Exercise Interventions Delivered Through Telehealth to Improve Physical Functioning for Older Adults with Frailty, Cognitive, or Mobility Disability: A Systematic Review and Meta-Analysis

Introductions: This study assessed the effects of telehealth-delivered exercise interventions on physical functioning for older adults and explored implementation measures related to program delivery. Methods: We conducted a systematic review of studies investigating effects of exercise interventions delivered through telehealth in adults 60+ years of age with frailty, mobility, or cognitive disability on mobility, strength, balance, falls, and quality of life (QoL). Electronic databases (MEDLINE, CINAHL, SPORTSDiscus, and Physiotherapy Evidence Database) were searched from inception until May 2022. Evidence certainty was assessed with Grading of Recommendations, Assessment, Development, and Evaluation and meta-analysis summarized study effects. Results: A total of 11 studies were included, 5 randomized controlled trials, 2 pilot studies, and 4 feasibility studies. The overall certainty of evidence was rated as “low” or “very low.” Pooled between-group differences were not statistically significant, but effect sizes suggested that telehealth produced a moderate improvement on mobility (n = 5 studies; standardized mean difference [SMD] = 0.63; 95% confidence interval [CI] = −0.25 to 1.51; p = 0.000, I2 = 86%) and strength (n = 4; SMD = 0.73; 95% CI = −0.10 to 1.56; p = 0.000, I2 = 84%), a small improvement on balance (n = 3; SMD = 0.40; 95% CI = −035 to 1.15; p = 0.012, I2 = 78%), and no effect on QoL. Analysis of implementation measures suggested telehealth to be feasible in this population, given high rates of acceptability and adherence with minimal safety concerns. Discussion: Telehealth may provide small to moderate benefits on a range of physical outcomes and appears to be well received in aged care populations.


O
lder adults, 60 years of age and older, are the most sedentary age group in our community. 1Aged care service recipients, who usually have frailty, and/or mobility or cognitive disability, are particularly inactive. 2Physical inactivity is associated with negative health outcomes such as increased risk of falls, frailty, mobility disability, and death. 35][6] The Australian 2021 Royal Commission into Safety and Quality in Aged Care identified falls and mobility decline as a serious problem in residential aged care, exacerbated by poor access to exercise health professionals such as physiotherapists. 7imilar issues have been identified in other countries. 8orld Health Organization has defined telehealth as the delivery of health care services, where patients and providers who are separated by distance use information communication technologies to diagnose and treat diseases and injuries. 9elehealth is a rapidly growing service delivery model that could enhance access and facilitate delivery of exercise programs to older adults receiving aged care services in their home or in residential care.Uptake of telehealth has been accelerated in response to the COVID-19 pandemic, especially in aged care settings. 10ittle is known about the effectiveness of telehealth use to deliver exercise programs in aged care settings or how best to implement telehealth exercise in this complex population.A recent rapid review investigating the use of websites and apps to assist older adults engage with balance and strength training found low to moderate evidence that the use of digital technology improved physical activity and reduced fall risk. 11However, this review excluded studies that targeted aged care populations.A discrete choice experiment in 2017 concluded that telehealth was acceptable among older adults receiving rehabilitation, but excluded aged care populations. 12his systematic review sought to investigate the use of telehealth exercise for older adults who are receiving aged care services or have frailty, mobility, or cognitive disability.This review aimed to summarize (1) the effects of exercise interventions delivered through telehealth on mobility, strength, balance, falls, and quality of life (QoL) and (2) implementation outcomes and determinants related to the delivery of telehealth exercise programs in this population.

Methods
Our systematic review with meta-analysis followed the methods described in the Cochrane Handbook for Systematic Reviews of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 13,14This review was registered through the PROSPERO international prospective register of systematic reviews on May 1, 2022 (CRD42022322469).

SEARCH STRATEGY
Optimized searches were completed through four electronic databases (MEDLINE, CINAHL, SPORTDiscus, and Physiotherapy Evidence Database [PEDro]) from inception to May 1, 2022.Four reviewers (I.M.R., C.Y.C., A.S., and M.B.) searched one database each.Keywords, MeSH, and other index terms were used to construct the search strategy (see online Supplementary Table S1 for examples of search terms).Articles sourced by hand searching were included.Articles were independently screened in two stages: screening of title and abstracts and screening of full-text articles by two pairs of reviewers (I.M.R./A.S. or C.Y.C./M.B.) using the eligibility criteria.Disagreements regarding the eligibility of studies were resolved through discussion.Conference abstracts and dissertations that reported data suitable for analysis were included.

