Seven English-language databases (PubMed, Web of Science, Embase, CINAHL, Cochrane Library, JBI, and Scopus) and 3 Chinese-language databases (China National Knowledge Infrastructure, SinoMed, and Wanfang) were systematically searched from inception to June 25, 2025. We conducted a repeated search on February 8, 2026, and found no more eligible literature. Google Scholar was retrieved as a gray literature to minimize publication bias. We retrieved the CENTRAL (Cochrane Central Register of Controlled Trials) database through the Cochrane Library, including resources from trial registries such as ClinicalTrials.gov and ChiCTR. No published search filters were used. The search terms “dysphagia,” “game,” and “swallowing exercise” were combined in each database using both free-text words and Medical Subject Heading terms, if available. We originally developed a scientific and rigorous search strategy based on our research question, under the guidance of a professor specializing in medical information retrieval. No previously published search strategies were adapted or reused in this review. The snowball method was used to retrieve references from eligible studies to identify potential literature and avoid missing studies. No language or other restrictions were applied to the search. The search strategy adhered to the PRISMA-S (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Search Extension) checklist to ensure transparency [22] as seen inChecklist 2 . Multimedia Appendix 2 provides the specific search strategies for each database.

The PICOS (population, intervention, comparison, outcomes, and study design) framework established the inclusion and exclusion criteria for the systematic review of randomized controlled trials (RCTs). The inclusion criteria were as follows: (1) population: patients with dysphagia aged 18 years or older, irrespective of the origin of the swallowing impairment; (2) intervention of interest: gamified swallowing exercises, defined as swallowing exercises carried out in the form of games, including video games, virtual reality games, etc; (3) comparison: individuals in the comparison group were provided with conventional swallowing exercises or nongamified exercises; (4) outcomes: swallowing function (including swallowing performance, dysphagia severity, and dysphagia screening), adherence, nutritional status, or quality of life; and (5) study design: RCTs (including crossover, cluster, and pilot studies). The exclusion criteria were as follows: (1) editorial articles, conference abstracts, study protocols, comments, and letters and (2) studies with essential data that remained unavailable if data conversion was not possible and the corresponding authors did not respond to our requests.

All retrieved studies were imported into EndNote 21 software for automatic and manual deduplication. After the removal of duplicates, 2 independent reviewers (JL and MY) screened the titles and abstracts before proceeding to assess the full texts against the established eligibility criteria. Disagreements during the study selection process were addressed through discussions within the research team.

Two reviewers (JL and MY) performed data extraction, with accuracy verification by a third researcher (MO). Information on the eligible studies was extracted, including study characteristics (authors, publication year, and country), participant characteristics (age and sample size), interventions (components, dose, timing, frequency, and length), controls, and outcomes. For studies with more than 1 intervention group where gamification was used in only 1 of the intervention groups, a comparison was conducted between the gamified group and the other 2 groups. If the authors did not report the mean and SD, we contacted the corresponding authors via email to request raw datasets. If unavailable, we converted them to the mean and SD based on the median and IQR [23,24]. Otherwise, the data were classified as missing data.

The revised Cochrane Risk of Bias Tool, version 2.0, was used to assess the methodological quality of the RCTs [25]. Five domains were assessed: randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. The evaluation results were divided into low-risk, some concerns, and high-risk groups. Additionally, the GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) framework was used to grade the certainty of evidence for all outcomes [26]. The certainty was divided into high, moderate, low, and very low based on the risk of bias, imprecision, inconsistency, indirectness, and publication bias [27]. The overall quality of evidence was evaluated using GRADEpro GDT software. Two reviewers (JL and MY) independently assessed the quality of the included studies. Disagreements were addressed through discussion and/or consultation with a third reviewer (MO) to achieve consensus.

Data were analyzed using R software (version 4.4.2). According to the Cochrane Handbook for Systematic Reviews of Interventions [28], for the 3-arm RCTs, we combined the 2 nongamified groups into 1 group, and pooled calculations were performed using the following formula:

A meta-analysis was conducted on outcomes measured in 2 or more studies. For continuous variables measured using the same scale, we estimated pooled effect sizes using the pooled mean difference (MD); otherwise, the pooled standardized mean difference (SMD) was computed as the pooled effect size. For studies that used multiple instruments to assess the same outcome, we established a priori “scale prioritization” rules based on the psychometric properties of each instrument (such as Gugging Swallowing Screen > Standardized Swallowing Assessment > Water Swallowing Test) [29], and extracted data accordingly for effect size synthesis. We employed the Hartung-Knapp-Sidik-Jonkman method for random-effects model analysis to calculate the pooled estimate and its corresponding 95% CI, addressing the issue of limited study inclusion and high heterogeneity between studies [30]. The heterogeneity of outcomes was tested using Cochran Q (χ² test) and I² statistic [31]. In detail, P<.05 was considered to indicate heterogeneity.

