This systematic review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [25]. A protocol was prospectively registered with the PROSPERO database (CRD42024539802) prior to study initiation.

Eligible studies were original research articles investigating gamified interventions for laparoscopic skills training in simulation-based or clinical settings. Gamified interventions were defined as the use of game-like elements to motivate and engage learners in laparoscopic education. Included studies addressed either basic laparoscopic tasks or procedural skills related to laparoscopic surgery, particularly within gastrointestinal, hepatobiliary, or general surgical contexts. Participants were required to have a medical background, including medical students, residents, or practicing physicians. Outcomes were required to have indicators evaluating the laparoscopic skills, including objective metrics like completion time and subjective assessments like Global Operative Assessment of Laparoscopic Skills and FLS.

The exclusion criteria were as follows: (1) non-English publications; (2) full text not available; (3) studies involving veterinary or nonhuman participants; and (4) review articles, conference abstracts, posters, technical notes, and basic science research.

We searched 6 databases (PubMed, Embase, Cochrane Library, Scopus, Web of Science, and ClinicalTrials.gov), using the free text words and subject headings to conduct the search strategy. The search time limit was January 1, 2000, to July 1,2026. The full reporting of all literature search components, including database platforms, search syntax, citation tracking, deduplication, and search date, was conducted in accordance with the 16-item PRISMA-S (Preferred Reporting Items for Systematic Reviews and Meta-Analyses literature search extension) checklist [26]. All databases were searched individually rather than via a unified integrated retrieval platform, and all relevant literature was exclusively retrieved from electronic bibliographic databases. Hand searching of reference lists of all included studies was performed to identify additional cited and citing relevant articles for potential inclusion. No attempts were made to contact study authors, clinical experts, device manufacturers, or other personnel to request supplementary unpublished studies, raw data, or additional unpublished results. No published validated search filters (either original or modified versions) were adopted in this review. Besides, no complete or partial search strategies from previously published systematic reviews were adapted or reused for database retrieval. The detailed search strategy and PRISMA-S checklist are presented in Multimedia Appendix 1. Articles included after full-text screening require further reference retrieval. Zotero 6.0.36 (Corporation for Digital Scholarship, Roy Rosenzweig Center for History and New Media at George Mason University) is used to manage all included studies.

Three authors (XM, ZW, and MS) initially screened all records, excluded clearly irrelevant titles and abstracts, and retrieved the full texts of potentially eligible articles. Two authors (XM and WL) independently reviewed the remaining articles, assessed reference lists, and performed final screening of titles, abstracts, and full texts. Disagreements were resolved through discussion, with a third author (YL) arbitrating when consensus could not be reached.

Data extraction was performed independently by 1 reviewer (XM) using Microsoft Excel. Extracted information included study characteristics (author, year, and design), participant demographics, intervention details, types of gamification approaches, laparoscopic skills, and outcomes. Measurement outcomes included video game rankings and laparoscopic skill performance. For the qualitative synthesis, studies were grouped according to the following four prespecified dimensions: (1) participant population (medical students, surgical residents, or practicing surgeons), (2) intervention type (video game–based training vs gamified instructional contexts), (3) targeted laparoscopic skills (basic tasks such as peg transfer vs advanced procedural skills such as suturing or cholecystectomy), and (4) study design (randomized controlled trials [RCTs], nonrandomized interventional studies, and observational studies). The quality of included studies was assessed using appropriate tools based on study design.

All meta-analyses were performed using R software (version 4.6.1; R Core Team). For all continuous outcomes, the standardized mean difference (SMD) was calculated as the summary effect measure, with Hedges g correction applied to account for small-sample bias. A random-effects model with Hartung-Knapp-Sidik-Jonkman [27] adjustment was applied. Because the included studies assessed laparoscopic performance using different instruments and scales (eg, task completion time in seconds and peg transfer scores), all outcome measures were assumed to assess a common underlying construct of basic or advanced laparoscopic proficiency, thereby justifying the use of SMD to standardize results across heterogeneous metrics. The Nagashima correction [28] was used when calculating prediction intervals

Heterogeneity among studies was quantified using the between-study variance (τ²) and described by the I² statistic, which expresses the inconsistency of results—the proportion of total variability due to true between-study differences rather than sampling error—as a relative measure, not an absolute measure of the magnitude of heterogeneity. Cochran Q test was used to test the null hypothesis of homogeneity [29].

