Feasibility and effectiveness of physical exercise for older adults delivered remotely via videoconferencing—systematic review and meta analysis

Abstract

Background

Physical activity promotes healthy ageing. Videoconferencing enables delivery of accessible exercise therapy, supporting older adults in engaging in structured physical activity.

Objectives

To investigate the feasibility and effectiveness of remotely supervised videoconferencing exercise therapy for adults over 65.

Data sources

Nine databases were searched, with manual search completed in March 2025.

Study selection

Randomised controlled trials assessing remote exercise effectiveness or feasibility, delivered via videoconference with professional real-time interaction.

Data extraction and synthesis

Two authors independently screened studies, extracted data, and assessed methodological quality. Meta-analysis used random effects models, with primary conclusions based on high methodological quality studies.

Results

Twenty-eight studies (n = 2086) were analysed. Results are given for the 18 high-quality studies, including 7 low and 11 moderate risks of bias studies. Average attendance and completion rates were 79.7% and 92.8%, respectively. In videoconference exercise, the number of adverse events was comparable to control groups. Videoconference exercise showed significant improvements over no intervention in: physical function [standardised mean difference (SMD) = 0.82; 95% CI 0.46–1.19; P < .00001], muscle strength (SMD = 0.86; 95% CI 0.46–1.26; P < .0001), emotional status (SMD = 0.72; 95% CI 0.35–1.09; P = .0001) and quality-of-life (SMD = 0.61; 95% CI 0.19–1.03; P = .004). High-quality studies showed comparable effectiveness to alternative delivery methods across all measures.

Limitations

Only 25% of studies met the highest methodological standards. High heterogeneity in outcome measures and small sample sizes may affect generalizability.

Conclusion

Videoconference exercise therapy appears safe, feasible and effective for improving physical function, strength, emotional status and quality-of-life compared to no intervention. Its equivalence to alternative delivery is established for all measures.

Key Points

• Videoconferencing-based exercise for community-dwelling older adults demonstrates high adherence rates, with no severe adverse events reported in the reviewed studies.

• Evidence suggests videoconferencing exercise may be comparable to traditional exercise methods in improving physical, muscle strength, emotional and quality of life measures among older adults.

• Notable gaps remain in the current literature, highlighting the need for additional high-quality studies that adhere to World Health Organization recommended exercise doses and utilise standardised reporting tools to enable more robust meta-analyses.

Introduction

Sedentary lifestyles among older adults are associated with functional limitations and increased morbidity, while physical activity offers numerous benefits, including improved functional abilities, cognitive function, mood and reduced fall risk [1]. It can also alleviate loneliness through group exercise [2]. The World Health Organization (WHO) recommends older adults to engage in at least 150–300 minutes of moderate-intensity or 75–150 minutes of vigorous-intensity physical activity weekly, along with balance and strength exercises, to maintain functional capacity and reduce fall risk [3].

Most exercise programmes are delivered face-to-face. These programmes have been shown to be effective when delivered in groups or individually [4]. The effectiveness is greater when administered by health professionals, particularly physiotherapists [5]. Despite the many benefits of physical activity and its positive implications for healthy ageing, adherence is still low [6]. Factors preventing participation include poor transportation, time restrictions and other accessibility issues [2].

Remote exercise has gained popularity, offering an accessible alternative to traditional programmes, potentially improving adherence [6, 7]. A review of 26 studies with over 5000 participants examined various remote interventions, including mobile apps, websites and videoconferences. Results showed reduced fall rates and improvements in strength, functional capacity and health-related quality of life (QOL) [8]. Another meta-analysis found that remotely supervised exercise, where healthcare providers offer feedback and encouragement via telephone, improved performance, adherence, fitness and reduced falls [9].

Videoconferencing facilitates real-time communication with instructors and fellow participants, effectively alleviating loneliness and providing an accessible alternative for those unable to attend traditional in-person exercise programmes [10]. Moreover, it provides older adults with the opportunity to engage with healthcare professionals when face-to-face interactions are not feasible [11]. Additionally, technology and the internet have become an integral component of the instrumental activities of daily living (IADL) for older adults [12].

A recent systematic review examined the effects of remote exercise for older adults via videoconferencing [13]. However, we believe a more comprehensive review is warranted due to several limitations in that study: the search was restricted to PubMed, thus, only nine studies met the eligibility criteria, and these included adults under 65 years of age. Additionally, the search was conducted in 2021, which predates many remote studies conducted during and after the COVID-19 pandemic [14–32]. Therefore, the aim of our systematic review is to explore recent evidence on the feasibility and effectiveness of remotely supervised videoconferencing exercise therapy for adults over 65.

Methods

Protocol and registration

This study was designed as a systematic review with meta-analysis, following the guidelines of the Cochrane Handbook [33] and is reported according to PRISMA guidelines [34]. The protocol was prospectively registered in PROSPERO on 10 October 2023: CRD42023467045: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023467045).

Selection criteria

This review included studies examining feasibility and effectiveness of videoconferencing exercise for older adults. Eligible studies included any type of published and unpublished randomised controlled trials (RCTs) (parallel, crossover, non-inferiority, etc.) or secondary analyses of RCTs in English only. Study protocols were excluded.