To explore potential sources of heterogeneity, subgroup analyses were conducted based on different assessment tools, kinds of games, different additional equipment, and disease duration. The Instrument for the Credibility of Effect Modification Analyses (ICEMAN) was applied to evaluate the credibility of subgroup effects [32] when the interaction P<.1. The credibility of subgroup effects was categorized as “very low,” “low,” “moderate,” or “high” based on 8 core evaluation items. An interaction P≤.005 was used as the threshold to minimize the likelihood of chance findings. Results for subgroups containing only a few studies were interpreted with caution.

Upon detecting substantial heterogeneity, 2 types of sensitivity analyses were conducted to evaluate the robustness of the findings as follows: (1) a leave-one-out method, where each study was sequentially excluded to assess changes in heterogeneity and result stability and (2) an alternative measurement instrument analysis, in which alternative measurement results were used for studies that assessed the same outcome using 2 different instruments. Funnel plots were used to assess small-study effects [33]. In addition, when the number of included studies was small, we calculated the 95% prediction interval using the Nagashima correction to illustrate the practical implications of heterogeneity and present the expected range of true effects across studies [34].

The initial search identified 2400 publications, including 2366 records from 9 databases (the JBI database yielded no publications) and 34 from the gray literature database. After removing 387 duplicate articles, the titles and abstracts were scanned, and 1976 articles were excluded owing to irrelevant topics. After reviewing the full text, 31 articles were excluded because of their irrelevant study design, participants, interventions, outcomes, or unavailability of data. Ultimately, 6 studies [17-19,35-37] were included in this review. A flow diagram is presented in Figure 1, and the details of excluded studies are shown in Multimedia Appendix 3.

Table 1 summarizes the characteristics of the 6 studies included in this review. All studies were published between 2019 and 2025. Four studies [17,19,36,37] were conducted in China, 1 [18] in Korea, and 1 [35] in Turkey. Five studies [18,19,35-37] were 2-arm RCTs, and 1 [17] was a 3-arm RCT. In total, the studies involved 330 participants with poststroke dysphagia. The disease duration ranged from 8 days (approximately 0.27 mo) to 5.18 months.

Table 1 summarizes the intervention characteristics of the 6 studies included in this review. The content and form of the gamified swallowing training varied. Three studies [17,18,37] included only 1 game module, and 3 studies [19,35,36] designed various games targeting specific swallowing movements. In terms of exercise content, 6 studies included tongue exercises (n=2), chin tuck against resistance exercises (n=2), lip movements (n=2), effortful swallowing (n=2), the Mendelsohn maneuver (n=2), and jaw exercises (n=1) [17-19,35-37].

All 6 studies required additional equipment (eg, surface electromyography, resistance bars, and computers equipped with face recognition–driven technology or artificial intelligence technology for intervention delivery) to conduct the interventions.

The gamified swallowing exercises in all studies were conducted for a period of 1 to 4 weeks [17-19,35-37], with only 1 study conducting a 4-week follow-up [36].

Table 1 summarizes the characteristics of the control groups of the 6 included studies. Swallowing exercise protocols were identical between the control and intervention groups, except that gaming components were excluded in the controls. The usual care or conventional therapy typically encompassed oral motor exercises, sensory stimulation techniques, compensatory strategies, and supportive care (eg, vital sign monitoring, routine airway humidification, and nutritional management) [17-19,35-37].

Swallowing function was the primary outcome in 6 studies [17-19,35-37]. Swallowing function was categorized into swallowing performance, dysphagia severity, and dysphagia screening based on scale items. Swallowing performance was primarily assessed through the Functional Oral Intake Scale (n=6) [17-19,35-37], which measures the patients’ ability to eat orally. The Dysphagia Outcome and Severity Scale (n=1) [35] and videofluoroscopic Dysphagia Scale (n=1) [18] were used to measure dysphagia severity. The measurement tools of dysphagia screening included the Gugging Swallowing Screen (GUSS; n=2) [19,36], the Standardized Swallowing Assessment (SSA; n=3) [19,36,37], and the Water Swallowing Test (WST; n=2) [17,37]. Physiological measures included submental muscle amplitude, tongue pressure, and surface electromyographic value [17,37].

