Effectiveness of technology-based exercise on motor function and quality of life in individuals with multiple sclerosis: an umbrella review

Abstract

This umbrella review evaluated the effects of Technology-Based Exercise (Tele-exercise) on balance and Quality of Life (QoL) in individuals with multiple sclerosis (MS). A comprehensive search across multiple databases identified meta-analyses and individual studies published up to June 3, 2025. Eligible studies included meta-analyses assessing tele-exercise interventions, while reviews with insufficient data were excluded. Methodological quality was assessed using the Measurement Tool to Assess Systematic Reviews 2 (AMSTAR-2) checklist. The pooled findings indicate that tele-exercise is not associated with meaningful improvements in balance among individuals with multiple sclerosis (SMD = 0.587, 95% CI: -0.533–1.706, N = 243 based on SMD). Likewise, tele-exercise shows minimal impact on quality of life (QOL) (SMD = 0.169, 95% CI: -0.025-0.362, N = 407 based on SMD). Although trends toward benefit were observed, findings remain inconclusive due to variability in protocols, methodological limitations, and low patient engagement. Overall, tele-exercise is feasible and may support patient-centered care, but current evidence does not confirm significant improvements in balance and QoL for people with MS. Standardized protocols, greater personalization, and higher-quality trials are needed to strengthen future clinical guidelines.

Introduction

Multiple sclerosis (MS) is a persistent, inflammatory, and progressive disorder affecting the central nervous system, commonly diagnosed between the ages of 20 and 40, although it may occasionally present in children and older adults [1]. While the precise etiology of MS remains unclear, it is thought to arise from a complex interplay between hereditary predisposition and environmental influences [2]. Demyelination disrupts neuronal signal conduction, leading to symptoms such as visual disturbances, fatigue, pain, paresthesia, and impairments in motor coordination and balance [3]. These challenges often result in reduced physical activity and a sedentary lifestyle, which further diminishes, Quality of Life (QoL) due to both motor and cognitive limitations [4].

Daily functional activities such as eating, dressing, grooming, toileting, and mobility are essential for well-being and survival [1, 5]. Among the estimated 50 million individuals affected by MS worldwide, 25% to 74% require partial or full-time caregiving, highlighting the profound impact of MS on functional independence [4]. Functional dependence is a key metric in evaluating the success of MS rehabilitation programs [6, 7]. Neuroimaging has shown structural changes in the bilateral amygdala, accumbens, hippocampus, and right frontal lobes after the onset of MS, including both delayed maturation and neurodegeneration [8, 9].

As a result, various rehabilitation strategies have been created to improve both motor and cognitive function in MS patients. These include physical and occupational therapies that emphasize high-intensity, repetitive, and task-oriented interventions across different stages of disease progression [10]. Pharmacologic treatments—such as psychostimulants—and non-pharmacologic approaches like psychotherapy and cognitive-behavioral therapy are commonly employed [11]. However, challenges such as medication tolerance, side effects, and lack of therapeutic response highlight the need for alternative approaches.

Tele-rehabilitation (TR) has surfaced as an encouraging modality, leveraging technology to provide accessible care for MS patients. TR can help improve physical function and quality of life, especially for those with limited access to healthcare due to geographic or physical limitations [12]. Still, some studies report no significant differences in outcomes such as QOL, balance, or fatigue when comparing TR with traditional in-person interventions [12, 13]. Technologies, including telephone, internet-based videoconferencing, and mobile applications, enhance patient-provider communication [14]. Both virtual reality and TR have shown potential in supporting self-care and caregiver involvement in MS management [14].

Comparative studies assessing TR effectiveness frequently use various outcome measures [14]. One analysis identified 30 different tools employed in randomized controlled trials (RCTs) for MS treatment [15]. For a new treatment to be considered viable, it must demonstrate effectiveness comparable to or exceeding current standards [12]. Therefore, standardized and context-appropriate outcome measures are crucial for both research and clinical implementation [16–18].

In conclusion, three essential considerations guide the effective use of TR in MS care: (a) employing standardized outcome measures, (b) identification of context-specific interventions, and (c) integration of organizational metrics. Despite its cost-effectiveness and accessibility—especially valuable during the COVID-19 pandemic and in rural settings—there is still no consensus on standardized outcome measures for TR [19]. Given these gaps, this umbrella review aims to evaluate the success of TR in enhancing patient outcomes in patients with MS.