Participants

Community-dwelling older adults aged 65 years and over. Studies on hospital inpatients, institutionalised patients and long-term care facility residents were excluded.

Types of interventions

Studies were considered eligible if the remote online exercise was delivered via videoconference with real-time interaction with human professionals and participants were engaged in the exercise programme, either individually or in a group, from their home or a community centre.

Comparisons

Eligible studies included comparisons with non-videoconference exercise, such as face-to-face sessions, minimally supervised exercise without video, remote technology without real-time interaction, or non-intervention control groups.

Outcome measures

Outcome measures were divided into feasibility and effectiveness of videoconferencing exercise for older adults. Feasibility was evaluated based on programme adherence and safety. Programme adherence was measured through completion rates (the percentage of participants who completed the training period) and attendance records (calculated as the percentage of attended sessions) [35]. Safety was measured by the incidence of adverse events. Effectiveness measures were both objective and self-reported. Due to the large number of reported outcome measures and measurement tools, related to Effectiveness, we categorized those with sufficient data into four groups for the meta-analysis: physical function (balance, walking capacity and walking speed), muscle strength, QOL and emotional status (fear of falling, anxiety, loneliness, psychological resilience, depression and stress level). When studies reported multiple measures within a category, we applied a hierarchical selection process. For physical function, we prioritised the Short Physical Performance Battery (SPPB) as the primary measurement tool, followed by the Timed Up and Go (TUG), Sit-to-Stand (STS), Romberg and finally, the 6-minute walk test. For muscle strength, we prioritised the arm curl, grip strength, muscle power and quadriceps strength. For emotional status, we prioritised anxiety and depression as the primary measures, followed by loneliness, fear of falling and resilience when multiple measures were reported. Falls and fall-related outcomes (e.g. fall rate, fear of falling) were pre-specified in the review protocol (registered in PROSPERO), but due to inconsistent reporting across studies, only fear of falling could be meaningfully synthesised and is therefore reported under emotional status. The full list of outcomes, measurement tools and categorisation is listed in Appendix A.

Information sources and search strategy

Nine databases were searched on 24 November 2023, and was subsequently updated on 17 February 2025: PubMed (National Library of Medicine), CINAHL- Cumulative Index to Nursing and Allied Health Literature (EBSCO), AgeLine (EBSCO), PEDro- Physiotherapy Evidence Database, Scopus (Elsevier), Embase- Excerpta Medica Database (Elsevier), Web of Science (Clarivate), Cochrane CENTRAL- Cochrane Central Register of Controlled Trials (Wiley), and ProQuest Dissertations & Theses Global. The search strategy, developed with an expert librarian, included keywords and database-specific index terms for older adults, videoconferencing, exercise and RCTs. Detailed search strategies are in Appendix B. Manual search was conducted after full-text screening and continued until March 20, 2025. Citation tracking of eligible studies was performed via Scopus, and two relevant journals (Age and Ageing, BMC Geriatrics) were browsed for the past eleven years.

Data collection and study selection

Identified studies were exported to Covidence [36]. The Cochrane RCT Classifier, with 99% accuracy in identifying non-RCTs, was used for initial filtering [37]. Two authors (ATS and RF/ALS) independently screened titles/abstracts and full texts based on selection criteria. Disagreements were resolved through discussion or consultation with a third author (HSB). Authors of unavailable full-text articles or relevant conference materials were contacted via email, with a follow-up after 2 weeks. Studies were excluded if full data remained unobtainable.

Assessment of risk of bias in included studies

Risk of bias was assessed using the Cochrane Collaboration’s tool, evaluating six domains: selection, performance, detection, attrition, reporting and other biases [33]. Two authors (ATS and RF/ALS) independently conducted the assessment, with conflicts resolved through discussion or consultation with a third author (HSB). Studies were classified as follows: low risk—no high or unclear risk ratings in any domain (excluding performance bias due to the inability to blind exercise delivery); moderate risk—one or two domains with high or unclear risk (excluding performance bias); and high risk—three or more domains with high or unclear risk (excluding performance bias). Primary conclusions were drawn only from studies with low to moderate risk of bias.

Data extraction and synthesis

Data extraction was independently performed by two authors (ATS and RF/ALS), with disagreements resolved through discussion. Extracted data included study details, intervention characteristics, outcome measures and study conclusions. A narrative synthesis summarised findings, focusing on aggregating evidence and identifying generalizable lessons.

For meta-analysis, studies were categorized by control group: those receiving exercise intervention (face-to-face, pre-recorded or minimally supervised) and those without exercise intervention (no intervention, usual care or exercise recommendations). A random effects model in RevMan software was used. Forest plots visualised differences between videoconference exercise and control groups. Standardised mean differences (SMDs) calculated effect sizes due to diverse rating scales. Heterogeneity was assessed using I² values: low (<50%), moderate (50%–75%) and high (>75%) [38]. A sensitivity analysis was conducted to assess the impact of excluding the high-risk of bias studies.

