The Effects of ICT-Based Interventions on Physical Mobility of Older Adults: A Systematic Literature Review and Meta-Analysis

Systematic literature review and meta-analysis were conducted to integrate and analyze intervention studies dealing with the effects of information and communications technology- (ICT-) based interventions on the physical mobility of older adults in the community. The PubMed/MEDLINE, Embase, CINAHL, and Cochrane CENTRAL databases were searched for studies published from January 2000 to December 2022. We used the Risk of Bias 2 (RoB 2) tool to evaluate the quality of the randomized controlled studies in the systematic review. The meta-analysis was performed using a random-effects model. The model was used to calculate the standardized mean difference (SMD) and 95% confidence interval (CI) for both effect measures. I2 tests were used to measure the presence of heterogeneity. Thirty-seven randomized controlled trials were included (2,419 intervention participants), of which 23 were included in the meta-analysis. ICT interventions significantly improved Timed Up and Go (TUG) as a marker of physical mobility variable in older adults (SMD = −0.33, 95% CI: −0.57 to −0.10, p=0.005, I2 = 74.7%). A sensitivity analysis was performed on subgroups, and interventions were found to be effective in improving TUG in the exergame group (SMD = −0.40, 95% CI: −0.72 to −0.08, p < 0.001, I2 = 75.0%) and in the exergame with virtual reality (VR) group (SMD = −0.33, 95% CI: −1.01 to 0.35, p < 0.001, I2 = 91.0%) but both groups showed high heterogeneity. A meta-analysis was also performed on Short Physical Performance Battery (SPPB) but statistically significant results were not found (SMD = −0.19, 95% CI: −0.61 to 0.23, p=0.375, I2 = 87.7%). For the Berg Balance Scale (BBS), the post-intervention scores were significantly better than baseline (SMD = 1.52, 95% CI: 0.48 to 2.57, p=0.004, I2 = 93.5%). However, the number of studies included in the meta-analysis was small and heterogeneity was high, so follow-up studies are needed. This study confirmed that exergames, telecommunication, e-health, information applications, and robots were used as effective ICT-based interventions for improving the physical mobility of older adults. It is necessary to develop and apply more diverse ICT-based interventions that will prevent impairments of mobility and encourage older adults to live more independently, with a higher quality of life, based on extensive research on ICT-based interventions.


Introduction
Te prolonged life expectancy and rapidly growing worldwide population of older adults have brought age-related physical, cognitive, and psychosocial health issues into the societal spotlight [1].Older adults experience declining physical function (e.g., reduced muscle strength), impaired sensory function (e.g., vision and hearing), and decreased mobility caused by multiple factors, including reduced social activities after retirement [2].In particular, maintaining mobility is an important goal for older adults to maintain independence and quality of life [3].In older adults, reduced physical mobility is likely to have negative impacts on their life, including an increased likelihood of falling and hospitalization [4,5], as well as placing them at higher risk for depression, social isolation, and loneliness [6,7].It has been established that senior citizens capable of standing for extended periods or traveling to various locations tend to have a lower risk of death [8], be more independent, and have a better quality of life [3].
Te term "mobility" has various meanings depending on the context in which it is used [9].Mobility, usually understood as a component of overall function, is defned as the ability to move or be moved easily and freely [10,11].Mobility is also used in a broader meaning, encompassing not only ambulation but also to participation in daily life or leisure activities, exercise, and using a variety of public transport modes [12,13].Given the multifaceted nature of mobility, methods to measure it are highly diverse.For example, physical activity, physical performance, muscle mass and strength, and balance and gait performance have been used to assess the level of mobility [3,14].Among several assessment tools, the Timed Up and Go (TUG) test, Short Physical Performance Battery (SPPB), and Berg Balance Scale (BBS) have frequently been used to evaluate older people's mobility [9].In this review, mobility was defned as "physical mobility," focusing on a person's physical ability to change his or her location or position or move from one place to another by walking and basic ambulation.
Recent advances in information and communications technology (ICT) have allowed the healthcare and medical sector to utilize the benefts of ICT in many ways, with impacts including reduced medical expenses, improved administrative tasks, maintaining patients' medical history, and reduced traditional paperwork [15].A report released by Statista, a global statistics portal service, estimated that the global digital healthcare market in 2018 was worth USD 84.9 billion and was expected to grow to USD 504.4 billion by 2025 [16].Te use of telemedicine rapidly increased during the COVID-19 pandemic [17], and the COVID-19 pandemic has also resulted in enhancing digital acceptance among older adults [18].Following this trend, ICT has been increasingly incorporated in various interventions for older adults to help with their daily routines and reduce healthcare costs [19].According to a systematic literature review on ICT-related interventions for seniors, interventions using computers and the Internet, robotics, telemedicine, virtual reality, video games, and sensor technology have proven to be efective in lowering fall risk and social isolation, improving quality and satisfaction of life, increasing gait speed, and reducing depression [20].Studies on ICT-based interventions to improve physical mobility in older adults include a pilot study that used an interactive smartphone application to boost physical activity in older adults [21] and a study on gait performance during wearable robot-assisted gait in older adults [22].Another systematic literature review reported that exergame technology and interactive interventions contributed to higher mobility and enhanced balance in older adults [23,24].A more comprehensive analysis is needed to understand the efects of other types of ICT-based interventions on enhancing physical mobility among older adults, as more ICTbased interventions will be performed in the future.
Various types of ICT-based interventions are performed to promote physical mobility, which has positive efects on older adults' quality of life [25].Tis study presents a systematic literature review and meta-analysis of studies on ICT-based interventions to promote mobility among older adults to provide a comprehensive and objective conclusion on the topic.Tis study will help understand the impacts of ICT-based interventions on improvements in physical mobility in the older population, given the trend for technological advancements, and can provide a foundation to promote successful aging through improving physical mobility and the quality of life in the older population.