Methods

Search strategy

This umbrella review synthesizes findings from published meta-analyses to evaluate the impact of technology-based exercise (tele-exercise) interventions on balance and QoL in individuals with multiple sclerosis (MS). The methodology adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and recommendations from Aromatases et al. Two independent reviewers conducted screening, data extraction, and study evaluation.

Ethical approval for this study was granted by the Ethical Committee of Tehran University of Medical Sciences (IR.TUMS.MEDICINE.REC.1403.514).

A systematic search was performed by two reviewers (HA, MGH) across databases including Web of Science, EMBASE, Cochrane Library, PubMed, SPORT Discus, Google Scholar, and Cumulative Index to Nursing and Allied Health Literature (CINAHL). The search covered studies from database inception to June 3, 2025. The selection process involved a stepwise review of titles, abstracts, and full texts. References of included studies were also screened. Conflicts were addressed through discussion or adjudication by the corresponding author (CHPS).

Accordingly, the literature search was restricted to article titles and guided by keywords associated with “tele rehabilitation,” “remote exercise,” “telehealth,” “home exercise,” and “technology exercise,”. Additionally, terms related to “motor function” were included, such as “balance”, “walking”, “gait”, “center of pressure”, and “quality of life” as dependent variables. The scope of the search was additionally narrowed to studies containing “systematic review with meta-analysis” in the context of MS. A full description of the search procedures, with the exact terms and database-specific syntax, is provided in Supplementary Material 1.

Study selection

Inclusion criteria targeted meta-analyses and systematic reviews of prospective studies (e.g. RCTs), evaluating remote physical interventions and their impact on motor function, QoL, or MS-related outcomes. No language restrictions were applied. Exclusion criteria included reviews not focused on MS, lacking quantitative data, or limited to animal studies or conference abstracts. Where duplicate meta-analyses existed, those with higher AMSTAR 2 scores were prioritized [20]. If scores were similar, the study with more included trials was selected. Overlapping data were managed by excluding duplicates, following Cochrane guidelines [21].

Conversely, scoping reviews, narrative reviews, and systematic reviews without meta-analysis, qualitative reviews, and studies lacking a clearly defined methodology for estimating prevalence were excluded. This decision was warranted, as such formats fail to provide primary findings from multiple RCT studies in a meta-analysis. Similarly, studies conducted in narrowly defined populations (e.g., exclusively in patients with MS) were pooled, while systematic reviews with meta-analysis that combined several neurological conditions, such as Parkinson’s disease, stroke, and MS, were excluded except in cases where the analysis allowed the MS subgroup to be isolated and extracted. Duplicate publications, as well as studies in which data could not be separated to isolate the effects of remote exercise on MS were also excluded, thus avoiding double-counting results or inaccurate data in the umbrella review.

Data extraction

The PRISMA flow diagram (Fig. 1) describes the search process and selection criteria (PRISMA checklist is provided in the supplementary material 2). Two reviewers (HA, MGH) extracted data independently, with discrepancies resolved by a third reviewer (ASH). For every meta-analysis, we extracted data on the first author, year of publication, sample sizes, adjusted effect estimates with 95% confidence intervals, and study designs. Analytical methods such as standardized mean difference (SMD) and mean difference (MD) were documented. Risk of bias assessments and methodological quality scores were also included.

Fig. 1.

The Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) flow chart of the literature search and screening process

Of the 510 studies identified, 290 duplicates were removed. After full-text screening, 115 studies were excluded for reasons including irrelevant outcomes, inadequate data, and absence of meta-analytic methods. Ultimately, three eligible meta-analyses were selected—all based on RCTs and reporting outcomes using SMDs. Table 1 summarizes the details of the included studies.

Table 1.