Results

Study selection

A total of 41 659 studies were retrieved from nine databases (39 941 studies), and through citations and manual searches (1718 studies). After excluding 15 522 duplicates and 9868 studies deemed ineligible by automated tools in Covidence [37], 16 269 studies underwent initial screening based on their titles and abstracts. Out of 304 fully reviewed studies, 275 were excluded mainly due to population discrepancies, resulting in 29 papers describing 28 studies, as one study was included in two papers presenting primary and secondary analyses [20, 21]. The study selection process is illustrated in a PRISMA flow chart (Figure 1).

Figure 1.

Study flow diagram.

Risk of bias

All studies resulted in a high risk of bias in the blinding of participants and personnel criterion, which is inherent to exercise intervention research, where concealing the intervention arm from participants and instructors is impossible. Excluding this criterion, 7 out of 28 studies demonstrated low risk of bias across all other criteria [21, 24, 28, 29, 39–41]. Nearly half of the studies were rated as high risk for incomplete outcome data. Figure 2 presents a risk of bias assessment across various domains. Assessments for individual studies are presented in the forest plots and in Appendix C, while the justification for risk of bias judgements is detailed in Appendix D.

Figure 2.

Risk of bias criteria by frequency for the collected studies

Study characteristics

Table 1 summarises the characteristics of the 28 included studies while Figure 3 shows their geographical distribution. Sample sizes varied from 15 to 283 home-dwelling older adults, totalling 2086 participants. Publication dates range from 2010 to 2025, with a notable increase in the past three years; 82% (23 out of 28) of the studies were published between 2021 and 2025. Notably, 17 studies are COVID-19 related.

Table 1.

