 7 of 10 Fig. 4 Funnel plot analysis for examination scores unclear for a lack of description or data. Potentially high learners and interventions in researches in this review, in- risk of incomplete reporting bias was identified in some consistent methodological method makes it difficult to studies. However, results of sensitivity and subgroup ana- draw accurate conclusions. lyses were nonsignificant for variables (year, country, In the subgroup analysis for levels of learners, the learners, course, intervention, comparator, and duration) source of high heterogeneity could be diverse phases of on the outcome variables. Since the different types of participants’ medical education among included studies. Learners are first-year medical students in two studies [28, 31], while learners in another two studies are forth- Table 2 Summary statistics for moderators related to year medical students [23, 24]. Of course, the longer examination scores learners acquired more knowledge of anatomy, which Subgroup n SMD 95% CI p value I leads to comparing results complex or paradoxical. As region Hattie et al. had concluded in 2015, the different degrees USA 7 1.14 0.56, 1.72 0.00 79.8% of expertise of learners are remarkable in education [38]. others 8 0.03 −0.57, 0.63 0.92 89.6% Therefore, medical students could be more easily moti- vated and effective in front of the fictitious scenarios of learners VR because they have fewer clinical experiences com- medical students 12 0.51 0.02, 1.01 0.04 89.6% pared to residents. In addition, various organs or body residents 3 0.67 −0.45, 1.79 0.24 77.8% parts learned present different levels of complexity, lead- course ing to the heterogeneity in results. For example, learning skeletal anatomy 6 −0.07 −0.95, 0.81 0.88 91.4% the anatomy of the brain was demonstrated harder than neuroanatomy 4 0.52 −0.04, 1.10 0.07 84.9% learning skeletal parts [31, 33]. In terms of duration, the results of this review showed that a course for 1 day or others 5 1.34 0.52, 2.14 0.00 87.8% longer had a larger effect size than a course for several intervention hours (0.71 vs 0.35). Thus, the learning duration has in- 3D interactive models 8 0.64 0.47, 0.81 0.00 82.5% fluenced the educational efficiency of VR methods, VR 7 −0.09 −0.37, 0.18 0.50 89.2% which should be considered and adjusted in practice. comparator Types of comparator is another source of variation. traditional methords 5 0.81 0.15, 1.47 0.02 82.6% Only five of 15 studies were found where this technology was compared to traditional methods such as lectures, dis- other digital methods 10 0.35 −0.25, 0.95 0.25 90.2% section or textbooks. However, it would be more mean- duration ingful to conduct evaluations of studies that compare the < 1 day 10 0.35 0.18, 0.52 0.00 89.4% different features of digital-based methods rather than ≥1 day 5 0.71 0.42, 1.10 0.00 84.4% those which compare digital-based to traditional methods Zhao et al. BMC Medical Education (2020) 20:127 Page 8 of 10 Table 3 Meta-regression analysis for exploration of the sources third of participants found the VR methods disorienting of heterogeneity factors and frustrating [27]. Using virtual reality could result in Factors Coefficient Standard error 95% CI p value cybersickness, such as nausea, disorientation and head- ache [41]. Thus, more studies should focus on the ad- year −0.12 0.20 −3.06, 0.67 0.21 verse effects such as blurred-vision and disorientation country −1.19 0.95 −2.99, 0.54 0.17 caused by VR. learners 1.08 1.24 −1.35, 3.52 0.38 As a fast-moving technology, the cost of VR will be a course −0.26 0.89 −2.01, 1.49 0.77 critical aspect when considering to apply it into education intervention −0.33 0.79 −0.53, 0.27 0.67 especially for low-income settings. In this review, only one comparator 0.29 0.86 −1.40, 1.99 0.73 study is from a lower income setting [21], which reduces the applicability of innovative educational methods to de- duration 