Summary of the included studies

Authors & year, country	Number of included trials	Total sample size	Intervention type	Outcomes	Results	Adverse effects
Di Tella S. et al. 2019, Italy	Nine studies in MS	716	Integrated telerehabilitation approach (ITA)	Motor disability, gait and balance, cognition performance, depression, fatigue, daily functioning, quality of life and self-efficacy	Large effect on motor disability, gait and balance, small effect on cognition performance, medium effect depression, fatigue, daily functioning, quality of life and self-efficacy	No mention
Federico S. et al.2023, Italy	Eight studies in MS	407	Physiotherapy intervention through TR	Quality of life	The results suggest a significant improvement in QoL in patients who underwent TR	No mention
Truijen S. et al. 2022, Nigeria	Seven studies in MS	287	Home-based virtual reality (VR) training and telerehabilitation (TR)	Balance the Berg Balance Scale (BBS)	There was an improvement in BBS scores over time in both experimental and control groups	No mention

These studies primarily focused on balance and QoL. Balance was assessed using tools such as the Berg Balance Scale and Choice Stepping Reaction Time. QoL was measured with instruments like the Multiple Sclerosis Quality of Life-54 (MSQOL-54), the Multiple Sclerosis Impact Scale, and the Hamburg Quality of Life Questionnaire. Two of the studies reported no significant improvements in balance or QoL (P > 0.05), and none found significant effects of TR on other outcomes such as mobility, upper limb function, or depression. Thus, this umbrella review focuses on evaluating the effects of TR on balance and QoL in MS patients.

Study quality assessment

The AMSTAR 2 checklist was used to assess all included meta-analyses [20]. Table 2 provides a complete summary of key characteristics from each meta-analysis, including effect sizes, risk of bias assessments, methodological quality, and the sample size per group. Of the three secondary studies evaluated, half were judged to be of moderate to high quality (Table 2). As shown in Table 3, the quality of systematic reviews with meta-analyses of RCTs was assessed using 16 evaluation questions. In line with Shea et al. 2017, the overall confidence in a review’s findings was categorized as follows: High (No or one non-critical weakness; the review offers an accurate and comprehensive synthesis of the available evidence relevant to the research question), Moderate (more than one non-critical weakness but no critical flaws; the review may still provide a reasonably accurate summary of available studies). Low (one critical flaw with or without non-critical weaknesses; the review may not provide a reliable or comprehensive summary of the evidence), critically low (multiple critical flaws with or without non-critical weaknesses; the review cannot be considered a dependable source of evidence) [20].

Table 2.

Effects of technology-based exercise on motor function and quality of life in individuals with multiple sclerosis

Authors & Year	Outcome	Measure	Cases	Studies	P-value	95% CI	I²	Egger (Z, P-value)	Assessment
Di Tella S. et al. 2019	Balance	BBS	74	2	0.025	−2.16 to −1.20	80.1%	Z = 0.56, P = 0.575	High
Di Tella S. et al.2019	Quality of Life (QOL)	MSQOL-54	171	2	0.932	−0.36 to 0.54	0%	Z = 0.38, P = 0.702	High
Truijen S. et al.2023	QOL	MSQOL-54	407	7	0.590	−0.37 to −0.03	0%	Z = 1.71, P = 0.090	Critically low
Truijen S. et al. 2022	Balance	BBS	169	4	0.230	−1.71 to 2.64	31%	Z = 0.76, P = 0.450	Moderate

Outcome: The specific variable or health-related measure analyzed in the study

Measure: The tool or scale used to assess the outcome (e.g., BBS for Balance, MSQOL-54 for QOL)

Cases: Total number of participants included in the analysis

Studies: Number of studies included in the meta-analysis for the outcome

P-value: Statistical significance from the random-effects model

95% CI: Confidence interval for the effect size, showing estimate precision

I²: Percentage of variation across studies due to heterogeneity rather than chance

Egger: Egger’s test statistic and p-value for assessing small study effects (publication bias)

Assessment: Overall evaluation of heterogeneity, small study effects, or excess significance bias

AMSTAR: Assess Systematic Reviews 2, CI: Confidence interval, SMD: standard mean difference, RCT: Randomized Controlled Trails, BBS: Berg Balance Scale; MSQOL-54: multiple sclerosis quality of life 54

Table 3.

AMSTAR 2 checklist

Data analysis

Umbrella review framework and statistical analysis

This study was conducted as an umbrella review in accordance with PRISMA guidelines and established recommendations for umbrella analyses. Data were synthesized from six studies evaluating balance and seven studies evaluating QoL, comprising 243 participants (132 cases, 111 controls) and 407 participants (205 cases, 202 controls), respectively. All individual effect estimates from included meta-analyses are provided in Supplementary Table S1.