Figure 2 summarizes the risk of bias for the 6 included studies. Two reviewers independently assessed the methodological quality, achieving 95% agreement (κ=0.91). Two studies had a low risk of bias, 3 raised some concerns, and 1 was high risk. Three studies were considered to have some concerns because they did not report allocation concealment. One study had a high risk of bias because the outcomes were assessed without blinding, and the results, particularly those based on subjective measures, were therefore susceptible to influence by the outcome assessors. For a detailed assessment, please refer to Multimedia Appendix 4.

Regarding adherence, 2 studies involving 110 participants demonstrated that, compared with the control group, gamified swallowing exercises were associated with increased adherence among adults with dysphagia (MD=2.4; 95% CI 1.8‐2.9; I²=0; Figure 6).

Quality of life was reported in 3 studies involving 143 participants. The results of the pooled analysis indicated that gamified swallowing training did not significantly improve the quality of life in adults with dysphagia compared with the control group (SMD=–2.3; 95% CI –9.6 to 5; I²=97.5%; Figure 7).

Given the preliminary findings of the meta-analysis, dysphagia screening was selected as the primary outcome for subgroup analysis. Subgroup analysis was conducted according to different measurement tools, kinds of games, additional equipment, and disease duration to explore the sources of heterogeneity. A significant between-subgroup difference was observed for the measurement tools (test for subgroup differences: χ²=86.5; 95% CI –5.4 to 1.8; P<.001). In contrast, no significant between-subgroup differences were found for kinds of games or additional equipment (χ²=0.8; 95% CI –5.4 to 1.8; P=.40) or disease duration (χ²=1; 95% CI –5.4 to 1.8; P=.30). The subgroup analysis showed that for patients with dysphagia after stroke, when using the SSA (k=2) to assess dysphagia screening, using a single game or using sEMG as the exercise equipment in dysphagia screening, the heterogeneity between studies significantly decreased (I²=0%), suggesting that such studies were relatively consistent in terms of design or implementation. However, the CI for the combined effect included the null value, indicating that the intervention effect was not statistically significant in this subgroup. Other subgroups, such as multiple games, other additional equipment (such as facial recognition devices), and disease duration ≥1 month, still had high heterogeneity (>95%). The remaining 3 subgroups (WST, GUSS, and disease duration <1 mo) had only 1 study, resulting in overall very low credibility of the findings, which should be interpreted with caution. Since the P value for the interaction effect in the kinds of games and additional equipment subgroup was >.1, it did not meet the prerequisites for applying the ICEMAN tool; therefore, no further evaluation was conducted. Detailed results of the subgroup analysis are illustrated in Multimedia Appendix 5.

To assess the impact of measurement tools on the robustness of the results, we performed an alternative measurement tool analysis, replacing the dysphagia screening scales used in Zhang study [36] and Kang study [37] with another scale. Specifically, we substituted GUSS with SSA in the Zhang study [36] and SSA with WST in the Kang study [37]. The direction of the pooled effect estimates remained unchanged after these substitutions. The leave-one-out sensitivity analysis indicated that the result of the swallowing performance was robust. For dysphagia screening, after excluding the Zhang study [36], the effect size of the remaining results changed (from ineffective to effective), and the heterogeneity decreased to 0, suggesting that the robustness for dysphagia screening was limited. For quality of life, if any 1 study was excluded, the effect size of the remaining results remained unchanged; however, after excluding the Zhang study [36], the heterogeneity significantly decreased (from 97.5% to 21.8%). Multimedia Appendix 6 presents the details of the sensitivity analysis.

However, due to the limited number of included studies, the estimation of the 95% prediction interval may be unstable, potentially leading to an overestimation of the true effects [38]. Therefore, the results of the prediction intervals should be interpreted with caution.

Due to the limited number of studies included in each funnel plot (n<8), the observed symmetry (or asymmetry) lacked reliability. Consequently, no meaningful assessment of funnel plot symmetry could be performed. The relevant funnel plots are provided in Multimedia Appendix 7.

Among the included studies, only 1 [36] examined the effectiveness of gamified swallowing exercises on nutritional status, which precluded the ability to conduct a meta-analysis on nutritional status. The intervention group exhibited significantly higher nutritional scores compared with the control group (MD 1, SD 0.5, 95% CI 0.1-1.9; P=.03). However, the 4-week follow-up revealed no statistical significance in nutritional status between groups. These findings suggest that individuals with dysphagia who received gamified swallowing exercises achieved significantly better short-term nutritional status than those who received conventional swallowing exercises; however, further investigation is required to determine the long-term effects.