ELIGIBILITY CRITERIA
Type of study.For our first aim, we included randomized controlled trials (RCTs), and for our second aim, we also included quasi-randomized, feasibility, and qualitative studies.
Participants.We included studies investigating participants with a mean age of 60+ years with frailty, mobility, or cognitive disability, or who were aged care service users.Any type of health condition was included.Studies involving participants being treated in hospital were excluded as these participants receive more support than older people receiving aged care services in their home or in residential aged care.
Interventions.Studies were included if they evaluated exercise interventions delivered through synchronous telehealth (virtual interactions between a participant and a health professional that occurs in real time) or asynchronous telehealth (sharing of data, educational materials, or online programs to assist a participant to exercise at a time of their choosing) that aimed to increase balance, strength, and/or physical functioning. 15We excluded studies involving no telehealth, that did not include participants with frailty, mobility, or cognitive disability or aged care service users, or involved only wearable technology.Interventions of any length and any follow-up period were included.
Comparator.For our meta-analysis, we included RCTs that compared telehealth to any comparator.
Outcomes.Our outcomes to address aim 1 included measures of mobility, balance, strength, falls, and QoL.Outcome data were extracted for baseline and post-intervention periods.Studies were included in meta-analysis if their data were

EXERCISE INTERVENTIONS DELIVERED THROUGH TELEHEALTH
presented as or could be converted into mean/standard deviation (SD) pre-intervention and post-intervention scores to facilitate quantitative pooling.To enable inclusion of as many studies as possible in the meta-analysis, we pooled results across multiple assessment tools for the same outcome.
For studies that reported results for more than one assessment tool for the same outcome, we selected one tool per outcome using a pre-defined order of priority. 16The order of priority was as follows: for mobility (Short Physical Performance Battery, Physical Performance Test, Timed Up and Go test, and walking speed); for balance (Berg Balance Scale, 4-Stage balance test, and timed step test); strength (only timed sit to stand was used); and QoL (no priority was required as there was only one tool used in each study).Falls were assessed separately using the risk of falling that is, number of adults who experienced one or more falls and rate of falls that is, falls per person-year.
Outcomes to address aim 2 included any implementationrelated outcome and determinant identified through an analysis of studies' results relating to the intervention's reach (proportion of participants who were successfully screened and consented to participate), 17 feasibility (proportion of participants who completed the follow-up assessment), 18 adherence (proportion of participants who completed the agreed number of planned intervention sessions), 19 acceptability (measure of satisfaction), dose (hours of exercise completed over study period), and safety (reporting of adverse events [AE] such as falls and pain directly related to the intervention), as well as barriers and facilitators determined using mixed methods. 20

DATA EXTRACTION
A data extraction sheet was developed, pilot-tested, and modified accordingly.For each study, two pairs of investigators (I.M.R./A.S. or C.Y.C./M.B.) extracted the data, and two investigators (R.D./W.S.K.) checked the data.Information extracted from each study comprised a description of participants, details of the intervention, and outcome measures (baseline and at first follow-up).Preintervention and post-intervention scores were used when available.Authors of the included studies were contacted by email if the study reports were incomplete, or data were missing.If the author did not reply, then the available data were used.For our meta-analysis, the pooled difference was calculated as a mean and 95% confidence intervals (CIs) for each outcome in the simple stratified analysis.

METHODOLOGICAL QUALITY ASSESSMENT FOR RCTS
Data pertaining to the risk of bias were extracted by two pairs of investigators (I.M.R./A.S. or C.Y.C./M.L.) and assessed using the PEDro scale. 21The PEDro scale evaluates 11 items: inclusion criteria and source, random allocation, concealed allocation, similarity at baseline, subject blinding, therapist blinding, assessor blinding, completeness of follow-up, intention-to-treat analysis, between-group statistical comparisons, and point measures and variability. 21Item 1 refers to external validity and does not contribute to the final score; thus, the final scores ranged from 0 to 10.