Effect sizes for each outcome—balance or QoL—were extracted directly from the included meta-analyses. Table S1 presents the original standardized mean differences (SMDs) and their 95% confidence intervals as reported in the source publications. No re-pooling of primary trials was performed; instead, the published effect sizes from each meta-analysis were reproduced to ensure transparency and comparability. All extracted values were cross-checked against the original publications to confirm accuracy.

To standardize effect size estimation and enhance replicability, the meta umbrella R package was used [21, 22]. Unlike conventional narrative umbrella reviews, meta umbrella requires extraction of the individual study-level data from the included meta-analyses (e.g., author, year, sample size, reported effect size, and 95% CI). This process does not involve conducting a new independent meta-analysis of randomized controlled trials; rather it allows the umbrella review to replicate each meta-analysis result within a standardized framework and consistently re-estimate pooled effect sizes across reviews [23]. Following package guidelines, all extracted data were formatted into structured datasets to enable computation of standardized mean differences (SMDs), heterogeneity (I²), prediction intervals, and small-study effects. This approach minimized methodological heterogeneity between reviews and facilitated assessment of the credibility of evidence using uniform statistical criteria.

From the eligible meta-analyses [14, 24, 25], we identified six primary studies contributing to balance outcomes and seven primary studies contributing to quality of life (QoL) outcomes. These trials were not independently identified; rather, they represent studies already included in the selected meta-analyses. Their characteristics and effect estimates were harmonized using the metaumbrella workflow, which re-computes pooled effects from the study-level data reported in each meta-analysis. This procedure allows for consistent synthesis and replication of the published meta-analyses while maintaining the umbrella review framework.

Given the anticipated heterogeneity across trials, overall effect sizes were recalculated using a random-effects model. Effect sizes were presented as both the original standardized mean difference (SMD) and an equivalent odd ratio (eOR) to facilitate cross-scale interpretation.

Continuous outcomes were primarily expressed as standardized mean differences (SMDs) with corresponding 95% confidence intervals. The metaumbrella package also generates equivalent odds ratios (eORs) using the transformation formula eOR = exp (SMD × π/√3), which assumes a logistic distribution with equal standard deviations across groups [22, 23]. This conversion improves interpretability across different effect size metrics but should be regarded as approximate rather than exact. Consistent with best practice, SMDs are reported as the primary effect size, with eORs presented only as supplementary information to aid interpretation [22].

Study heterogeneity was measured using the I² statistic, which estimates the proportion of total variance due to heterogeneity rather than chance. An I² value above 50% was considered indicative of heterogeneity. Subgroup analyses explored the differential effects of tele-exercise by MS severity, while sensitivity analyses tested the robustness of pooled results by successively removing each study and recalculating effect estimates.

All statistical tests were two-tailed, with p-values < 0.05 considered significant. Effect estimates were reported with 95% confidence intervals (CIs), indicating the range within which the true effect is expected to lie with 95% certainty.

Egger’s regression asymmetry test was performed only when a meta-analysis included at least 10 studies, consistent with established guidelines, as results based on fewer studies may be unreliable [23]. A p-value < 0.05 was considered indicative of potential small-study effects, though findings should be interpreted cautiously, especially in the presence of high heterogeneity or limited study numbers. Between-study heterogeneity was quantified using the I² statistic and classified as low (< 25%), moderate (25–49%), high (50–74%), and very high (>75%) [22]. These thresholds were applied consistently across all analyses to distinguish true heterogeneity from random variation.

Software

All statistical analyses were performed using R version 4.4.0 (R Core Team, 2023) with the RStudio interface (RStudio Team, 2023).

Results

Literature search

The results of the umbrella review are summarized in Fig. 2, which illustrates the effects of tele-exercise interventions on balance and quality of life in individuals with MS. Panel (a) presents the pooled standardized mean differences (SMDs) with 95% confidence intervals, the primary effect measures in this review. Panel (b) displays the equivalent odds ratios (eORs) automatically generated by the metaumbrella package, included only as supplementary values and intended for approximate interpretation.

Fig. 2.

Pooled SMD effect size (a) and estimated eOR (b)

For balance, the pooled standardized mean difference (SMD) was 0.59 (95% CI: − 0.53 to 1.71), reflecting a moderate but non-significant effect (Fig. 2a). The wide confidence interval crossing zero reflects indicates uncertainty, and heterogeneity was moderate (I² = 48.3%). Corresponding equivalent odds ratios (eORs) are provided in Fig. 2b as secondary, approximate measures.

