This meta-analysis was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. The PRISMA checklist is provided in Multimedia Appendix 1.

We conducted a comprehensive literature search across 8 databases, namely China National Knowledge Infrastructure (CNKI), Wanfang database, China Science and Technology Journal Database, China Biomedical Literature database (SinoMed), Cochrane library, Web of Science, PubMed, and Embase. We used a combination of subject terms and free terms tailored to each database. The search was performed for studies published from inception up to December 2023. Two researchers independently executed the searches and managed the retrieved literature using EndNote software. Detailed search strategies for each database are provided in Multimedia Appendix 2.

Textbox 1 shows the inclusion and exclusion criteria.

Two researchers independently extracted data from the included studies, including details such as author, publication year, country, sample size, intervention specifics, and outcome measures. Studies were initially screened based on titles and abstracts, followed by a full-text review to ensure they met the inclusion criteria. Discrepancies between the 2 researchers were resolved through discussion, and if consensus could not be reached, a third researcher was consulted for final arbitration.

To ensure comprehensive data extraction, we contacted study authors by email to request missing or unclear data. Studies with unresolved data issues were excluded from the meta-analysis.

The quality of included studies was assessed using the Cochrane Collaboration’s risk of bias tool. We evaluated studies on several criteria, such as random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other potential biases. Each criterion was classified as low risk, unclear risk, or high risk of bias [11].

Statistical analyses were performed using RevMan (version 5.4) software and State (version 17) provided by the Cochrane Collaboration. We analyzed both continuous and dichotomous variables, reporting relative risk with 95% CIs. A fixed-effects model was used for the meta-analysis of outcome indicators. We assessed heterogeneity among studies using the chi-square test and I-squared (I²) statistic, with I² values of 25%, 50%, and 75% indicating low, moderate, and high levels of heterogeneity, respectively. Publication bias was evaluated through funnel plots and Egger regression test [12]. A 2-tailed P value of ≤.05 was considered statistically significant [13]. Due to the limited number of studies, further subgroup and sensitivity analyses were not performed.

In cases where data were incomplete or unclear, we sought additional information from the authors. Studies for which relevant data could not be obtained were excluded from the systematic review and meta-analysis.

A total of 4850 articles were initially retrieved from the literature search, comprising 80 from CNKI, 4 from Wanfang database, 121 from China Science and Technology Journal Database, 38 from SinoMed, 34 from the Cochrane library, 2921 from Web of Science, 656 from PubMed, and 996 from Embase. After a systematic screening process, 12 articles were ultimately included. The detailed screening process and the reasons for exclusion are illustrated in Figure 1.

The included studies were conducted across 7 countries: namely, Poland [14], Spain [15,16], China [17-21], Qatar [22], South Korea [23], Germany [24], and Brazil [25]. One study used a 3-arm design [24], while the remaining 11 studies used a 2-arm design. Of the 12 trials, 2 focused on patients undergoing PD, while the remaining 10 involved patients undergoing HD [17,21]. Five trials compared VR training with conventional care [18-21,23], 2 trials included specified interventions in the control group [14,16], and the remaining trials compared VR training with other types of interventions, such as printed materials [17], exercise programs based on daily habits [15,18], and nurse-led exercise programs [16,22]. Summary of included studies are provided in Multimedia Appendix 3. Details regarding the number of participants, interventions, and outcomes are presented in Table 1. Further information on the interventions under experimental and control conditions is detailed in Multimedia Appendix 2.

According to Cochrane criteria, the risk of bias assessment is illustrated in Figures 2 [14-25] and 3. Of the 12 studies, 7/12 (58.3%) studies [14,15,17,18,21,22,24] were rated as having “some concerns” of bias, and 5/12 (41.7%) studies [16,19,20,23,25] were rated as high risk of bias. In the domain of “random sequence generation,” 5/12 (41.7%) studies [16,17,22,24,25] had low bias, 5/12 (41.7%) studies [14,15,18,20,21] had unclear risk of bias, and 2/12 (16.7%) studies [19,23] had high risk of bias. In the domain of “allocation concealment,” 6/12 (50%) studies [16-18,22,24,25] had low bias, 4/12 (33.3%) studies [14,15,20,21] had unclear risk of bias, and 2/12 (16.7%) studies [19,23] had high risk of bias. In the domain of “blinding of participants and personnel,” 3/12 (25%) studies [21,23,24] had low bias, 7/12 (58.3%) studies [14-19,22] had unclear risk of bias, and 2/12 (16.7%) studies [20,25] had high risk of bias. Blinding participants was not feasible due to the nature of the intervention methods used in all trials. In the domain of “blinding of outcome assessment,” 6/12 (50%) studies [16,19-21,23,24] had low bias and 6/12 (50%) studies [14,15,18,20,22,25] had unclear risk of bias. The absence of blinding outcome assessors or data analysts was observed in studies with some bias. In the domain of “incomplete outcome data,” 9/12 (75%) studies [14,15,17-23] had low bias, 1/12 (8.3%) studies [24] had unclear risk of bias, and 2/12 (16.7%) studies [16,25] had high risk of bias. In the domain of “selective reporting,” 12/12 (100%) studies [14-25] had a low bias.