Sensitivity analyses were performed to assess the robustness of the pooled estimates. Given the small number of studies within each subgroup (k=2‐4), formal sensitivity analyses were limited. Where feasible (subgroups with k≥3), a leave-one-out analysis was conducted to examine the influence of individual studies on the pooled effect. For the subgroup comprising only 2 studies (k=2), sensitivity analysis could not be meaningfully performed and results are interpreted with caution.

In this review, to reduce publication bias, only RCTs with SD and q value were included in meta-analysis, and other studies only adopted qualitative research methods. For studies that only reported median and IQR without SD, we estimated SD using the formula previously described by Wan et al [30]. Studies with severely skewed data and limited sample size were excluded from quantitative synthesis and only narratively summarized.

Included studies were categorized into interventional studies and observational studies. The ROBINS-I (Risk of Bias In Non-randomized Studies-of Interventions) quality assessment tool [31] was used to evaluate the quality of nonrandomized studies and pre-post studies, with ratings classified as low, moderate, high, and no information. RCTs were evaluated using the RoB 2 (Cochrane Risk of Bias Tool 2) quality assessment scale [32], with assessment categorized as low, some concerns, and high. For observational cohort and cross-sectional studies, methodological quality was assessed using the National Institutes of Health (NIH) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies [33]. The overall quality of each study was graded as good (low risk of bias), fair (moderate risk), or poor (high risk) according to the NIH official criteria. Given the qualitative nature of the synthesis, all studies meeting the eligibility criteria were included in the narrative summary. For the main conclusions, greater interpretive weight was placed on studies with stronger methodological quality (rated as “Good” on NIH tools or “Low risk of bias” on RoB 2 and ROBINS-I), larger sample sizes, and direct relevance to the review question as defined by the inclusion criteria. Studies rated as “Poor” or “High risk of bias” were not excluded but were interpreted with caution, and their influence on the overall conclusions was explicitly noted in the Discussion section.

The certainty of evidence for each critical and important outcome was assessed using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) framework [34].

The initial database search yielded 12,012 records. After title and abstract screening, 93 articles were selected for full-text review. Moreover, 43 records were identified from citation searching, and after removing duplicates, 4 records were selected for full-text review. Following full-text assessment, 39 articles [16,18,20-22,24,35-67] met the inclusion criteria and were included in the final analysis (Figure 1). The included studies consisted of 29 interventional [16,18,20,22,35-37,39,47-67] and 10 observational [21,38-46] studies. The publication years ranged from 2005 to 2026, with the highest number published in 2017 (n=5, 12.82%). Detailed study type of each study is presented in Table 1.

A total of 2102 participants were included across all studies. All participants had a medical background, including 814 medical students and 905 physicians (eg, surgeons, residents, and interns; Table 1). Handedness was reported in 723 participants—628 were right-handed, 93 left-handed, and 2 ambidextrous [35]. Participants represented various specialties, including general surgery (n=179), obstetrics and gynecology (n=90), urology (n=33), and vascular surgery (n=10). Moreover, 89 participants reported experience playing musical instruments [16,18,36,37].

Of the 10 observational studies, 7 were rated fair quality of risk, and 3 good quality. Detailed assessment results are presented in Table 2. Among the 29 interventional studies, 21 were rated as some concerns or moderate risk of bias, 3 were rated as high risk of bias, and 5 were rated as low risk of bias. Detailed assessment results are presented in Figure 2.

Gamified strategies in laparoscopic education were categorized into game-based learning and gamified context-based approaches, with 34 articles [18,22,35-45,47-65] and 5 articles [20,21,46,65,66] included, respectively.

Among game-based learning approaches, “Underground” was the most frequently selected game, featured in 9 out of 34 (26.47%) studies [16,39,41,43,44,47-50], necessitating specific Nintendo equipment. The next most popular was “Super Monkey Ball 2,”(developed by SEGA) selected in 6 [44,45,53,55-57] of 34 (17.65%) studies, compatible with Nintendo [44,45,51,52] or intelligent devices [53]. Specific game titles and platforms are listed in Tables 3 and 4 and Textbox 1. The 4 gaming platforms provided are Nintendo, intelligent device, PlayStation, and Xbox.