0.09 0.95 −1.77, 1.97 0.91 veloping regions. Unfortunately, no randomized con- trolled studies reported on cost-effectiveness of VR [19]. Dissection is regarded as the standard teaching compared with other teaching methods. method for anatomy. In this review, only two of 15 studies compared VR with dissection for anatomy teaching [19, Strengths and limitations 23]. In fact, VR could be used as an adjunct to dissection VR is currently a new visualization technique, so there in class with fewer lab hours or resources. For example, in was no high-quality evidence on the effectiveness of VR- 2006 SN Biasutto et al. demonstrated that the best possi- based technology. It is hard to offer an overall conclu- bility in teaching anatomy is the correct association of ca- sion of the efficacy of these strategies. The strengths of daver dissections and computerized resources based on this meta-analysis included detailed search on random- their studies [39]. ized controlled studies, and the data was drawn out by For satisfaction scores, the pooled results of the com- two of authors independently. Because of the variability parison of VR versus others was significantly in favor of in studies, we also assessed the risk of bias, sensitivity VR, which could be due in part to the novelty of the analyses and meta-regression analyses on outcomes from method. Most of the participants in the studies reported articles. Results of sensitivity and subgroup analyses that the VR methods were easier and more enjoyable to were nonsignificant, indicating that the findings were use. In 2011, researchers had revealed that there was a robust. significant positive correlation between motivation and This review also has several limitations. First, the in- academic record of students [40]. However, due to the cluded researches mainly reported post-intervention infor- complicated anatomical configuration, in one study, one mation, so we did not compute pre-to post-intervention Fig. 5 Sensitivity analysis assessing the influence of each study on the pooled analysis Zhao et al. BMC Medical Education (2020) 20:127 Page 9 of 10 modification. The validity of the different assessments Received: 9 January 2020 Accepted: 4 March 2020 used in the included studies might constitute a bias. Gen- der information was not easy to obtain in the current meta-analysis, but it is an important factor influencing References 1. McLachlan JC, Patten D. Anatomy teaching: ghosts of the past, present and teaching effect [42]. In future studies, information on gen- future. Med Educ. 2006;40(3):243–53. der ratio for treatment and control group may be collected 2. Wainman B, Wolak L, Pukas G, Zheng E, Norman GR. The superiority of for analysis. Another limitation was that, none of the stud- three-dimensional physical models to two-dimensional computer presentations in anatomy learning. Med Educ. 2018;52(11):1138–46. ies assessed the cost of setup and maintenance of the VR- 3. Yammine K, Violato C. The effectiveness of physical models in teaching based intervention. Further research should evaluate the anatomy: a meta-analysis of comparative studies. Adv Health Sci Educ. 2016; effectiveness of VR in a variety of settings and evaluate 21(4):883–95. outcomes such as attitude, adverse effects, and cost- 4. Sugand K, Abrahams P, Khurana A. The anatomy of anatomy: a review for its modernization. Anat Sci Educ. 2010;3(2):83–93. effectiveness. 5. Hu-Au E, Lee JJ. Virtual reality in education: a tool for learning in the experience age. Int J Innov Educ. 2017;4(4):215–26. 6. Hawkins DG. Virtual reality and passive simulators: the future of fun. In Conclusions Biocca F., & Levy MR (Eds.), Communication in the age of virtual reality. As an emerging and new technology, VR has the potential Hillsdale: Lawrence Erlbaum; 1995. p. 159–89. in transforming medical teaching. In this meta-analysis, 7. Pilot A. Virtual reality and cardiac anatomy: exploring; 2018. 