Methods
2.1.Research Question.Systematic literature review and meta-analysis were conducted to verify the efects of ICTrelated interventions on the physical mobility of older adults.Tis study was conducted according to the systematic literature review guidelines of Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) [26].
Detailed data selection criteria were established as described below using PICO-SD, the key question strategy recommended in guidelines on systematic literature reviews [26].Te participants were community-dwelling older adults aged 65 and above, without physical limitations.Tose with severe cognitive impairments were excluded.Te interventions analyzed in this study utilized ICTmodalities (e.g., the Internet, wireless networks, cell phones, and other equipment and technologies).Interventions using robotics, telemedicine, sensor technology, video games, smartphones, mobile applications, and medication-dispensing devices were included in this study [20,27].Tis study only included studies with control groups.Te control groups for comparisons in this study were older adults (aged 65 and older) who did not receive interventions or those who received usual-care interventions for ethical purposes.Te outcomes included measured variables of senior physical mobility.Physical mobility for the purpose of this study denoted the physical ability to move from one place to another (i.e., physical performance and physical activity).In this review, variables measuring physical mobility were classifed into fve categories: physical activity, physical performance, muscle mass and strength, and balance and gait performance [3,14].A meta-analysis was conducted on physical performance and balance, since these were the categories with sufcient studies to enable a meta-analysis [28].Table 1 shows the measurement variables according to the classifcation of physical mobility used in this study.Te study type was limited to randomized controlled trials (RCTs) only.Studies conducted among adults aged 65 and under or older adults residing in facilities including nursing homes, interventions not utilizing ICT, interventions that used ICT simply as a tool for contacting participants, and interventions with the main purpose of treating or rehabilitating a particular disease were excluded.

Search Strategy.
Tree researchers who had experience in meta-analyses and literature searches conducted the literature search for this study after receiving IRB approval (IRB No. KYU-2020-145-01), and its protocol has been registered in PROSPERO (No. CRD42021225483).

2
International Journal of Clinical Practice Te search formula was created with a combination of terms representing the older population aged 65 and older (P) and ICT-based interventions (I).Four databases (PubMed/ MEDLINE, Embase, CINAHL, and Cochrane CENTRAL) were selected based on the COSI model suggested by the National Library of Medicine (NLM).Relevant full publications and conference abstracts were identifed by electronic searching of the four online databases using both text words and exploded Medical Subject Heading (MeSH) terms: (aged) AND (locomotion OR exercise OR Physical Functional Performance OR Walking Speed OR Muscle strength OR Postural Balance OR Mobility Limitation) AND (telemedicine OR information technology OR Information Science OR Robotics OR Video games OR Cell Phone OR Smartphone OR Mobile Applications).Te results were limited to RCTs published in English between January 2000 and December 2022.In addition to the MeSH terms, text search terms were entered in the search.Te detailed search formula is outlined in the Supplementary Material (Supplementary Appendix S1).