In the gamified contexts, 4 studies [20,21,46,65] implemented competitive elements (80%), and one study [66] implemented collaborative elements (20%).

Among the included studies, 34 articles applied video games as either training tools or assessment tools for laparoscopic skill development. Out of 34 studies, 26 [16,18,23,35-37,47-64,67] (76.47%) used video games as substitutes for traditional training, with interventions categorized into the following 3 modes: during training, pretraining, and warm-up. The most commonly used mode was during-training (17/26 [16,23,35,37,48,49,51,54-60,64,67], 65.38%). Details are summarized in Table 5. During training refers to the use of self-directed or structured gameplay between 2 tests, pretraining [18,36,50,62] involves completing designated game tasks prior to performance testing, without time constraints, and warm-up [47,52,53,67] refers to short gaming sessions (typically 10‐18 minutes) immediately before testing, without specific game-task requirements.

Among the 34 studies, 9 [22,36-40,42,44,45] (30%) used video game performance or ranking scores as surrogate assessment tools, exploring correlations with laparoscopic skill levels. Detailed outcomes are summarized in Table 6. Kulkarni et al [42] reported strong correlations between scores on XBOX or Nintendo games and performance in laparoscopic tasks, such as Sugar Cube Stacking, Cutting, and total composite scores. IJgosse et al [68] found that participants with greater laparoscopic experience achieved higher scores in the Underground game, although no statistically significant difference was observed between expert and intermediate groups.

Creating a competitive context mainly involves rewarding top-ranked participants and regularly updating the leaderboard. Websky et al [20] developed a peer-group benchmarks group to enhance collaborative settings, which resulted in better performance in efficiency and training motivation relative to the control group. Specific interventions and results are listed in Tables 7 and 8.

Of all 39 studies included in this review, 9 studies [20,35,48,49,51,56,58,63,65] matched the criteria for inclusion in the meta-analysis. Of the total, 3 [51,58,63] studies compared game-based learning with a no-intervention (blank) control group, 3 [48,49,56] compared it with conventional laparoscopic training, and 2 [20,65] examined gamified vs nongamified learning. Only 1 study [35] compared different video game platforms.

Subgroup meta-analyses were conducted to compare the effectiveness of game-based learning against different control conditions on laparoscopic skill performance. The SMD was used as the effect measure, and data were pooled with a random-effects model. Furthermore, 4 studies [20,48,58,65] required the use of the formula by Wan et al [30] to estimate the SD from IQR. The detailed data were shown in Multimedia Appendix 2.

As illustrated in Figure 3, 3 studies [51,58,63] (95 participants) evaluated game-based learning vs no training. The pooled Hedges g was –0.04 (95% CI –7.14 to 7.06), with a 95% prediction interval ranging from −13.95 to 13.61. The I² statistic indicated noteworthy inconsistency (I²=96.7%), with a between-study variance of τ²=7.89 (Q=59.88; P<.001).

As illustrated in Figure 4, 3 studies [48,49,56] (108 participants) compared game-based learning with conventional laparoscopic training. The pooled Hedges g was 0.49 (95% CI –1.39 to 2.36), with a 95% prediction interval ranging from –3.60 to 4.51. The I² statistic indicated noteworthy inconsistency (I²=79.4%), with a between-study variance of τ²=0.45 (Q=9.69; P=.008).

As illustrated in Figure 5, 2 studies [20,65] (86 participants) compared gamification training with nongamification training. The current outcome was task completion time, where smaller values correspond to better performance. Raw time data were negated before calculating Hedges g to align effect directions with other end points. A positive effect size thus indicates better performance among participants receiving training. Under the random-effects model, the pooled Hedges g was 0.87 (95% CI –5.08 to 6.81). The I² statistic indicated inconsistency (I²=61.6%), with a between-study variance of τ²=0.29 (Q=2.61; P=.11).