This meta-analysis encompassed 12 trials involving 625 participants, of whom (99/625; 15.8%) were PD patients. All studies used VR training as the intervention, while control groups received conventional methods such as health education, routine care, and exercise adapted to patients’ daily routines. Of these 12 studies, 5 reported the use of random sequence generation.

This meta-analysis offers a thorough evaluation of the effectiveness of VR training in comparison with traditional methods across various outcome measures, including motor abilities, balance, and psychological well-being. The findings indicate that VR training significantly enhances walking ability, as measured by the 6MWT, with an SD of 29.36. This effect is notably more pronounced compared with gait speed and other motor ability measures, where VR training did not produce significant improvements. Furthermore, VR training was linked to significant reductions in symptoms of depression and anxiety, along with enhancements in social functioning and self-efficacy. However, VR training did not show significant benefits for balance, Sit-to-Stand tests, or the TUG.

The substantial improvement observed in the 6MWT with VR training stands in stark contrast to the absence of significant results in gait speed. The 6MWT evaluates endurance and walking capability over a period, which may be well-suited to the interactive and engaging nature of VR environments. The immersive qualities of VR training may enhance patients’ motivation and engagement during prolonged walking tasks, thereby contributing to better performance on the 6MWT. Silva et al [26] demonstrated that visual and auditory stimuli in virtual environments encourage active participation in training, leading to increased exercise, walking distance, and time, which enhances the effectiveness of the 6MWT and improves patients’ walking ability. Research indicates that gait rehabilitation programs using VR can aid dialysis patients with muscle weakness and joint pain in improving walking speed and quality by simulating diverse walking environments and obstacles, thereby enhancing their independence in daily life [27]. The inclusion of dynamic, goal-oriented tasks in VR may better simulate real-world walking scenarios, motivating participants to exert greater effort and maintain activity for extended periods. In contrast, gait speed, a measure of immediate walking performance, may not benefit as substantially from the immersive features of VR training. The effectiveness of VR in enhancing gait speed may be influenced by the intensity and duration of training sessions, which varied across studies. Consequently, the lack of significant differences in gait speed improvement may suggest a need for more specific training protocols or supplementary interventions to optimize gait speed outcomes.

The observed significant improvement in walking ability among dialysis patients underscores the potential of VR training to address key aspects of patient care. Improving walking capability can significantly impact daily functioning and overall quality of life, particularly in individuals with mobility impairments [28]. The capacity to enhance physical function while concurrently addressing psychological factors, such as depression and anxiety, highlights the comprehensive benefits of VR training. Research by Schröder et al [29] further demonstrates that VR training can alleviate patients’ negative emotions, including anxiety and depression. By offering a multifaceted therapeutic approach, VR has the potential to enhance both physical performance and mental well-being, essential components of holistic patient care. Furthermore, the effectiveness of VR in reducing depression and anxiety suggests that VR training may serve as a valuable tool for mitigating the broader psychosocial impacts of chronic illness [30]. The immersive and interactive features of VR provide an engaging and motivating environment that fosters positive changes in mood and self-perception, thereby contributing to improved overall patient outcomes [31].

The findings of this meta-analysis demonstrate that VR training significantly improves social functioning in patients, as evidenced by a SD of 16.2. This result suggests that VR training can substantially enhance social interaction and integration, which are crucial aspects of overall well-being. The positive effect of VR on social functioning may be attributed to various factors inherent to the VR experience. VR environments frequently offer engaging and interactive scenarios that promote social engagement and collaboration, even among individuals with limited mobility or social anxiety. VR training usually includes interactive tasks and group-based activities that encourage active participation, thereby enhancing patients’ social skills and interactions. The immersive nature of VR provides opportunities for patients to practice social scenarios in a controlled environment, thereby enhancing their confidence and comfort in real-world social interactions. Furthermore, participation in VR-based social activities may foster a sense of community and support, further enhancing social functioning. Despite these promising results, it is crucial to consider the variability in social functioning outcomes reported across studies. The observed heterogeneity (I²=72%) indicates that factors such as the design of VR interventions, the nature of social interactions, and the duration of training may affect the effectiveness of VR in enhancing social functioning. Future research should focus on exploring these variables in detail and identifying the specific components of VR training that most effectively improve social interaction and integration [32].