Data Extraction. Te items of the data extraction form
for systematic literature review were decided by consensus among the three researchers.Te data analysis form included the author, published year, country of the study, place of the study, characteristics of inclusion/exclusion criteria, age and gender of participants, ICT intervention type, devices used, whether the study analyzed an individual or group intervention, the duration and frequency of the intervention, the intervention provider, the duration of follow-up, efect variables, and devices used for outcome measurements.If there were inconsistencies in the results among researchers, fnal decisions were made after reviewing and discussing the original studies.
2.4.Quality Assessment.Te 37 selected studies were evaluated using the revised Cochrane risk-of-bias tool for randomized trials (RoB 2) for RCT studies developed by the Cochrane Bias Method Group [29].Te RoB 2 tool consists of 22 questions in fve areas including randomization process, intended interventions, missing outcome data, measurement of the outcome, and reported results.Te choices for answering each question were "yes," "probably yes," "probably no," "no," and "no information."Each researcher decided whether the risk of bias was "low risk," "some concerns," or "high risk" and reevaluated the literature for questions where they had disagreements.Te researchers reached a conclusion after sharing and discussing each other's evaluation records for these questions.

Statistical Analysis.
For studies that were suitable for meta-analysis, the efect size and homogeneity of the ICT interventions were calculated using R version 4.2.1.A metaanalysis was conducted when four or more studies reported data on the same outcome variable [30].Terefore, a metaanalysis was conducted on the TUG, SPPB, and BBS, which are commonly used to measure physical mobility [30].Te standardized mean diference (SMD) was used to quantify the efect size of outcome variables reported with diferent measurement tools or units, and mean diference (MD) was used when the measurement tools and units were the same.
For a crossover study [31], since data at each starting point and end of follow-up were presented, each time point was regarded as a separate study and the standardized mean diference (SMD) was obtained and analyzed.In addition, for multiarm studies [32][33][34], the groups were combined and then analyzed [35].A random-efects model was used under the hypothesis that each study would have diferent participants, intervention methods, and research environment.Heterogeneity was estimated using the forest plot, and statistical hypothesis testing was conducted using the I 2 index to quantify the dispersion among the studies.An I 2 value of higher than 75% means considerable heterogeneity, 25% < I 2 ≤ 75% indicates moderate heterogeneity, and an I 2 value of 25% or less means low heterogeneity [30].If the outcome variables were measured twice or more, the value measured immediately after the intervention was adopted, considering that the results may be distorted with time, and the statistical signifcance of the efect size was evaluated using 95% confdence interval (CI) and a 5% level of signifcance.Te MD between two groups was considered insignifcant if the 95% CI included 0, while it was considered signifcant if the 95% CI did not include zero.Te interpretation of the efect size was based on Cohen's standardized mean diference, where 0.20 ≤ d ≤ 0.50 denotes a small efect, 0.50 ≤ d ≤ 0.80 indicates a medium efect, and d ≥ 0.80 denotes a large efect [36].Te funnel plot, Begg and Mazumdar's rank correlation test, and Egger's linear regression test methods were used to evaluate publication bias.

Results
Tis study reviewed the existing literature to identify the efects of ICT-based interventions on the physical mobility of older adults.Te online database search yielded a total of 6,496 studies, including 2,493 from PubMed/MEDLINE, 1,719 from CINAHL, 2,154 from Embase, and 130 from the Cochrane CENTRAL.Te number of overlapping studies from the frst search was 2,131.Of the 4,365 studies, 50 were selected after applying the inclusion and exclusion criteria upon reviewing the titles and abstracts.Tirty-seven studies were fnally selected for analysis, removing four studies that did not match the age criteria, three non-RCT studies, fve studies that did not have eligible outcomes and research environment criteria, and one study that did not ft in terms of the intervention.23 studies with measurements of the same variables were fnally selected for the meta-analysis (Figure 1).[38-40, 51, 62-64, 67], South America (one study) [61], Europe (10 studies) [1,41,43,44,48,53,54,57,60,65], Middle East (four studies) [45,49,50,68], Asia (nine studies) [31-33, 47, 52, 55, 56, 58, 66], Australia/Oceania (four studies) [34,37,42,61], and Africa (one study) [59].Te participants were both male and female in 35 studies, while one study had female participants [46], and one study had male participants [57].Eight studies identifed only the gender of the total participants and did not specify the gender ratio of the intervention group and control group [34,40,45,54,56,57,59,61].
Telecommunications interventions mainly used telephones [40,51,57,64], although smartphones [53] and video conferencing units [53,67] were used, too.In applicationbased interventions, tablets and smartphones were used, and accelerometers, personal computers, smartphones, and tablets were used in web-based communication interventions [38,60,66].A balance exercise assist robot and an exoskeletal hip-assist robot were used in robot interventions [32,56], and wearable motion sensors were used in a study of wearable devices [55].