A study by Ju et al [35] showed a further comparison between different video games. Both groups achieved statistically significant improvements in the bead transfer task (Wii: mean score improvement 62, SD 14; P<.001; PS2: mean score improvement 40, SD 14; P=.01), yet the between-group difference in improvement magnitude did not reach statistical significance. For the intracorporeal suturing task, neither group showed a statistically significant change in performance after the intervention. Double counting in the meta-analyses did not occur.

To determine whether the pooled estimate was driven by a single study, we performed leave-one-out sensitivity analyses (Figure 6). The SMD reversed direction after omitting one of the following studies: (1) study by Andrews et al [63]: 1.52 (95% CI –9.96 to 13.00), (2) study by Byrns et al [51]: –0.38 (95% CI –36.04 to 35.28), (3) study by Giannotti et al [58]: –1.28 (95% CI –25.46 to 22.90), and every 95% CI still spanned the null, indicating a nonrobust pooled estimate.

The pooled estimate for subgroup 2 was highly sensitive to the included studies (Figure 7). The overall nonsignificant result for group 1 was attributable mainly to the study by Lin et al [56]: after its exclusion, the pooled SMD shifted to 0.92 (95% CI 0.17 to 1.67; P=.04) with I²=0%, whereas excluding the study by Mirmoghtadaie et al [49] or Graafland et al [48] did not change the nonsignificant conclusion (SMD 0.29, 95% CI –8.29 to 8.88; and SMD 0.27, 95% CI –7.54 to 8.07, respectively).

GRADE certainty assessment was performed for three comparisons of laparoscopic training interventions. Evidence for gamification versus nongamification was moderate certainty. Evidence for game-based learning vs no training was low certainty. Evidence for game-based learning versus conventional laparoscopic training was very low certainty. Detailed GRADE assessments are provided in Multimedia Appendix 3.

This systematic review evaluated the feasibility and effectiveness of gamified interventions, including game-based learning and gamified context-based approaches, in laparoscopic skills training. Video games were primarily used as supplementary training tools or assessment platforms, whereas gamified contexts incorporated motivational elements, such as competition and collaboration, to enhance learner engagement and participation [16,39,45,49,69,70]. Overall, qualitative evidence suggested that gamified interventions may provide educational benefits in laparoscopic training, particularly in improving psychomotor performance, cognitive outcomes, and learner motivation [16,20,21,49,58,65]. However, the magnitude of these effects appeared to vary depending on the training objectives, intervention characteristics, and assessment domains.

Previous studies have reported that video game-based training may improve basic psychomotor abilities, including task completion time, movement efficiency, and error reduction among surgical trainees [35,37,59,62]. Serious games, such as Touch Surgery (developed by Kinosis Ltd) and Dr. Game, Surgeon Trouble (developed by Amsterdam UMC), have also shown potential value in supporting theoretical knowledge acquisition, procedural understanding, and intraoperative decision-making [18,36,48]. Nevertheless, the transferability of these benefits to advanced laparoscopic performance remains uncertain. Current evidence suggests that video games may facilitate basic visuospatial and psychomotor skill development but have limited ability to reproduce the procedural complexity, instrument manipulation, and cognitive demands required for advanced operative tasks [22,50,64].

Compared with conventional laparoscopic training, game-based learning demonstrated comparable performance outcomes in selected basic skill tasks but has not consistently demonstrated superiority for procedure-specific training [56]. This may be related to differences in intervention design, insufficient surgical specificity, variable training exposure, and heterogeneous outcome assessment methods across studies [16,42]. Therefore, video games are more appropriately considered complementary training resources rather than replacements for established simulation-based curricula.

Gamification strategies may provide additional educational value by modifying learner motivation and engagement through structured feedback, competition, or collaboration [23,46,65,66]. Competitive gamification has been associated with improvements in dexterity, performance efficiency, and practice motivation. For example, Hashimoto et al [65] reported reduced movement variability and shorter instrument paths following competitive training, while El-Beheiry et al [21] observed improved task completion and increased voluntary practice in competitive environments. In contrast, collaborative gamification may facilitate teamwork, communication, and sustained learner participation, suggesting that different gamification elements may influence distinct aspects of surgical education [20,54].