For QoL, the pooled SMD was 0.17 (95% CI: − 0.025 to 0.362) (Fig. 2a), suggesting a weak, non-significant effect. Compared with balance, the narrower confidence interval indicates greater consistency across studies, although statistical significance was not achieved (p = 0.087). No heterogeneity was observed (I² = 0%). As with balance, eORs are reported in Fig. 2b only as supplementary information.

In summary, tele-exercise showed a non-significant effect on both balance and QoL. The balance outcome demonstrated greater ambiguity due to high variability and wide confidence intervals. While QoL outcomes showed more consistent estimates, they still lacked statistical significance.

Discussion

This umbrella review found no significant effects of tele-exercise on balance or QoL in patients with MS. As the first synthesis focused specifically on these outcomes in the tele-exercise context, the findings highlight the limited evidence supporting its efficacy. Consistent with the predefined methodology, we present standardized mean differences (SMDs) with 95% confidence intervals are reported as the primary effect estimates, while equivalent odds ratios (eORs), generated by the metaumbrella package are provided only as supplementary information.

Egger’s tests were conducted only for analyses including at least 10 studies; for these, Z-scores and p-values were reported to ensure transparency regarding potential small-study effects. In this review, none of the performed Egger’s tests indicated significant small-study effects.

Among the three systematic reviews analyzed, only one addressed both outcomes, while others included related variables such as depression, upper limb function, and mobility. None demonstrated significant improvements in balance (n = 243) or QoL (n = 407) [14]. Balance is a key motor function affecting QoL in MS, yet reviews by [24] and [25] concluded that tele-exercise is not the most effective intervention for improving this domain [24, 25].

Levels of heterogeneity differed across the included meta-analyses. For balance outcomes, [24] reported very high heterogeneity (I² = 80.1%), reflecting considerable variability between the two included studies. In contrast, [25] found only moderate heterogeneity (I² = 31%) across four balance studies, indicating some inconsistency but less pronounced. For quality of life (QOL), both [24] and [14] reported low heterogeneity (I² = 0%), reflecting highly consistent findings across the included studies.

Assessment of small-study effects using Egger’s regression was restricted to analyses with at least 10 studies, in line with methodological recommendations. Within this review, none of the tests performed showed statistical significance (p > 0.05), indicating no apparent evidence of publication bias. Egger’s test was not applied to analyses with fewer than 10 studies due to its limited reliability in small samples. Taken together, these results suggest that heterogeneity and potential small-study effects varied by outcome: balance analyses demonstrated greater variability, whereas QOL outcomes were more stable and consistent across studies.

Participation in daily activities and frequent session engagement have demonstrated the ability to decrease motor impairment and improve QoL. [14] found that while TR may benefit QoL in some neurological conditions, the effects for MS are modest and influenced by individual factors and intervention design [14]. Although the current review found a weak trend toward improvement in QoL (p = 0.08), consistent tele-exercise programs might still yield long-term benefits.

Guidance from trained therapists, alongside caregiver involvement, can enhance outcomes, especially when sessions are frequent [25]. Even without in-person presence, interactive TR enables therapists to provide real-time supervision, fostering patient confidence and safety. This collaborative care model can improve engagement and outcomes for MS patients [26–28]. Importantly, clinicians can still adapt existing protocols by emphasizing higher-frequency, shorter-duration sessions (e.g., 3–5 times per week), prioritizing interactive video-based delivery over phone-only models, and incorporating multimodal exercise components such as aerobic, resistance, and balance training. Tailoring interventions to patient preferences and functional status may improve adherence and increase the likelihood of meaningful benefit despite the current lack of definitive evidence.

Consistent with previous research, this review supports TR as a feasible and well-received strategy that may enhance QoL in patients with MS [25, 26]. Many QoL focused studies integrated physical therapists into care teams. Their involvement has been shown to: (a) increase adherence and motivation; (b) strengthen therapeutic relationships that boost patient confidence and self-efficacy [25, 29, 30]. (c) improve cost-effectiveness. For example, Llorens et al. (2015) reported that TR was $654 less expensive than in-clinic programs, with comparable efficacy, especially when session frequency was higher [31]. This clearly indicates that TR can reduce travel costs, positively impacting the overall healthcare system expenses [32].