Exergames with VR.
Tree studies used exergames with VR, and the intervention group had 101 participants (range, 13 to 45) including 10 females (except for those that did not report gender), compared to 41 (range, 11 to 15) in the control group, including 10 females.Te sample size ranged from 24 to 60 participants in total [33,34,41].Te mean age was 76.29 in the intervention group and 77.90 in the control group.
In terms of the contents of interventions, two studies provided exercise interventions [33,41], and one study provided exercise and cognitive function training [34].Te interventions were provided by a trainer/researcher [33] or physiotherapists [34,41].Te mean frequency and duration of the interventions were 25.33 minutes/session and 3 sessions/week for 7 weeks.
Among the mobility variables measured in this study, TUG was most frequently reported to be signifcant in all three studies.It was confrmed that there was a signifcant improvement compared to the pre-test in the 10 m walk test and single-leg stance test, and signifcant changes were also reported in variables such as the Performance Oriented Mobility Assessment (POMA) [34,41] and muscle strength [33].It was also reported that the VR intervention group exhibited improved physical mobility compared to the control group [33,34].

Telecommunication. Six studies used telecommunication
technologies for their interventions [40,51,53,57,64,67].Te number of participants in the intervention and control groups was 354 (range, 28 to 234) including 231 females and 317 (range, 13 to 160) including 210 females, respectively (with the exclusion of one study because it did not report the number of participants) [40].Te sample size ranged from 41 to 351 participants in total, and the mean age was 73.79 in the intervention group and 73.87 in the control group.
In terms of the contents of interventions, two studies provided exercise interventions [53,67], three studies provided motivation for exercise [40,51,57], and one study provided motivation for exercise and nutritional advice [64].In one study [64], the intervention was self-conducted by the participants, while in other studies, the interventions were provided by researchers [40,51] or counselors [57], or were self-conducted by the participants with an educationalprofessional physical trainer [53].Te interventions were provided for an average of 44.67 minutes, 1.79 times/week, for 14.67 weeks.
Tese studies demonstrated efectiveness in terms of improving single leg stance (SLS), medial-lateral foot center of pressure (ML-COP), and TUG scores [67] and increasing International Journal of Clinical Practice physical activity [51,64,67].Venditti et al. [64] reported that SPPB, gait speed, and 5-chair-stand components demonstrated efectiveness compared to the pre-test, but there were no signifcant diferences between groups.Furthermore, Langeared et al. [53] reported that a videoconferencingbased exercise intervention was partially efective in strengthening muscles (knee and lower limb) but was not as efective as a face-to-face exercise intervention in improving knee fexion isometric strength.Two studies provided motivation for exercise [37,65], one study motivated exercise and provided health advice [42], and other study provided exercise, cognitive training, and nutritional advice [43].Te intervention was selfprovided via the app or the web page [42,43,65], and one study was provided by an expert [37].Te interventions lasted for 11.75 weeks.
Delbaere et al. [42] demonstrated efectiveness in terms of improving standing balance and functional mobility (such as TUG, the 5 times sit-to-stand test, 10 m walk, and SPPB) compared to the pre-test, but there were no signifcant diferences between groups.Alley et al. [37] reported that the intervention improved moderate to vigorous physical activity (MVPA) outcomes in comparison with a control group.[38,66] and one study used an application with wearable devices [60].Te number of participants in the intervention group was 259 (range, 80 to 150) including 164 females, while the control group had 662 participants (range, 60 to 264) including 415 females, and the sample size ranged from 157 to 263 participants in total.Te mean age was 69.94 in the intervention group and 70.80 in the control group.