Virtual reality training demonstrated a significant improvement in self-efficacy, with an SD of 20.47, indicating that VR interventions substantially enhance patients’ confidence in their ability to manage various tasks and challenges. Self-efficacy, defined as the belief in one’s ability to achieve goals, is a crucial determinant of both physical and psychological health. The observed improvement in self-efficacy with VR training can be attributed to the interactive and engaging nature of VR environments, which frequently include goal-setting, feedback, and achievement [33]. VR training typically integrates elements such as progress tracking, rewards, and virtual challenges, which can enhance patients’ confidence in their abilities. By offering immediate feedback and a sense of accomplishment, VR can reinforce positive behaviors and attitudes, thereby contributing to improved self-efficacy. Furthermore, the immersive and supportive features of VR environments may help patients overcome fears and build confidence in a controlled setting, potentially translating into improved self-efficacy in real-world situations [34]. However, the high heterogeneity (I²=99%) in the self-efficacy outcomes suggests that the effectiveness of VR in enhancing self-efficacy may be influenced by several factors, including the design of the VR interventions and individual patient characteristics. Future research should focus on identifying the most effective VR components for enhancing self-efficacy and investigating how different patient characteristics may influence the outcomes of VR training.

The use of VR training among PD patients has been relatively limited compared with other patient populations. Several factors are likely to contribute to this limited use. PD patients face distinct challenges, such as the requirement for specialized equipment and potential complications related to fluid management, which may complicate the integration of VR training. In addition, this patient population may have unique physical and psychological conditions that necessitate tailored interventions. Current research on VR training for this population is still emerging, and the application of VR may need adaptation to address these specific needs effectively.

Furthermore, PD patients may experience varying levels of physical capability and psychological stress compared with other patient groups, which may influence the outcomes of VR interventions. Future research should focus on developing and testing VR programs specifically tailored for this population, considering their unique requirements and limitations to ensure that the interventions are both safe and effective.

This meta-analysis highlights the potential of VR training as an adjunctive therapy for enhancing walking ability and psychological well-being in dialysis patients. The substantial improvements in walking endurance and psychological outcomes observed in this analysis suggest that VR training may be a valuable addition to existing therapeutic strategies. Nevertheless, the observed variability in results and heterogeneity in certain outcomes underscores the need for further research.

The strengths of this study include (1) it is the first meta-analysis to assess the impact of VR training on dialysis patients and (2) we assessed the effectiveness of VR training using 12 outcome measures: gait speed, balance, 6-minute walk test score, STS-5, STS-10, STS-60, TUG, depression, anxiety, fatigue, social functioning, and self-efficacy, providing a comprehensive reference for future research.

This study also has limitations: (1) it included only published articles in Chinese and English, which may impact the meta-analysis results, (2) some studies did not provide details on sequence generation, allocation concealment, and blinding, and (3) the 12 studies included in the analysis used different control interventions and the intervention durations across the studies varied significantly, which may have contributed to significant heterogeneity among the studies.

Future research should focus on refining VR interventions, exploring their optimal implementation, and addressing the specific needs of diverse patient populations. Larger and more diverse sample sizes, coupled with standardized training protocols, will be crucial for comprehensively evaluating the benefits and limitations of VR.

In this study, we used multiple databases for literature retrieval to ensure comprehensive coverage of relevant research in the field. Our selected databases included CNKI, Wanfang, China Science and Technology Journal Database, SinoMed, Cochrane Library, Web of Science, PubMed, and Embase. These databases offer a wealth of research outputs, particularly the Chinese databases, which aggregate a substantial amount of high-quality studies conducted by Chinese scholars, reflecting the region’s active contribution to this field. However, we recognize that our current search strategy may not fully represent the entire research landscape, particularly the lack of studies from North America.

Although our initial search did not yield relevant North American research, this may be attributed to the limitations of our search strategy and keyword selection. We acknowledge that this omission could impact the representativeness of our results and potentially influence the conclusions drawn from the study. Therefore, we emphasize this limitation within the article and recommend that future research consider broader literature retrieval strategies to ensure a more holistic understanding of the field.

This reflection provides important guidance for subsequent studies, ensuring that future efforts can better integrate research findings from diverse regions, thereby enhancing the breadth and depth of analysis.

This meta-analysis offers a thorough evaluation of the use of VR training in dialysis patients. The findings indicate that VR training effectively enhances patients’ physical fitness, alleviates anxiety and depression, and improves adherence, thereby facilitating better self-management. However, research focusing on VR training in peritoneal patients is relatively limited compared with that on hemodialysis patients. Future research should involve PD patients and use rigorous, large-scale studies to validate the findings of this review for both hemodialysis and PD populations.