8. Erolin C, Reid L, McDougall S. Using virtual reality to complement and results showed that when compared with traditional or enhance anatomy education. J Vis Commun Med. 2019;42(3):1–9. 2D digital methods, VR can potentially improve teaching 9. Maresky H, Oikonomou A, Ali I, Ditkofsky N, Pakkal M, Ballyk B. Virtual reality effectiveness of anatomy. However, our results are not cer- and cardiac anatomy: exploring immersive three-dimensional cardiac imaging, a pilot study in undergraduate medical anatomy education. Clin tain for lack of standardized measures and high hetero- Anat. 2019;32(2):238–43. geneity among studies, and the appropriate mode of 10. Yammine K, Violato C. A meta-analysis of the educational effectiveness of integrating VR into class needs to be further explored. To three-dimensional visualization technologies in teaching anatomy. Anat Sci enhance the teaching quality, VR as an implement could Educ. 2015;8(6):525–38. 11. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for be considered on the medical teaching situations by uni- systematic reviews and meta-analyses: the PRISMA statement. Ann Intern versities and hospitals. Med. 2009;151(4):264–9. 12. Fernández-Palacios BJ, Morabito D, Remondino F. Access to complex reality- Abbreviations based 3D models using virtual reality solutions. J Cult Herit. 2017;23:40–8. CI: Confidence Interval; VR: Virtual reality; SMD: Standardized mean difference; 13. Higgins JP: Cochrane handbook for systematic reviews of interventions 3D: three-dimensional version 5.0. 1. The Cochrane Collaboration. 2008. http://www.cochrane- handbook.org. Acknowledgments 14. Landis JR, Koch GG. The measurement of observer agreement for This work was funded by research fund program of Xijing university categorical data. Biometrics. 1977;33(1):159–74. (XJ180203). The authors also thank the editor William Brooks, and reviewers 15. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in Caroline Erolin and Melissa Taylor for their detailed and constructive meta-analyses. BMJ. 2003;327(7414):557–60. comments that greatly improved the manuscript. 16. Kong Y, Seo YS, Zhai L. Comparison of reading performance on screen and on paper: a meta-analysis. Comput Educ. 2018;123:138–49. Authors’ contributions 17. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of HLJ and XLX involved in designing the study, collecting the data, analyzing testing and adjusting for publication bias in meta-analysis. Biometrics. 2000; the data, and writing the manuscript. YD and JJZ contributed to the study 56(2):455–63. methodology and corrected the final version of the manuscript. The 18. Orwin RG. A fail-safe N for effect size in meta-analysis. J Educ Stat. 1983;8(2): author(s) read and approved the final manuscript. 157–9. 19. Codd AM, Choudhury B. Virtual reality anatomy: is it comparable with Funding traditional methods in the teaching of human forearm musculoskeletal The authors declare that they have no source of funding for the research. anatomy? Anat Sci Educ. 2011;4(3):119–25. 20. Battulga B, Konishi T, Tamura Y, Moriguchi H. The Effectiveness of an Availability of data and materials Interactive 3-Dimensional Computer Graphics Model for Medical Education. The datasets used and/or analysed during the current study are available Interact J Med Res. 2012;1(2):e2. from the corresponding author on reasonable request. 21. de Faria JWV, Teixeira MJ, Júnior LD, Otoch JP, Figueiredo EG. Virtual and stereoscopic anatomy: when virtual reality meets medical education. J Ethics approval and consent to participate Neurosurg. 2016;125(5):1105–11. Not applicable. 22. Ellington DR, Shum PC, Dennis EA, Willis HL, Szychowski JM, Richter HE. Female Pelvic Floor Immersive Simulation: A Randomized Trial to Test the Consent for publication Effectiveness of a Virtual Reality Anatomic Model on Resident Knowledge of Not applicable. Female Pelvic Anatomy. J Minim Invasive Gynecol. 2019;26(5):897–901. 