Applications. Two studies used applications
In terms of the contents of interventions, all studies provided motivation for exercise and presented health or nutritional advice.Te interventions were self-conducted by the participants in two studies [38,66] and by a medical doctor for 24 weeks in another study [60].
Regarding outcome variables, inconsistent results were reported for gait step-namely, Bickmore et al. [38] reported that the intervention group participants walked signifcantly more than control participants, but Recio-Rodríguez et al. [60] reported that there was no signifcant diference between the groups.Wang et al. [66] reported that the intervention improved skeletal muscle mass, but that it was particularly efective in the group where exercise and nutritional counseling were also provided.
3.2.6.Robots.Robotic technology was used in two intervention studies, and the intervention group had 29 participants (range, 14 to 15), compared to 26 (13 participants each) in the control group.Te sample size ranged from 27 to 58 participants in total [32,56].One study reported only total participants' sex ratio and age (20 females among 29 participants; mean age, 73) [56], and another study's intervention group had 21 females (mean age, 74.2) compared to 7 females (mean age, 76.4) in the control group [32].All studies provided exercise interventions where participants performed the interventions themselves with the help of researchers.Te mean frequency and duration of the interventions were 40 minutes/session and 2.5 sessions/ week for 8 weeks.Te outcome variables showed inconsistent results; Ozaki et al. [56] reported that the TUG, FRT, muscle strength (abduction, extension, adduction, and fexion-hip and knee), and gait speeds exhibited a signifcant efect compared to the control group.Lee et al. [32] reported that the SPPB, TUG, FRT, and muscle strength (abduction, extension, adduction, fexion-hip and knee, and trunk) demonstrated efectiveness compared to the pretest, but there were no signifcant between-group diferences.

Wearable Devices.
Wearable devices were used in one intervention study [55], where participants performed the intervention with assistance from researchers.Te intervention group had 36 participants including 16 females (mean age, 69.3), and the control group had 34 participants including 16 females (mean age, 68.8).
Te exercise interventions were provided for 60 minutes/ session and 3 times/week for 12 weeks.Te study reported that the Chair-Stand-30 (CS-30) scores and other physical function outcomes (TUG and CS-30) improved in both the intervention and control groups, but there was no statistically signifcant diference between the groups.

3.3.
Quality Evaluation of the Literature.Te 37 publications were evaluated for their risk of bias using Microsoft Excel tool in accordance with RoB 2, provided by the Cochrane Group.In the frst area of the randomization process, 16 studies (42.1%) were evaluated as low risk because they followed the randomization process relatively well with a detailed outlining of the randomization process and hiding the assignment order, while 21 studies (55.3%) were evaluated as having some concerns because they did not detail the process of randomization (Figure 3).One study (2.6%) was randomized but not blinded to participants and classifed as high risk [32].
In terms of the intended interventions, 28 studies (76.3%) were evaluated as low risk because they included statistical information on participants who were excluded from the intended intervention or dropped out but had a small impact on the outcomes.Tree studies (7.9%) were evaluated as high risk because 5% or more of the participants were excluded from the analysis.Seven studies (15.8%) were evaluated as having some concerns about carrying out a modifed intention-to-treat protocol because they included all participants except for those with missing outcomes or it 14 International Journal of Clinical Practice was unclear whether the participants or researchers were aware of the randomization.Te evaluation of missing outcome data estimated that three studies (7.9%) had some concerns because the dropout rate of the initial participants was 5% or more, and all others were evaluated as low risk.For measurements of the outcomes, seven studies (18.4%) had some concerns because a double-blind study was not conducted; thus, the outcome assessors may have been aware of the interventions that the study participants received.Tere was deemed to be a low risk of bias for the reported results (Figure 3).
A subgroup analysis was performed because the original meta-analysis was performed with diferent kinds of ICT interventions in a single analysis, which could lead to bias in interpreting the outcomes.In the subgroup analysis, in 12 exergame studies [31,45,46,48,49,52,58,59,61,62,68] and three exergame with VR studies [33,34,41], TUG signifcantly decreased in the post-test compared to the pre-test, proving that the interventions had efects.However, both groups showed high heterogeneity (exergame I 2 � 74.7%, VR with exergame I 2 � 91.1%) (Figure 5).To assess the impact of an individual study on the pooled estimates, a sensitivity analysis was conducted by excluding one study at a time.For TUG, the sensitivity analyses yielded similar results, indicating that no individual study infuenced the TUG (Supplementary Figure 1).