Study characteristics

Author, Year, Country	N	Study population	Type of exercise and instructor	Group/ Individualised	Mode of remote delivery	Frequency & duration of intervention	Control
Akbuga Koc et al. 2024, Turkey [26]	60	Older adults diagnosed with mild to moderate Alzheimer’s disease, with Montreal Cognitive Assessment (MoCA) 13–25. Mean age 77.8 ± 6.7 years, 67% women.	Range of motion, stretching, strengthening, balance, flexibility and aerobic exercises, delivered by physiotherapist.	Individualised	Zoom and WhatsApp	60 minutes, twice a week, over 12 weeks.	Control 1-no intervention Control 2- physical exercise and cognitive stimulation via videoconference.
An et al. 2021, Korea [42]	60	Older adults scheduled bilateral Total knee arthroplasty. Mean age 70.5 ± 2.7 years, 100% women.	Preoperative telerehabilitation programme- lower limb strengthening and ROM exercises: mobility, flexibility, strength & balance, delivered by physiotherapist.	Individualised	Unspecified	30-minute sessions, 2 times a day, 5 days a week, over 3 weeks.	Control 1- self-individualised unsupervised exercise with daily motivational and educational phone call Control 2- usual care- daily exercise recommendation.
Baklouti et al. 2023, Tunisia [15]	200	Community dwelling older adults. Mean age N/A. For the 160 participants who completed the study, 75% aged 65–75 years, and 25% aged 76–85 years, 39% women.	Hatha yoga, delivered by a certified Hatha yoga instructor.	Group	Google meet	80-minute sessions, twice a week, over 8 weeks.	No intervention.
Barnes et al. 2024, USA [27]	97	People with cognitive impairment. Mean age 76.4 ± 10.6. 43% women.	Functional movement exercises, including seated, standing and dyadic exercises performed by participant-caregiver pairs, focusing on balance, transitions and daily function, delivered by trained ‘Moving Together’ instructors.	Group	Unspecified	60 minutes, twice a week, over 12 weeks.	No intervention.
Beauchamp et al. 2021 Canada [28]	241	Low active older adults. Mean age 73.0 ± 5.4 years, 76% women.	Moderate intensity exercises, including strength, flexibility, balance and aerobic components, delivered by older adult exercise instructors that had considerable experience delivering older adult exercise classes.	Group	Zoom	50–60 minutes, offered daily, over 12 weeks. Participants were encouraged to attend at least three classes each week.	Control 1-no intervention Control 2- Personalised pre-recorded classes.
Bernocchi et al. 2018, Italy [39]	283	Older adults, with one or more chronic diseases (cardiac, respiratory, neuromuscular or neurologic) who have been discharged home from rehabilitation and are at medium/high risk of fall. Mean age 79.0 ± 6.6 years, 59% women.	Fall prevention exercises: based on the Otago Exercise Programme, delivered by physiotherapist.	Individualised	Unspecified	Twice a month (at least), over 6 months * An exercise booklet for self-exercise + phone call from physiotherapist every 2 weeks, with home visit if needed. Monitoring exercise sessions via videoconference.	Usual care- written information on fall risk factors, recommendation of exercises and walking twice a week.
Chan et al. 2024, China [29]	91	Community-dwelling older adults at risk of falls. Mean age 76.1 ± 6.5 years, 88% women.	Fall prevention exercises: based on the Otago Exercise Programme, delivered by a fitness instructor.	Group	Zoom	45 minutes, 3 days a week for 12 weeks.	Face-to-face exercise.
Coletta et al. 2024, Canada [30]	32	Sedentary community-dwelling older adults. Mean age 70.9 ± 4.5 years, 78% women.	Strengthening, balance, and aerobic exercises, delivered by a physiotherapist and kinesiologists.	Group	Zoom	60 minutes, 3 days a week for 8 weeks.	No intervention.
de Araújo Silva et al. 2023 & Maranhão et al. 2024, Brazil [20, 21]	38	Community-dwelling older adults who did not experience structured physical activity over the previous six months. Mean age 68.0 ± 6.5 years, 81% women.	Exercises primarily targeted muscle strength: weight-bearing exercises, delivered by certified professionals in exercise sciences.	Group	Google meet	25-minute sessions, 3 times a week, over 12 weeks.	Minimally supervised group: YouTube channel monitored through weekly text messages.
Elmas et al. 2024, Turkey [25]	30	Healthy older adults. Mean age 68.8 ± 3.6 years, 43% women.	Calisthenic exercises delivered by physiotherapist.	Unclear	WhatsApp	25 to 45 minutes (exercise duration increased with the number of repetitions), 3 days a week for 6 weeks.	Prerecorded videos of same calisthenic exercises with weekly phone call.
Giovannini et al. 2022, Italy [16]	15	Frail older adults, diagnosed with stable heart failure, able to ambulate safely. Mean age 81.3 ± 6.6 years, 53% women.	breathing exercises, aerobic and strengthening exercises, delivered by physiotherapist.	Group	Unspecified	45–60 minute sessions, twice a week, over 4 weeks, interspersed with a further 4 weeks of pause. * In addition, audio-video exercises were sent for self-exercise with supervision of the caregivers every 2 weeks.	Control 1-same programme, face-to-face Control 2- No intervention.
Granet et al. 2023, Canada [17]	83	Inactive but not frail, community-dwelling older adults,, walking independently without walking aids. Mean age 70.1 ± 5.2 years, 70% women.	Exercises that aimed to improve muscle function, flexibility and cardiovascular capacity, delivered by certified exercise instructor (kinesiologist).	Group	Zoom	55-minute sessions, 3 times a week, over 12 weeks.	Prerecorded video for individualised exercise on website platform.
Hong et al. 2017, Korea [43]	26	Community-dwelling older adults who did not experience regular exercise over the previous six months. Mean age 81.8 ± 5.0 years, 57% women.	Resistance exercise 3 sets, 6–8 repetitions at 75%–80% of 1 RM, delivered by trained and experienced exercise instructor.	Individualised	Skype	20–40 minute sessions, 3 times a week, over 12 weeks.	No intervention. Nutrition and exercise education was provided for all the participants 1 every 4 weeks.
Hong et al. 2018, Korea [44]	30	Community-dwelling older women, at increased risk of falls, who did not experience regular exercise over the previous six months. Mean age for the 23 participants who completed the study 80.0 ± 5.3 years, 100% women.	Resistance and balance exercises, delivered by trained and experienced exercise physiologist.	Individualised	Web Real-Time Communication (WebRTC) technology, self-developed	20–40 minute sessions, 3 times a week, over 12 weeks.	No intervention. Nutrition and exercise education was provided for all the participants 1 every 4 weeks.