Among the included outcomes, swallowing performance demonstrated moderate certainty of evidence; dysphagia severity and adherence had low certainty of evidence; and dysphagia screening, nutritional status, and quality of life showed very low certainty of evidence.

The certainty was downgraded primarily due to a high risk of bias, wide CIs, limited sample size, and statistical heterogeneity. Full GRADE (Grading of Recommendations Assessment, Development, and Evaluation) assessment details are available in Table 2.

Gamified swallowing exercise is an emerging and promising approach in the field of swallowing rehabilitation. To the best of our knowledge, this study is the first systematic review and meta-analysis to rigorously and comprehensively evaluate the effectiveness of gamified swallowing training. The findings suggest that gamified swallowing exercises may help improve swallowing performance, reduce dysphagia severity, and increase adherence among individuals with dysphagia. However, due to the limited number of studies and considerable heterogeneity, the effects of gamified swallowing training on dysphagia screening, nutritional status, and quality of life remain uncertain.

The meta-analysis indicated that gamified swallowing exercises can effectively enhance swallowing performance, reduce dysphagia severity, and improve adherence. Low heterogeneity and narrower prediction intervals also support higher consistency of the effects across different settings [39]. This positive outcome may be attributed to the integration of gamification elements into traditional swallowing exercises without altering the core components of the rehabilitation program. Compared to conventional approaches, the inherent mechanisms of gamification can stimulate participants to maintain their motivation, engagement, and interest, facilitating sustained participation in the process [40]. Moreover, immersive game-based interventions address key challenges in rehabilitation, such as monotony, participant discomfort, and resistance, which can foster positive attitudes toward swallowing exercises and promote long-term adherence to beneficial health behaviors [41].

Gamified swallowing rehabilitation systems typically employ additional equipment, such as EMG [17,35,37] or facial recognition devices [19,36], to monitor participants’ swallowing activity, including movements of the tongue, lips, jaw, and throat muscles. They then present the electrical activity of the relevant muscles or allow the participant to control a game character through these movements, providing real-time visual and auditory feedback. This equipment-assisted feedback mechanism helps participants promptly correct their training posture and reduces the workload of nurses and therapists [36]. Notably, the above equipment is specialized and can only be used in hospitals or rehabilitation institutions. Future development may focus on integrating this method with widely accessible mobile technologies, such as smartphones, to enable remote dysphagia management in community or home settings. As an innovative and scalable intervention model, gamified swallowing rehabilitation holds considerable promise for boosting engagement, promoting long-term recovery, and advancing the application of technology-assisted rehabilitation in real-world contexts.

Regarding dysphagia screening, this meta-analysis did not demonstrate a significant effect of gamified swallowing training, and wide prediction intervals indicate that the effect varies differently. During the subgroup analysis of dysphagia screening, the observation of zero heterogeneity in some subgroups points to consistent intervention design; however, the lack of significant therapeutic effect indicates that the intervention’s mechanism itself may need refinement. For example, swallowing is a multistage collaborative process, and a single game typically focuses on muscle training in one stage, which may limit the overall effectiveness of swallowing training. The limited number of studies in each subgroup reduced the statistical power, and the observed differences may reflect random error rather than true subgroup effects. These findings are exploratory and should be interpreted with caution, pending further validation through high-quality studies. Substantial statistical heterogeneity was observed within some subgroups (ie, multiple games, disease duration ≥1 mo), which may be attributable to the inclusion of a pilot study [19] with a small sample size, resulting in high uncertainty around its effect estimates. Furthermore, one included study [19] was a pilot version of another included trial [19,36]; however, these 2 studies exhibited markedly different standard deviations, with no overlap in confidence intervals, which further contributed to the heterogeneity.

The sensitivity analysis showed that removing the study by Zhang et al [36] would affect the direction of the results for dysphagia screening, indicating that, in the context of a limited number of studies, a single large-sample study might disproportionately impact the combined effect, reflecting the statistical vulnerability of meta-analyses with small samples rather than indicating a true difference in intervention effects.

Furthermore, several studies [17,19,37] had methodological limitations, primarily the lack of allocation concealment and blinding of outcome measurement. The former may introduce selection bias, whereas the latter can cause judgment bias in outcome assessments relying on subjective evaluations, which may lead to over- or under-estimation of effects and undermine the robustness of the combined results. These methodological flaws should be carefully considered when interpreting the research findings.