Despite these encouraging findings, the current evidence remains limited by considerable variation in intervention protocols, target skills, participant characteristics, and outcome measurements [16,45]. Although gamified approaches are accessible and potentially cost-effective compared with traditional simulation platforms, many commercial video games lack sufficient surgical specificity and may not adequately reproduce the technical and cognitive demands of laparoscopic procedures [16,45]. Further well-designed studies with standardized training protocols, validated assessment tools, and clinically relevant outcomes are required to determine the optimal role of gamification in surgical education.

Pooled effects were small to moderate and none reached statistical significance. Inconsistency among studies was substantial (I²=61.6%‐96.7%); because I² measures inconsistency rather than the magnitude of heterogeneity, between-study dispersion is better reflected by the between-study variance (τ²) and, above all, by the prediction intervals, which were far wider than the CIs and crossed the null in every comparison. The true effect may therefore range from harmful to beneficial across settings, and a uniform benefit of gamification cannot be assumed. It is important to distinguish the average effect from the distribution of true effects. Although the three pooled estimates were small (near-zero for game-based learning versus no training, or moderate, none was statistically significant, and the 95% prediction intervals were far wider than the CIs and crossed the null in every comparison. This indicates that, beyond the average effect, the true effect in a given setting is likely to vary widely—from negligible or harmful to substantial benefit, so the benefit of gamification cannot be assumed to be uniform across populations, training protocols, or outcome domains. Notably, the very wide prediction intervals may reflect not only genuine between-study dispersion but also the small number of studies in each subgroup, since with so few studies the between-study variance (τ²) is estimated imprecisely and the intervals become correspondingly uncertain. Reporting only I² would have obscured this, because a high I² is a proportion and indicates neither the magnitude nor the clinical relevance of the dispersion.

The I² statistic indicated substantial inconsistency across all 3 subgroups. Differences in outcome types may represent the most important potential contributor. Such heterogeneity was expected given the diversity of training interventions, assessment methods, and participant characteristics included in this review. The outcomes included distinct dimensions of laparoscopic performance, including time efficiency, accuracy, error rates, and technical skills. Laparoscopic performance is a multidimensional construct encompassing psychomotor coordination, efficiency, accuracy, and cognitive decision-making. Therefore, combining these outcomes into a single estimate may obscure domain-specific effects and contribute to substantial heterogeneity.

The pooled studies also differed in what was measured. This included outcomes from different performance domains (time-based efficiency, error counts, kinematic measures such as path length and movements, technical scores, problem-solving and situational awareness, and nontechnical measures such as state anxiety [16,20,37,48,59,65]), tasks at different levels (basic transfer and suturing [22,56,58] vs cholecystectomy and other procedural tasks [20,49,65]), and inconsistent outcome direction, since “lower is better” measures (time, errors, and path length) required reversal before pooling. Pooling such heterogeneous outcome constructs into a single standardized mean difference therefore limits the interpretability of the pooled estimate and likely contributes to the high between-study dispersion observed. In addition, training duration and training modalities were also important potential sources of variability. Longer training durations tended to be associated with greater performance improvements, suggesting a possible dose-response relationship between training exposure and skill acquisition. In contrast, video games used solely as warm-up activities were unlikely to produce significant improvements and might even have limited benefits for physical skill development.

The robustness of the pooled estimates was examined using leave-one-out sensitivity analysis. This revealed that the pooled estimate for the game-based learning vs no-training subgroup was not robust; the SMD reversed direction and every 95% CI still spanned the null after omitting any single study (Andrews and Hewett [63], Byrns et al [51], or Giannotti et al [58]). In the game-based learning vs conventional training subgroup, the overall nonsignificant result was driven largely by a single study; its exclusion shifted the pooled estimate to a statistically significant positive value and eliminated inconsistency, whereas excluding the other 2 studies did not change the nonsignificant conclusion. Given the small number of studies in each subgroup (k=2‐4) and the substantial between-study heterogeneity observed, these results should be interpreted with caution and are not sufficient to establish robust effectiveness.

According to the GRADE framework, the certainty of evidence was rated as moderate to very low across the evaluated outcomes. The main reasons for downgrading were substantial inconsistency among studies, methodological limitations, limited sample sizes, and imprecision of effect estimates. Although gamified interventions demonstrated potential benefits in laparoscopic training, the current evidence remains insufficient to support definitive conclusions regarding their superiority over conventional training approaches.