Hong et al. 2022, Korea [18]	31	Obese (> 30% body fat) older women. Mean age for the 29 participants who completed the study 80.0 ± 3.5 years, 100% women.	Stretching, Aerobic and resistance exercises, delivered by experienced exercise physiologist.	Individualised	Smartphone mirroring web real-time communication, self-developed	20–40 minute sessions, 3 times a week, over 12 weeks.	Face-to-face group exercise.
Kuldavletova 2021, France [45]	28	Community-dwelling older adult without vestibular, neurological, oculomotor or visual impairments and/or auditory pathologies and does not exercise regularly (more than 1.5 hours a week). Mean age 72 ± 3.8 years, gender distribution was not reported.	Aerobic, resistance exercises and balance exercises. The aerobic part of the training included exercises with ergo-cycles and steppers, delivered by specialists in physical activity.	Group	System developed by the European MOTION project	60 minute sessions, twice a week, over 16 weeks.	Face-to-face individualised exercise.
Langeard et al. 2022, France [19]	43	Independently living community-dwelling older adults. Mean age for the 41 participants who completed the study 72.5 ± 4.0 years, 66% women.	Muscle function, functional activities and aerobic capacity, delivered by professional physical trainers.	Group	System developed by the European MOTION project	60 minute sessions, twice a week, over 16 weeks.	Control 1- No intervention Control 2- same programme, face-to-face, individualised exercise.
Lee 2025, South Korea [31]	94	Pre-frail older adults who experienced a fall within the last 6 months or displayed impaired gait and balance. Mean age for the 90 participants who completed the study 84.9 ± 2.9 years, 47% women.	Balance, muscle strengthening, and gait exercises aimed at fall prevention, delivered by a physiotherapist.	Group	Zoom	50 minute sessions, twice a week, over 8 weeks.	Control 1- received an exercise booklet and a single educational session on fall prevention and physical activity, without structured exercise. Control 2- same programme, face-to-face.
Mascret et al. 2023, France [22]	40	Independently living community-dwelling older adults, walking independently without walking aids. Mean age 70.7 ± 4.3 years, 52% women.	Square-Stepping Exercise: a complex cognitive-motor task, delivered by qualified physical activity teacher.	Individualised	A mobile telepresence robot	One total session.	One session face-to-face.
Menengic et al. 2022, Turkey [32]	20	People with Alzheimer’s disease. Mean age 79.2 ± 5.8, 70% women.	Simple chair-based exercises with cognitive tasks (motor-cognitive dual-task), delivered by physiotherapists.	Individualised	Zoom	4–5 exercise sessions a week for 15–40 minutes per session. Total of 25 sessions over 6 weeks.	No intervention.
Okpara et al. 2023, Canada [23]	70	Mild to moderate frail older adults, walking independently with or without walking aids. Mean age 77.3 ± 6.5 years, 77% women.	Strength and balance exercises, delivered by physiotherapist.	Group	Zoom	60-minute session, twice a week, over 12 weeks * Additional weekly hour of self-directed home-based exercises to reach the 3-hour minimum for fall prevention.	Socialisation arm-once-week phone call to mitigate the impact of social isolation.
Pandya 2021, India [46]	140	Community-dwelling older woman who was interested in participating in Yoga classes. Mean age 67.7 ± 2.4 years, 100% women.	Yoga, delivered by certified yoga experts.	Group	Graphic user interface software- unspecified	30-minute synchronous sessions, every day, over 30 days *Additional asynchronous self-practice recommended daily.	No intervention.
Terkes et al. 2023, Turkey [24]	70	Older adults, diagnosed with Type 2 diabetes mellitus for at least 6 months who received oral antidiabetic and/or insulin. Mean age for the 65 participants who completed the study 68.8 ± 6.6 years, 48% women.	Mobility improvement, aerobic, muscle strengthening, balance and flexibility exercises, delivered by physiotherapist.	Group	Zoom	Approximately 60-minute session, 3 times a week, over 6 weeks.	Unsupervised-exercise training group + distance education.
Toledano-Shubi et al. 2024, Israel [40]	81	community-dwelling adults, who walked independently and were at risk of falling. Mean age 73.2 ± 4.0 years, 75% women.	Fall prevention exercises: based on the Otago Exercise Programme, delivered by physiotherapist.	Group	Zoom	45–60-minute sessions, twice a week, over 12 weeks.	Face-to-face exercise.
Tsai et al. 2017, Australia [41]	37	Patients with stable COPD who ambulate independently without a walking frame, weigh <150 kilogrammes, and did not participate in any exercise programme in the last 12 months. Mean age for the 36 participants who completed the study 74.0 ± 8.0 years, 50% women.	Lower limb cycling, walking training and Lower limb strengthening exercises, delivered by physiotherapist.	Group	videoconferencing software (VSee, CA, USA, http://vsee.com)	3 times a week, over 8 weeks.	No intervention.
Wu et al. 2010, USA [47]	64	Community-dwelling older adults, who were at risk for falls walking independently without walking aids. Mean age 75.4 ± 7.0 years, 90% women.	Tai Chi, delivered by a certified Tai Chi Chuan instructor.	Group	DocBox custom-made videoconferencing system	60-minute sessions, 3 times a week, over 15 weeks.	Control 1- community centre–group-based exercise Control 2- home video–individualised based exercise.
Yerlikaya et al. 2021, Cyprus [48]	52	community dwelling older adults, who were at risk for falls, walking independently with or without walking aids. Mean age for the 50 participants who completed the study 72.5 ± 7.2 years, 70% women.	simple balance and strengthening exercises, delivered by physiotherapist.	Group	WhatsApp and Google Meet	40-minute sessions, 3 times a week, over 8 weeks.	Control 1- individualised unsupervised home exercise Control 2- no intervention.
Zengin Alpozgen et al. 2022, Turkey [14]	30	Community-dwelling older adults who were in social isolation due to the COVID-19 pandemic. Mean age 68.2 ± 4.8 years, 57% women.	Strength, balance, and stretching exercises, delivered by physiotherapist.	Unclear	Skype	40–45-minute sessions, 3 times a week, over 6 weeks.	No intervention.