Despite the predominantly non-significant findings, it is notable that no studies reported strong positive effects. This may reflect limitations in intervention design, such as inadequate teaching strategies or reliance on phone-based rather than video-based delivery modes. Further research should aim to standardize tele-rehabilitation protocols, optimize session frequency, and explore patient-specific adaptations to maximize effectiveness.

Chen et al. 2023 demonstrated that patients undergoing TR exhibited significant increases in time-varying amplitude of low-frequency fluctuations, regional homogeneity, and functional connectivity in the precuneus. These findings suggest that TR may promote increased dynamic regional brain activity and strengthen functional connectivity within the precuneus. The precuneus is extensively interconnected with multiple cortical regions (e.g., medial parietal cortex, inferior and superior parietal lobules, prefrontal cortex, premotor area, and supplementary motor area) and subcortical structures (including the thalamus, dorsolateral caudate nucleus, and putamen). These connections support the integration of external and self-generated information and regulation of mental processes. However, the study by Chen et al. 2023 did not identify a strong association between increased dynamic changes in the precuneus and enhancements in motor function. This suggests that while increased spontaneous brain activity and connectivity within the precuneus could impact motor function via cognitive modulation, it is unlikely to be a primary driver of motor recovery [33].

Strengths of the study

This umbrella review employed a complete search strategy to identify relevant systematic reviews. A notable strength lies in its inclusion criteria, which were limited to systematic reviews with meta-analyses based on RCTs, thereby synthesizing high-quality evidence with reduced risk of bias. The review was conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions, ensuring methodological rigor.

Methodological limitations

Several of the included studies present methodological limitations, most notably the lack of control groups, which undermines the reliability of their conclusions. Differences in participant demographics and disease severity, further contribute to inconsistent findings on the impact of tele-exercise on balance and QoL [29]. Moreover, variability in exercise protocols and outcome assessment methods introduce additional heterogeneity, making comparisons across studies difficult. Subtle benefits of of tele-exercise may also go undetected, particularly in populations with differing levels of disability [34].

In our umbrella review, only three systematic reviews with meta-analyses were ultimately included, and no additional eligible studies on TR were identified. Related studies involving technology-based exercise interventions were either not aligned with remote delivery or overlapped with existing studies. This limited the ability to incorporate a broader evidence base.

A possible reason for the limited number of studies may be a declining research interest due to inconclusive or modest effects of TR on motor function and QoL in individuals with MS. This restricts the strength of current conclusions and emphasizes the need for renewed research efforts.

We also recognize the impact of heterogeneity across studies, stemming from differences in participant characteristics, intervention protocols, and control conditions. While subgroup analyses or meta-regression could have provided insight into these sources of variability, such analyses were not feasible due to data limitations. Therefore, our findings warrant careful interpretation, and future research should aim to produce more consistent and homogenous data.

To improve future research quality, we recommend the development of standardized intervention protocols and clearer definitions of TR modalities. These improvements would enhance the reliability and generalizability of findings in future umbrella reviews.

Furthermore, although assessing the methodological quality of individual trials could provide additional insights into potential biases, we focused on evaluating the quality of the systematic review themselves. Future work should include an in-depth assessment of individual studies to further strengthen the rigor of evidence synthesis.

Another limitation is the wider confidence intervals observed for Balance compared to Quality of Life, which can be explained by several factors. First, the meta-analyses for Balance were based on fewer studies with smaller sample sizes (N = 243) and exhibited substantial heterogeneity (I² = 80%), both of which increase statistical uncertainty and lead to wider intervals. In contrast, the analyses for Quality of Life included a larger number of participants (N = 407) and showed no heterogeneity (I² = 0%), resulting in more precise estimates with narrower confidence intervals. Second, the conversion of standardized mean differences into equivalent odds ratios can amplify uncertainty when the original estimates are based on heterogeneous or limited data, as was the case for Balance outcomes. Taken together, these methodological and statistical factors account for the observed differences in confidence interval widths between Balance and Quality of Life.