ASSESSMENT OF CERTAINTY OF THE EVIDENCE
The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) was used to assess the certainty of evidence for the primary outcomes of our metaanalyses for aim 1. 22 We used the GRADE system for all outcomes, which pooled results from three or more studies.We evaluated the quality of the body of evidence as ''High,'' ''Moderate,'' ''Low,'' or ''Very Low'' based on the presence or extent of four factors: study limitations, inconsistency of the effect, imprecision, and publication bias. 23

DATA ANALYSIS
Meta-analyses were completed using Stata Meta-Analysis software using the random-effects model for each outcome (mobility, strength, balance, and QoL). 24We gathered preintervention and post-intervention mean/SD and the sample size per group.We used the controlled trials that compared the telehealth intervention group with either usual care that included no active exercise (four trials) [25][26][27][28] or face-to-face inperson exercises (one trial). 29We calculated treatment effects using standardized mean differences (SMDs) (Hedges' g), standardized by post-score SD (or its estimate) with 95% CIs.SMD was calculated using the pre-mean and post-mean and SD or, when this was unavailable, we used the mean change score.Effect sizes were categorized as small (0.2-0.49), medium (0.5-0.79), or large (0.8 or greater). 30We visually inspected forest plots for evidence of heterogeneity with consideration of the I 2 and v 2 tests.For the implementation data, we collected median and range scores and conducted a thematic analysis of the authors' reported barriers and facilitators to implementing telehealth.

Results
After duplicates were removed, the electronic search retrieved 370 references.We completed full-text screening on 118 articles.We included 11 studies for the review, which included 5 RCTs, 4 feasibility studies, and 2 pilot studies.The five trials contributed to our meta-analysis (aim 1) and all the studies contributed to our implementation analysis (aim 2).Search results are presented in Figure 1.

TRIAL DESIGN AND PARTICIPANT CHARACTERISTICS
Participant characteristics are displayed in Table 1.A total of 546 subjects participated across all studies (302 in controlled studies and 244 in feasibility or pilot studies).Mean age was 77.4 -4.7 years and 65% of participants were female.Three studies recruited from residential aged care facilities 27,31,32 and one study recruited participants receiving aged care services in their home. 2836 In the controlled studies, two compared telehealth to usual care, 27,28 with one study compared to usual care plus fall education, 25 one study compared to seated stretching, 26 and one study compared to in-person balance training.29 The mean PEDro total score of the RCTs was good at 6.4.

EFFECTS OF INTERVENTIONS
The results of outcomes measured are summarized in Supplementary Table S2 with the forest plots and mean differences presented in Figures 2-5.The overall certainty of evidence for the effect of telehealth intervention on physical outcomes was rated as very low quality (Supplementary Table S3) and the pooled results for all outcomes did not reach statistical significance as all the CIs crossed the null line of effect.The pooled estimates (Hedges' g) of suggested effect size are as follows: a moderate improvement in mobility (5 studies; SMD = 0.63; 95% CI = -0.25 to 1.51; 302 participants, p = 0.000, I 2 = 86%, very low certainty), [25][26][27][28][29] a moderate improvement in strength (4 studies; SMD = 0.73; 95% CI = -0.10 to 1.56; 226 participants, p = 0.000, I 2 = 84%, very low certainty), [25][26][27][28] and a small improvement in balance   (3 studies; SMD = 0.40; 95% CI = -035 to 1.15; 199 participants, p = 0.012, I 2 = 78%, very low certainty). 25,26,29There was no indication of an effect of telehealth on QoL (3 studies; SMD = -0.09;8][29] We were unable to conduct a meta-analysis of fall outcomes as there were only two studies that could have their reported data translated into a fall rate ratio 27,29 and only one study was able to translate their fall data into risk ratio. 26

IMPLEMENTATION ANALYSIS
The results of the implementation analysis are summarized in Table 3.
Reach: Seven studies reported on the proportion of participants who were successfully screened and consented to participate, and their median reach was 55% (range = 9-82). 25,26,28,29,32,34,36easibility: All studies reported the proportion of participants who were included in the follow-up data collection with a median score of 87% (range = 67-100).
Facilitators: Three factors that supported high program adherence rates were reported by the authors in the ''Discussion'' section of the article.33]35 Participant appreciation of the conve-nience of telehealth exercise programs being delivered in their own home was reflected on by five authors. 28,29,31,32,36Positive effects of program flexibility on exercise adherence using asynchronous telehealth exercise programs as it allowed participants to choose the timing of when they exercised were observed by four authors. 25,27,32,33arriers: Authors' reports regarding the impact of the participant's technology hesitation and age on the outcomes of telehealth exercise programs were mixed.One study reported that high technology hesitation reduced recruitment rates. 31Two studies observed that higher levels of technology hesitation correlated with reduced telehealth satisfaction rates 26,35 and another study found that older participants had lower satisfaction rates. 35However, one study observed that technology hesitation or age was not related to the feasibility or acceptability of the telehealth intervention. 32   multiple sites is challenging and may inhibit the ability to deliver adequate exercise dose and intensity required to improve other physical outcomes. 31Brun et al. stated that asynchronous exercise programs did not enable sufficient exercise tailoring to enhance physical outcomes. 36