ROM, range of motion. RM, repetition maximum. COPD, Chronic Obstructive Pulmonary Disease. Low risk of bias; Moderate risk of bias; High risk of bias. Studies were classified as low risk of bias if only ‘blinding of participants and personnel’ was rated high risk. Moderate risk included 1–2 additional unclear/high-risk domains, while high risk included three or more additional unclear/high-risk domains.

Figure 3.

Characteristics of included studies. Four pie charts illustrating: (A) Geographical distribution of studies by country; (B) Types of interventions; (C) Group vs. individual exercise format; (D) Comparison groups, including studies with one or two control groups and types of control conditions used. Each section is labelled with the category and number of studies it represents.

Participants

The age of participants in the included studies varied from 65 to 98 years old, with a mean age ranging from 67.7 ± 2.4 to 84.9 ± 2.9 years. Women comprised the majority of participants in 17 studies, with proportions ranging from 52.5% to 84.4% [14, 16, 17, 19, 21–23, 26, 28–30, 32, 39, 40, 43, 47, 48]. Four studies exclusively enrolled women [18, 42, 44, 46], and one study did not report participant gender distribution [45].

Interventions and comparison

A variety of exercise programmes, in group or individual formats, with diverse control groups for comparison, were considered, all outlined in Figure 3 and detailed in Table 1. Interventions included fall prevention (such as strength and balance exercises) [14, 23, 29, 31, 39, 40, 44, 48], yoga [15, 46], specific health condition exercises [such as breathing exercises, aerobic and strengthening exercises for participants with heart failure [16], and aerobic and strengthening exercises for participants with Chronic Obstructive Pulmonary Disease (COPD) [41]], and general physical activity. Control groups ranged from no exercise intervention to face-to-face with minimal supervised exercise.

Feasibility of remote exercise via videoconference

Attendance data was reported in 11 studies, with rates ranging from 60.7% to 100% [17, 19, 21, 23, 28–30, 39–41, 47]. In methodologically sound studies (low and moderate risk of bias), the average attendance rate was 79.7% [21, 23, 28–30, 39–41]. The average completion rate was 92.8% [14, 21, 23–32, 39–42, 48]. Attendance and completion rates for all studies are presented in Appendix E. Adverse events were reported as an outcome in nine studies [17, 23, 27, 29, 30, 32, 39–41], of these, seven studies explicitly stated that no adverse event occurred, while two studies reported falls as adverse events, with incidence comparable to the control groups [23, 40]. The remaining 19 studies did not address adverse events.

Effectiveness of remote exercise via videoconference

One study [16] assessing all outcomes and another assessing physical function and emotional status [23] were excluded due to insufficient data despite author contact attempts. We present forest plots for each category, comparing videoconference exercises against both alternative exercise methods and non-exercise controls.

Physical function

Seventeen studies assessed comparable measures of physical function [14, 17, 18, 25, 26, 29, 31, 32, 39–45, 47, 48]. Ten of these studies compared physical function outcomes between 241 videoconference participants and 244 participants using alternative exercise methods (face-to-face, recorded and home exercise) [17, 18, 25, 29, 31, 40, 42, 45, 47, 48]. The pooled meta-analysis showed a small statistically significant difference in physical function outcomes favouring the videoconference group compared with alternative exercise methods (SMD = 0.25; 95% CI: 0.03–0.47; P = .03, I² = 29%). However, when restricting the analysis to high-quality studies, the difference between the two methods was no longer significant (SMD = 0.29; 95% CI: -0.06 to 0.64; P = .11, I² = 58%). Ten studies compared 273 videoconference participants to 274 control participants who received no exercise intervention [14, 26, 31, 32, 39, 41–44, 48]. The results showed a significant improvement in physical function favouring the videoconference group (SMD = 0.71; 95% CI: 0.39–1.03; P < .0001, I² = 61%). The sensitivity analysis yielded consistent results (SMD = 0.82; 95% CI: 0.46–1.19; P < .00001, I² = 65%). These findings are illustrated in the forest plot in Figure 4A, while the forest plots corresponding to the sensitivity analysis are provided in Appendix F.

Figure 4.

Forest plots comparing videconference exercise to alternative exercise methods (control training) and no exercise controls (control no training); 4A. Physical function.TUG: Timed Up and Go test, SPPB: Short Physical Performance Battery; 4B. Muscle strength. Nm: Newton-metres. Kg: kilogramme; 4C. Emotional status. DASS-21: Depression Anxiety Stress Scales—21 Items, FES: Falls Efficacy Scale, K10: Kessler Psychological Distress Scale, FES-K: Korean Falls Efficacy Scale, GAS-10: Geriatric Anxiety Scale—10 item version, BRS: The Brief Resilience Scale, M-FES: Modified Falls Efficacy Scale, HADS: Hospital Anxiety and Depression Scale, ABC: Activities-specific Balance Confidence scale, TAI: Trait Anxiety Inventory, LSE: Loneliness scale for the elderly; 4D. Quality of life. EQ-5D: Euro Quality of Life-5 Dimensions, WHOQOL: World Health Organization Quality of Life, WHO-5: WHO-Five Well-Being Index, CASP-19: Control, Autonomy, Self-Realisation and Pleasure, SF12: Short Form Health Survey consists of 12 questions, SF36: Short Form Health Survey consists of 36 questions, WHL: World Health Organization Quality of Life Instrument-Older Adults, NHP: Nottingham Health Profile. Risk of Bias legend: A-Randon sequence generation, B-Allocation concealment, C-Blinding of participant and personnel, D-Blinding of outcome assessment, E-Incomplete outcome data, F-Selective reporting, G-Other bias.