23. Hampton BS, Sung VW. Improving medical student knowledge of female Competing interests pelvic floor dysfunction and anatomy: a randomized trial. Am J Obstet The authors declare that they have no competing interests. Gynecol. 2010;202(6):601.e601–8. 24. Keedy AW, Durack JC, Sandhu P, Chen EM, O'Sullivan PS, Breiman RS. Author details Comparison of traditional methods with 3D computer models in the 1 2 School of Media, Xijing University, Xian, China. Department of Traumatic instruction of hepatobiliary anatomy. Anat Sci Educ. 2011;4(2):84–91. Surgery, Jining No.1 Peoples Hospital, Jining, China. Health Science Center, 25. Khot Z, Quinlan K, Norman GR, Wainman B. The relative effectiveness of Xi’an Jiaotong University, Xian, China. Department of Pharmacy, Xijing computer-based and traditional resources for education in anatomy. Anat Hospital, Fourth Military Medical University, Xian, China. Sci Educ. 2013;6(4):211–5. Zhao et al. BMC Medical Education (2020) 20:127 Page 10 of 10 26. Kockro RA, Amaxopoulou C, Killeen T, Wagner W, Reisch R, Schwandt E, Gutenberg A, Giese A, Stofft E, Stadie AT. Stereoscopic neuroanatomy lectures using a three-dimensional virtual reality environment. Ann Anat. 2015;201:91–8. 27. Moro C, Stromberga Z, Raikos A, Stirling A. The effectiveness of virtual and augmented reality in health sciences and medical anatomy. Anat Sci Educ. 2017;10(6):549–59. 28. Nicholson DT, Chalk C, Funnell WRJ, Daniel SJ. Can virtual reality improve anatomy education? A randomised controlled study of a computer- generated three-dimensional anatomical ear model. Med Educ. 2006;40(11): 1081–7. 29. Seixas-Mikelus SA, Adal A, Kesavadas T, Baheti A, Srimathveeravalli G, Hussain A, Chandrasekhar R, Wilding GE, Guru KA. Can image-based virtual reality help teach anatomy? J Endourol. 2010;24(4):629–34. 30. Solyar A, Cuellar H, Sadoughi B, Olson TR, Fried MP. Endoscopic sinus surgery simulator as a teaching tool for anatomy education. Am J Surg. 2008;196(1):120–4. 31. Stepan K, Zeiger J, Hanchuk S, Del Signore A, Shrivastava R, Govindaraj S, Iloreta A. Immersive virtual reality as a teaching tool for neuroanatomy. Int Forum Allergy Rhinol. 2017;7(10):1006–13. 32. Tan S, Hu A, Wilson T, Ladak H, Haase P, Fung K. Role of a computer- generated three-dimensional laryngeal model in anatomy teaching for advanced learners. J Laryngol Otol. 2012;126(4):395–401. 33. Drapkin ZA, Lindgren KA, Lopez MJ, Stabio ME. Development and assessment of a new 3D neuroanatomy teaching tool for MRI training. Anat Sci Educ. 2015;8(6):502–9. 34. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50:1088–101. 35. Harris RJ, Deeks JJ, Altman DG, Bradburn MJ, Harbord RM, Sterne JA. Metan: fixed-and random-effects meta-analysis. Stata J. 2008;8(1):3–28. 36. Shiozawa T, Butz B, Herlan S, Kramer A, Hirt B. Interactive anatomical and surgical live stream lectures improve students' academic performance in applied clinical anatomy. Anat Sci Educ. 2017;10(1):46–52. 37. Liu D, Bhagat KK, Gao Y, Chang T-W, Huang R. The potentials and trends of virtual reality in education. In: Virtual, augmented, and mixed realities in education. Singapore: Springer; 2017. p. 105–30. 38. Hattie J. The applicability of visible learning to higher education. Scholarsh Teach Learn Psychol. 2015;1(1):79. 39. Biasutto SN, Caussa LI, del Río LEC. Teaching anatomy: cadavers vs. computers? Ann Anat. 2006;188(2):187–90. 40. Kim K-J, Frick TW. Changes in student motivation during online learning. J Educ Comput Res. 2011;44(1):1–23. 41. Rebenitsch L, Owen C. Review on cybersickness in applications and visual displays. Virtual Reality. 2016;20(2):101–25. 42. Bleakley A. Gender matters in medical education. In: Patient-Centred Medicine in Transition: Springer, Cham; 2014. p. 111–26. Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Journal

BMC Medical EducationSpringer Journals

Published: Apr 25, 2020

There are no references for this article.