Discussion
This systematic review with meta-analysis assessed the effectiveness and implementation of exercise interventions delivered through telehealth for older adults, 60 years and older, who are receiving aged care services or have mobility, cognitive, or frailty disability.It is the first review of this approach in this population, to our knowledge.Pooled effects did not reach statistical significance outcomes, but suggested favorable effects of telehealth interventions to improve mobility, strength, and balance in older adults, which are likely to be clinically meaningful.Our analysis of implementation measures suggested that telehealth is feasible, as evidenced by high rates of acceptability and adherence with few safety concerns.
The utilization of telehealth to promote physical activity and improve physical functioning in aged care is emerging as an effective and acceptable mode of health care delivery. 37 2020 systematic review (n = 17 controlled studies) into community-dwelling older adults, 60 years of age and older, receiving a variety of health care interventions using synchronous telehealth found similarly high levels of feasibility, safety, and acceptability. 38Three of the included studies that focused on falls, exercise, or strength-based measures demonstrated significant improvements. 39- 41A scoping review in 2022 found that synchronous and asynchronous telehealth physiotherapy is safe, feasible, and acceptable to adults with complex comorbidities and has comparable effects to in-person care activity. 42Both reviews suggest that telehealth delivery increases access to exercise programs, especially for those who cannot travel to a treatment facility due to distance or disability. 38,42Hawley et al. concluded that telehealth physiotherapy may increase adherence to exercise by providing increased prompts reminding patients when and how to exercise. 42he positive effects emerging in our meta-analysis favoring telehealth-led exercise intervention could be explained by their alignment with some behavior change techniques known to facilitate exercise adherence. 43Several studies involved elements of monitoring and feedback where they were able to provide real-time feedback on individuals' levels of performance and offered clear instructions on how to perform the exercise. 26,27,29,32These elements combined with professional support have been shown to increase exercise adherence. 43 set of key components for successful fall prevention exercise programs has been identified in community-dwelling adults. 4Interventions that included an exercise dose of more than 3 h per week and included balance and strength exercises reduced falls by 34%. 4 However, adherence to in-person fall prevention exercise programs has previously been reported to be *50%. 44In our review, we found that the adherence of exercise interventions delivered by telehealth was high (86%), but the exercise dose achieved was low (1.3 h/week).Future research could investigate how to utilize telehealth's high adherence to support increased exercise dose to improve physical outcomes.
This review highlighted the importance of providing staff and participant training to improve the chances of successful program implementation.The importance of technology training was also reported in a recent qualitative study where telehealth was used to enhance mobility and physical activity for older adults receiving rehabilitation services. 45They found that patients engaged optimally with telehealth when they

DAWSON ET AL.
received sufficient training and support to use the technology and understood the potential benefits from using telehealth.Their study also reported the importance of sufficient therapist training to increase telehealth technology competence and overall program implementation.

Conclusions
This is the first systematic review and meta-analysis that provides a summary of the impact of telehealth on physical outcomes for older adults, 60 years of age and older, with mobility, frailty, and cognitive disabilities.We conducted this systematic review in accordance with PRISMA guidelines and followed a pre-specified protocol registered on PROSPERO.Furthermore, the controlled studies included were of sound methodological quality.However, due to the study heterogeneity and low number of controlled studies in this area, we were unable to uncover any significant finding.There was also insufficient follow-up data, which limited our ability to assess the long-term effects and sustainability of exercise interventions delivered through telehealth in this population.
Future RCTs are required to investigate the use and effects of synchronous different exercise programs delivered through telehealth, trials that investigate use effects of asynchronous telehealth versus asynchronous telehealth, trials that include cost-effective analyses, and trials that explore the implementation and sustainability of these telehealth interventions to ensure that wide scale uptake of telehealth in aged care is as safe, effective, and cost-effective as possible.Telehealth used to deliver evidence-based exercise intervention to our most vulnerable older adults has the potential to be an effective and acceptable addition to in-person exercise interventions.

Fig. 1 .
Fig. 1.Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart.

Fig. 5 .
Fig.5.Effect size (95% confidence interval) of telehealth versus control on quality of life using random-effects meta-analysis.

Table 1 .
Trial Design and Participant Characteristics (N a Physiotherapy Evidence Database (PEDro) Score only for RCTs.46 C, control; CI, cognitive impairment; I, intervention; IQR, interquartile range; mod, moderate; NA, not applicable; OA, osteoarthritis; QoL, quality of life; PD, Parkinson's disease; RCT, randomized controlled trial.Data are mean -SD, unless otherwise stated.
Bernard et al. commented that delivering synchronous exercise classes to