Muscle strength

Eleven studies assessed comparable measures of muscle strength [14, 17–19, 25, 26, 29, 42–45]. Seven of these studies compared muscle strength outcomes between 140 videoconference participants and 157 participants using alternative exercise methods [17–19, 25, 29, 42, 45]. The pooled meta-analysis showed no statistically significant difference in muscle strength change between these groups (SMD = 0.05;95% CI: -0.40 to 0.51; P = .82, I² = 72%). The sensitivity analysis, restricting to high-quality studies, presented consistent results (SMD = 0.01; 95% CI: -0.92 to 0.94; P = .98). Six studies compared 87 videoconference participants to 91 control participants who received no exercise intervention [14, 19, 26, 42–44]. The results showed a significant improvement in muscle strength, favouring the videoconference group (SMD = 0.52; 95% CI: 0.11–0.93; P = .01, I² = 44%). High-quality studies demonstrated comparable effect (SMD = 0.86; 95% CI: 0.46–1.26; P < .0001, I² = 0%). These findings are illustrated in the forest plot in Figure 4B, and the forest plots for the sensitivity analysis are provided in Appendix F.

Emotional status

Nineteen studies assessed emotional status [14, 15, 17, 21, 24, 26–32, 39–41, 44, 46–48]. Nine of these studies compared 314 videoconference participants to 314 participants using alternative exercise methods [17, 21, 24, 28, 29, 31, 40, 47, 48]. The analysis revealed no statistically significant difference in emotional status change between these groups (SMD = 0.02; 95% CI: -0.15 to 0.20; P = .80, I² = 16%). The sensitivity analysis, restricting to high-quality studies, presented consistent results (SMD = -0.01; 95% CI: -0.22 to 0.20; P = .90, I² = 29%). Thirteen studies compared 524 videoconference participants to 530 control participants who received no exercise intervention [14, 15, 26–28, 30–32, 39, 41, 44, 46, 48]. The results showed a significantly higher improvement in emotional status favouring the videoconference group (SMD = 0.93; 95% CI: 0.45–1.40; P = .0001, I² = 92%), with consistent results in the sensitivity analysis (SMD = 0.72; 95% CI: 0.35–1.09; P = .0001, I² = 80%). These findings are illustrated in the forest plot in Figure 4C, forest plots for the sensitivity analysis are provided in Appendix F.

Quality of life

Eleven studies evaluated participants’ QOL [14, 17, 21, 24, 26, 27, 39, 40, 46–48]. Seven of these studies compared 179 videoconference participants to 176 participants using alternative exercise methods [17, 21, 24, 26, 40, 47, 48]. The analysis, revealed no statistically significant difference in QOL improvement between these groups (SMD = 0.07; 95% CI: -0.14 to 0.28; P = .53, I² = 0%) with consistent results in sensitivity analysis (SMD = 0.08; 95% CI: -0.21 to 0.37; P = .58, I² = 24%). Five studies compared 273 videoconference participants to 253 control participants who received no exercise intervention [14, 27, 39, 46, 48]. The results showed a significantly higher improvement in QOL index favouring the videoconference group (SMD = 0.75; 95% CI: 0.34–1.16; P = .0003, I² = 76%), with consistent results in sensitivity analysis (SMD = 0.61; 95% CI: 0.19–1.03; P = .004, I² = 67%). These findings are illustrated in the forest plot in Figure 4D, forest plots for the sensitivity analysis, are provided in Appendix F.

As presented in Figure 4, heterogeneity was low (I² = 0%–29%) for physical function, emotional status and QOL and moderate (I² = 72%) for muscle strength, when comparing videoconference exercise to alternative exercise methods (control training). However, when videoconference exercise was compared to no-exercise control group (control no training), heterogeneity varied from moderate to high across the four outcome categories: I² = 61% for physical function, 44% for muscle strength, 92% for emotional status and 76% for QOL.

Discussion

This systematic review and meta-analysis investigated the feasibility and effectiveness of videoconference-based exercise for community-dwelling older adults. Due to significant methodological concerns in many included studies, we base our primary conclusions only on studies with low to moderate risk of bias [14, 21–32, 39–42, 48]. Our findings suggest that videoconference exercise is feasible, with an average attendance rate of 79.7% and a completion rate of 92.8%. Attendance rate is in line with the 85% reported in another review [13]. The rate of adverse events in videoconference-based exercise was comparable to that of in-person exercise. This finding aligns with a review on e-health exercise aimed at improving balance in community-dwelling older adults, which reported no major adverse events [7]. To ensure the safe implementation of remote exercise programmes, future research should clearly define appropriate participant criteria for older adults. Key factors to consider include cognitive status, balance capacity and the severity of medical conditions, helping to establish thresholds beyond which unsupervised remote training may be unsafe, ineffective or impractical. Such research would support clinicians in making evidence-based decisions regarding whether older adults can safely engage in remote exercise or require more intensive in-person supervision.

Regarding effectiveness, our analysis faced challenges due to the considerable variability in outcome measures across the included studies. To address this, we categorized outcomes and utilised SMD indices for analysis. Notably, the control groups across studies varied, with some employing alternative exercise methods and others offering no exercise intervention. Consequently, the analysis compared videoconferencing exercise to both types of control groups for each outcome category.