Short Physical Performance Battery (SPPB).
Of the 37 studies, six used SPPB as an outcome variable.A metaanalysis was performed of fve studies with a verifed efect size [32,43,44,63,64], unifying the efect size of each outcome using Hedges' g value and distribution.One study was excluded from the meta-analysis because standard Figure 3: Risk of bias summary in included studies.Te 37 publications were evaluated for their risk of bias using the tool to implement RoB 2 provided by the Cochrane Group.Te evaluation of missing outcome data estimated that two studies had some concerns because the dropout rate of the initial participants was 5% or more, and all others were evaluated as low risk.Tere was deemed to be a low risk of bias for the measurement of the outcome and the reported results.

Berg Balance Scale (BBS).
Of the 37 studies, six used BBS as an outcome variable.A meta-analysis was performed of fve studies with a verifed efect size [31,32,34,49,68], unifying the efect size of each outcome using Hedges' g value and distribution.One study was excluded from the metaanalysis because standard deviations were not presented [39].

Discussion
Te purpose of this study was to verify the efects of ICTbased interventions on the physical mobility of older adults through a systematic literature review and meta-analysis of International Journal of Clinical Practice RCTs.In total, 37 studies were selected for the systematic literature review.A meta-analysis was performed of the 23 studies that reported TUG, SPPB, or BBS outcome variables to identify the efect size.
Twenty-four of the 37 studies were published after 2020, refecting the recent trend for more publications as ICTbased interventions gained attention during the coronavirus disease 2019 (COVID-19) pandemic.Among the various types of interventions, exergames were used in 18 studies, accounting for almost half (47.4%; reaching 55.3% if including exergame using VR), followed by telecommunications, one of the most traditional types of ICT (six studies, 15.8%).Tis is in line with the results of previous systematic literature review studies suggesting that exergames are commonly used as an efective intervention for promoting physical activities or mobility among older adults [24,69,70].Exergames readily triggered interest and motivation among participants, bringing higher compliance and persistence than traditional exercise interventions, and are considered a cost-efective intervention to encourage physical activity [39,48,54].Research has suggested that contactless interventions based on the web or telecommunications had markedly higher compliance than in-person interventions because providing in-person interventions at a certain venue made it inconvenient for the participants to be there at a certain time, even if the physical distance was minimal [38,40,67].Tis is a meaningful fnding given the current circumstances with rising demand for the development and application of contactless healthcare interventions throughout the COVID-19 pandemic.It shows that applying ICT in various healthcare interventions can be efective for increasing compliance by triggering interest and facilitating convenience among participants.In particular, during the COVID-19 pandemic, ICT-related health intervention services were efective for supplementing faceto-face intervention services in situations where it was difcult to deliver health services face-to-face [71,72].However, it is difcult to comply with regulations (intensity, frequency, posture, etc.) for ICT services compared to faceto-face services, so related guidelines are needed [53].