Lastly, in this umbrella review, the methodological quality of the included meta-analyses was assessed using the AMSTAR-2 checklist, which is specifically designed to evaluate systematic reviews and meta-analyses. Unlike GRADE, which is primarily applied to assess the certainty of evidence derived from primary studies, AMSTAR-2 provides a structured appraisal of the rigor and reporting quality of secondary research. Since our review synthesized findings from published meta-analyses rather than directly analyzing primary randomized controlled trials, AMSTAR-2 was considered the most appropriate tool. While a GRADE assessment could further inform the certainty of the underlying evidence, this was beyond the scope of the present work. Future research may benefit from incorporating GRADE in parallel with AMSTAR-2 to provide a more comprehensive evaluation of both methodological quality and evidence certainty.

Recommendations

This umbrella review provides several recommendations for future research. Defining a minimal effective protocol, including timing, duration, and intensity (e.g., 150 min of moderate exercise per week), is essential. Research should also explore safe thresholds for volume and intensity in TR for individuals with MS. Determining which activity levels most effectively enhance motor function and improve QoL is a priority. Future studies should examine functional exercise interventions that mimic daily tasks and assess their impact across diverse cultural contexts.

We also hypothesize that heterogeneity in study outcomes may stem from variations in patient demographics, treatment protocols, and provider experience. However, given the observational nature of our review, we did not perform detailed analyses to assess these variables. Future research should investigate these factors more thoroughly. The use of outcome measures with limited sensitivity (e.g., ceiling effects in balance scales) may help explain why findings on tele-exercise remain inconclusive. The Italian SARA has been validated as a reliable instrument for assessing ataxia severity, including balance and mobility, in MS, and its strong psychometric properties support application in both clinical practice and research [35]. Similarly, the Fatigue Severity Scale (FSS) is a well-established tool for evaluating fatigue. Further work should examine the responsiveness of these measures, explore their performance across different settings, and investigate potential sources of heterogeneity. Establishing cutoff values in healthy populations would also be useful for contextualizing fatigue symptoms and clarifying differences in perception across groups [36].

Although this review contributes valuable insights into TR effectiveness, further studies, particularly targeted RCTs, are necessary to validate these findings. Such research should evaluate the impact of different TR modalities, dosages, and intensities on outcomes such as activities of daily living (ADL) and motor function. Greater clarity on optimal intervention parameters would provide actionable guidance for clinicians and enhance generalizability.

Lastly, while existing literature supports the safety and cost-effectiveness of TR, our review did not include direct data on costs or adverse events. Future research should address these areas and determine which TR approaches are most effective for patients with differing MS severities and needs, thereby guiding remote rehabilitation practices more effectively.

Conclusion

While tele-exercise holds promise for improving balance and QoL in individuals with MS, current evidence is inconsistent. Variability in study outcomes, methodological concerns, patient adherence challenges, and the need for tailored interventions highlight the complexity of evaluating TR effectiveness. When making decisions about TR implementation, it is crucial to prioritize large, well-designed studies with sufficient intervention duration and consistent methodology. Addressing these research gaps will be essential for optimizing TR strategies for individuals with MS.

Abbreviations

ADLs

Activities of daily living

AR

Augmented reality

CINAHL

Cumulative index to nursing and allied health literature

eOR

Estimated odds ratio

MD

Mean difference

PRISMA

Preferred reporting items for systematic review and meta-analysis

RCT

Randomized controlled trial

SMD

Standard mean difference

TR

Tele-rehabilitation

VR

Virtual reality

WHO

World Health Organization

Technology-Based Exercise

Tele-exercise

Author contributions

M.G.N., A.S. conceived the idea for the review. M.G.N., M.H., Z.B.E., and A.S. conducted the search, study selection, data extraction, and quality assessment. M.G.N., F.M., L.R., and A.S. drafted the initial manuscript. L.R. and A.S. contributed in review and supervision. L.R., N.M., K.B. and A.S. contributed in project administration and manuscript editing. All authors reviewed and approved the final manuscript.

Funding

This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets analyzed in this review are available from the corresponding author on reasonable request.

Code availability

Not applicable.

Declarations

Ethics approval and consent ot participate

Ethical code received from Ethical Committee of Tehran University of Medical Sciences (IR.TUMS.MEDICINE.REC.1403.514).

Competing interests

The authors declare no competing interests.

Supplementary Information

The online version contains supplementary material available at 10.1186/s41043-025-01158-w.