The sensitivity analysis indicated that study quality had an influence on the observed outcomes. Initially, the full analysis suggested an unexpected, small yet significant, advantage of videoconference intervention over alternative exercise methods (control training) in physical function. However, when high-bias studies were excluded, this difference was no longer significant, showing comparable effectiveness between the two approaches. Across other parameters—muscle strength, emotional status and QOL- videoconference intervention remained as effective as control training. Our findings align with those of other reviews and meta-analyses in the field. A recent review [13] also concluded that videoconference exercise is at least as effective as face-to-face exercise methods in improving physical function among older adults. However, future research with larger, well-designed RCTs is needed to strengthen the evidence base, as current studies are limited in number and exhibit substantial variability in both the type of exercise interventions delivered via videoconference (e.g. yoga, fall prevention and resistance training) and the nature of the control groups.

The limitations of existing research are further illustrated by our experience with fall prevention studies. Although our review identified eight studies related to fall prevention [29, 31, 39, 40, 44, 45, 47, 48], we encountered significant challenges in analysing fall-related outcomes, particularly fall efficacy and fall rate. The scarcity of data led us to combine fall-related confidence measures with other emotional assessments. In the context of digital health interventions, two recent reviews presented conflicting results on fall efficacy [8, 9]. These contradictory findings, coupled with limited data on videoconference exercise, highlight the need for more comprehensive reporting of fall-related outcomes in remote intervention studies. Fall rate data was limited in our review, with only 4 out of 8 fall-related studies reporting fall outcomes [29, 39, 40, 47], and analysis was hindered by control group variability. This data scarcity aligns with a review on eHealth balance exercise interventions [7], where fall effects could not be examined despite all studies addressing balance. A review of digital home exercises for older adults [8] reported effectiveness in improving balance and reducing fall rates. Similarly, a broader review of e-interventions [9] reported 11 out of 31 studies addressing falls, also showing positive outcomes.

Exercising together with telehealth support emerged as one of the most effective techniques for improving balance [9]. The researchers attributed this to the communication, support and motivation provided by instructors. We believe that these interactions also positively impacted emotional status and QOL, as shown in our review of videoconference exercise. Similarly, a review of digital exercise found improvements in QOL, with 17 out of 26 studies involving instructor interaction, and 8 of those included interaction in every exercise session [8].

The findings of this review highlight the importance of integrating videoconferencing into treatment plans to make exercise more accessible, aligning with the WHO’s recommendations for optimal ageing. Remote exercise can reduce time and travel costs for adults, provide care to multiple individuals simultaneously, and make efficient use of professional resources. To optimise programme effectiveness, future research should explore hybrid training methods that combine videoconference-based exercise with face-to-face sessions. This approach could enhance adherence to WHO-recommended exercise doses for healthy ageing while maintaining the convenience of remote delivery. To build a robust evidence base, more RCTs with rigorous methodologies are needed. These should employ consistent measurement tools like TUG and STS, monitor and report falls and implement intervention programmes at the recommended doses per the WHO [3]. Additionally, fall prevention exercise studies should adhere to ProFaNE’s reporting guidelines to create a uniform database [49], facilitating meta-analyses and overcoming challenges in multi-participant RCTs.

Strengths and limitations

This meta-analysis has several notable strengths, including a rigorous search strategy coupled with reporting in strict adherence to PRISMA guidelines. This methodological rigour enhances the reliability and reproducibility of our findings. Our review offers valuable insights for the responsible implementation of remote exercise to promote healthy ageing, highlighting an accessible exercise modality that leverages the benefits of group and interactive sessions. However, the study also faces limitations. The high heterogeneity in outcome measures and control groups across studies poses challenges in synthesising results and drawing definitive conclusions about videoconference exercise effectiveness. Additionally, the quality of included studies is a concern, with only 25% of included studies were at low risk of bias. Insufficient information in some trials led to the exclusion of potentially relevant studies due to missing data, despite attempts to contact authors. Finally, the typically small sample sizes in individual studies underscore the importance of this meta-analysis in validating findings, though this limitation should be considered when interpreting results.

Conclusion

This systematic review and meta-analysis demonstrate that videoconference exercise is feasible and effective for improving various aspects of health and well-being. Based solely on methodologically robust studies, videoconference exercise showed significant advantages in physical function, emotional well-being, QOL and muscle strength when compared to no intervention. While the effect sizes for these improvements were generally small to moderate, they highlight the potential of videoconferencing to deliver health interventions. Furthermore, videoconference exercise appeared comparable to alternative exercise methods in terms of effectiveness across all measured outcomes. This suggests that videoconferencing can be a valuable alternative when other exercise options are not feasible or accessible. Its flexibility is advantageous for individuals who cannot participate in face-to-face interventions.

Declaration of Conflicts of Interest

None declared.

Declaration of Sources of Funding

This work was supported by the Data Science Research Center at the University of Haifa, grant numbers 100008981, 100009419. The funding sources played no role in the conduct of this study.