To translate gamification and game-based learning into meaningful educational practice, careful instructional design is essential. Instead of using commercial video games in isolation, future efforts should prioritize the development or customization of task-aligned, evidence-informed gamified modules that complement existing laparoscopic curricula [44,59]. Embedding competitive or collaborative game elements within simulation-based training can promote learner autonomy, increase deliberate practice, and enhance motivation [21,65]. However, the design must avoid overemphasizing entertainment at the expense of clinical relevance or technical fidelity [50]. Alignment with validated assessment tools, such as OSATS (Objective Structured Assessment of Technical Skills) [71] or FLS metrics, can enhance educational value while ensuring objective evaluation [35,48,55,65]. Instructors should also consider learner variability when integrating gamified approaches—tailoring difficulty levels, feedback frequency, and reward mechanisms to the learner’s stage of training may optimize both engagement and skill acquisition. As interest grows in intelligent surgical training ecosystems, gamification could play a complementary role when combined with adaptive learning algorithms and real-time performance analytics [38,55,58].

Despite the promising findings, this review has several limitations. First, substantial heterogeneity existed in gamification formats, training duration, participant characteristics, assessment tools, and outcome measures, which limited the comparability among studies and complicated quantitative synthesis. Second, several potentially eligible studies could not be included in the meta-analysis because they reported medians and IQRs rather than means and SDs. Although statistical conversion methods are available, such transformations may introduce estimation uncertainty and potential bias. Consequently, subgroup analyses were restricted by the limited number of available studies. Third, the small number of studies in each subgroup (k=2‐4) was below the minimum threshold (typically ≥10) recommended for testing funnel plot asymmetry, so small-study effects could not be reliably assessed. Notably, such tests detect small-study effects—a pattern in which smaller studies report systematically larger effects—which may arise from publication bias but also from other causes such as poorer methodological quality in smaller studies, true heterogeneity, or chance. Because publication bias and other contributors to asymmetry cannot be excluded, pooled estimates should be interpreted cautiously. Fourth, although both interventional and observational studies were included, several studies had methodological limitations, particularly regarding randomization procedures, allocation concealment, and blinding, which may have affected the certainty of evidence. Fifth, most studies evaluated short-term skill acquisition rather than long-term retention or clinical transfer. Although gamification may enhance motivation and engagement during initial training, whether these benefits can be sustained over prolonged training periods remains unclear.

Future studies should adopt standardized outcome measures, clearly define training protocols, and incorporate longitudinal follow-up to evaluate long-term educational effects. Researchers should also report complete statistical parameters, including means and SDs, to facilitate future evidence synthesis. Furthermore, future investigations should explore whether adaptive gamification strategies, such as dynamic difficulty adjustment, personalized feedback, and progressive challenges, can maintain learner engagement and support sustained skill development throughout extended training programs.

In summary, this review confirms that gamified interventions and game-based learning are feasible adjuncts for laparoscopic skills education and may support basic psychomotor skill acquisition and learner motivation; however, the pooled meta-analytic estimates did not reach statistical significance and were not robust to leave-one-out analysis, so a definitive advantage over conventional training cannot be established. Game-based learning and gamified contexts each have distinct strengths; the former serves well as supplementary training or warm-up tools, while the latter offers a framework for curriculum-wide engagement design. A key finding is that generic video games do not reliably improve advanced procedural competence, underscoring the need for task-aligned, curriculum-integrated gamified modules rather than isolated commercial games. This review differs from prior work by comparing both major gamification approaches within a single analytical framework and by mapping their underlying theoretical foundations to the evidence base. Its contribution lies in providing educators with an evidence-informed rationale for selecting specific gamified strategies based on training objectives—competitive elements for technical proficiency, collaborative elements for team-based skills—and in highlighting implementation realities, including the motivation paradox whereby positive attitudes do not guarantee sustained voluntary practice. Looking ahead, the integration of gamified instructional design with adaptive learning algorithms and standardized assessment tools may represent the next step toward personalized, proficiency-based laparoscopic training. Until then, gamification should be viewed as a valuable complement to, rather than a replacement for established simulation methods.