Overall, despite exploring multiple dimensions such as measurement tools, game quantity, additional equipment, and disease duration, the source of heterogeneity remains unclear. The findings of this study should be interpreted with caution, and future research should focus on large-scale RCTs while improving methodological quality and reporting standards to validate the intervention effects and clarify the sources of heterogeneity.

This meta-analysis revealed a potential positive impact of gamified swallowing exercises on reducing dysphagia severity. A major limitation, however, is the scarcity of available trials (n=2) and the small overall sample size (n=70). These constraints likely led to low statistical power, suggesting that the observed positive result should be considered preliminary and requires validation through larger, more robust studies. Furthermore, poststroke dysphagia often results from neurological impairment [42]. Due to the inclusion of participants who have had a stroke for >3 months, a longer intervention period may be required to achieve meaningful improvement [35].

Quality of life is a crucial long-term outcome indicator for assessing improvements in swallowing function [43]. The meta-analysis did not demonstrate a significant effect of gamified swallowing exercises on improving quality of life, and the wide prediction intervals indicate that the effects could vary widely, potentially benefiting or harming quality of life in different settings. Quality of life, a highly subjective and multidimensional outcome, is influenced by a range of factors, including psychological status, disease duration, and level of social support [44]. Furthermore, most included studies were constrained by small sample sizes and limited follow-up, with outcomes primarily assessed immediately post-intervention, which likely fails to capture the long-term efficacy of the swallowing exercises. Consequently, the current evidence is insufficient to draw definitive conclusions regarding the effect on quality of life, and these findings must be interpreted with caution. Further high-quality studies with larger samples and extended follow-up periods are warranted to validate these effects.

Several limitations should be noted in this study. First, a total of 11 databases were systematically searched, and the search was updated before submission. Ultimately, only 6 RCTs involving 330 participants were included. The limited number of studies and small sample size may restrict the comprehensiveness and generalizability of the findings. Nevertheless, rigorous and standardized methods were applied to ensure the methodological quality and transparency of this review. Second, only 2 studies strictly implemented randomization and blinding, whereas the methodological quality of the remaining 4 studies was suboptimal and requires improvement.

This study indicated that gamified swallowing exercises have a positive impact on swallowing function and adherence in adults with dysphagia, making them a promising rehabilitation approach for clinical practice. When designing and developing games for swallowing training, it is necessary to consider the cause of dysphagia and the accessibility of rehabilitation. Game-based exercises can be precisely designed to align with key rehabilitation goals. Accordingly, oral swallowing disorders may be addressed through games that enhance the strength of the lip, tongue, and jaw muscles, while pharyngeal disorders may be targeted with games designed to strengthen the throat muscles. The 6 included studies all used specialized and complex equipment, such as EMG and facial recognition devices, and, considering the long-term rehabilitation needs of patients, decreasing equipment dependence for gamified swallowing rehabilitation as much as possible is necessary so that individuals with swallowing disorders can engage in gamified swallowing exercises at home.

The current evidence base is characterized by single-center designs and an inherent high risk of performance bias due to the inability to blind participants and therapists. Moreover, the evidence is markedly limited by the absence of long-term outcome data, with only 1 study reporting a 4-week follow-up, which would bring a significant concern given that swallowing rehabilitation is a protracted process. It is therefore imperative that future research prioritizes large-scale, multicenter RCTs with extended follow-up periods to minimize confounding factors and provide higher-quality evidence on the effectiveness and long-term effects of gamified swallowing exercises, as well as to further explore the underlying effect mechanisms of gamified swallowing training. Furthermore, research should also be expanded to develop etiology-specific protocols, moving beyond the current narrow focus on poststroke dysphagia.

This is the first systematic review and meta-analysis to evaluate and synthesize the effects of gamified swallowing exercises among adults with dysphagia. The findings demonstrated potential benefits of gamified swallowing exercises in improving swallowing performance, reducing dysphagia severity, and enhancing adherence among adults with dysphagia, while its effects on dysphagia screening, nutritional status, and quality of life warrant additional research to establish conclusive evidence. Given the limited number and small sample size of included studies, as well as the generally low quality of evidence, these results should be interpreted with caution. This study provides some reference and guidance for healthcare providers in the field of swallowing rehabilitation, such as integrating gamification elements, designing and developing etiology-specific games, and simplifying rehabilitation equipment. In the field of swallowing rehabilitation, helping adults with swallowing difficulties overcome technical barriers and improving the feasibility and scalability of home-based gamified swallowing rehabilitation might be a promising future research direction.