18
International Journal of Clinical Practice Terefore, in order to prepare for these issues, prior training on posture when providing ICT services, periodic feedback from experts, and measures for safety issues would be essential [71].Tis study also found that telecommunications-based interventions, as a type of classic ICT, were cost-efective in boosting physical mobility among community-dwelling older adults.Tis conclusion is supported by another study that found telephone counseling to be efective in encouraging moderate physical activity [51,57] and a study reporting that using both telephone and mail for exercise counseling was more efective in promoting exercise among participants than simply using mail only [40].Interventions that provided exercise information using applications or websites were also efective in promoting physical mobility among older adults [38,65], as were interventions using robots [56].A systematic literature review of these studies confrmed that most ICT-based interventions, including training sessions using exergames and robots, as well as information-based interventions such as web-based interventions, applications, and telecommunications, were efective in improving mobility among older adults.
Interventions that used computer games and gaming devices for cognitive training exclusively or in combination with exercise interventions were efective for improving physical mobility [44,61,62].Tis result is consistent with the outcomes of existing systematic literature reviews proving that individuals' physical mobility can improve as a result of utilizing ICT-based virtual reality for learning and processing diverse information in the central nervous system such as visual images, exercise planning, and motivation [52,61], and that general cognitive training can improve mobility among older adults, especially in the context of higher-order executive function (such as walking while talking) [73].Te study published by Smith-Ray et al. [62] also confrmed that cognitive training using ICT improved the range of feld of view, driving ability, and confdence, implying their efectiveness for mobility.However, considering the potential safety concerns such as fall risk and injury when applying ICT-based interventions for older adults in the community, it is important to make sure that the interventions are safely performed with the help of researchers or trained medical staf [39,48,52,56,61].In addition, an introductory session on how to safely use ICT devices should be provided prior to  International Journal of Clinical Practice ICT-based interventions so that participants can better understand ICT-related interventions and apply them in a safe manner and ensure higher compliance and lower dropout rates [31,38,49,52,54].
In particular, with the recent development of VR technology, VR exergames have been introduced.Compared to regular exergames, VR allows greater immersion in the situation, thereby increasing interest [33,34,41].Compared to exergames alone, physical mobility was further increased by exergame interventions with VR, most likely because VRbased exergames require much greater sensory integration and processing [33].Tis fact can be used as a motivational strategy for participation in mobility promotion interventions for the elderly.
Variables for measuring the physical mobility of older adults included measures of primary outcomes, such as physical activity (e.g., International Physical Activity Questionnaire (IPAQ), MVPA, etc.), physical performance (e.g.SPPB, TUG test, etc.), muscle mass and strength (e.g.30s chair stand test, skeletal muscle mass, muscle function, etc.), balance (e.g.BBS, standing balance (s), etc.), and gait performance (e.g., single-task walking (m/s), step count, 6MWT, etc.).In the meta-analysis, it was difcult to analyze the integrated efect size across the studies due to the inconsistent mobility measurement variables.21 out of the 37 selected studies reported TUG as an outcome variable, and a meta-analysis was performed on the 20 studies in which the efect size could be confrmed.ICT interventions were found to be efective in improving TUG as a physical mobility measurement variable.A sensitivity analysis was performed for each subgroup, and the exergame and exergame with VR groups showed high heterogeneity, but TUG decreased from the pre-test to the post-test, proving that it was efective.
In addition, a meta-analysis was performed on SPPB and BBS.For BBS, the post-intervention efect was signifcant.However, the number of studies used in the meta-analysis is small and heterogeneity is high, so follow-up studies are needed.Although TUG, SPPB, and BBS are widely used clinical assessment tools to evaluate balance and walking ability among the older adults, the concept of mobility in the elderly is a complex concept [3], so a tool that can evaluate these factors together is needed.
Tis study is meaningful in that it verifed the efects of ICT-based interventions on the physical mobility of community-dwelling older adults through a systematic and objective integration of individual studies.Most of the studies included in the analysis had a low risk of bias, and only RCT studies were included to ensure credibility.However, this study had limitations in that it only included studies published in the English language.In addition, this study included interventions with a small sample size (15 or fewer participants) [1,31,39,41,54,56,61] in the metaanalysis.Tis requires attention due to the possibility of overestimating the efect size.Terefore, to compensate for this issue, Hedges' g value was used in the meta-analysis.If the number of studies included in a meta-analysis is less than 10, the statistical test may not detect heterogeneity among the studies.For the outcome variables of SPPB and BBS, the number of included studies was less than ten; thus, there may have been heterogeneity that was not found in this study.Furthermore, for TUG, the limitations were supplemented by identifying heterogeneity-related factors through subgroup analysis.In this study, a meta-analysis was performed with only papers published in academic journals, which was motivated by the need to include high-quality articles on this topic.However, it has been pointed out that a meta-analysis can produce more reliable results when studies published in academic journals and unpublished studies are included in the analysis at a similar ratio [74].Tus, a follow-up meta-analysis incorporating unpublished studies at an appropriate ratio is recommended.
Most of the studies excluded from this study dealt with ICT interventions provided to older adults, focusing on rehabilitation from certain conditions, including dementia, stroke, and Parkinson's disease, as well as patients who had undergone musculoskeletal surgery, implying the need for further research on developing and expanding ICT-based interventions efective for promoting mobility among healthy community-dwelling older adults.Additionally, this study included interventions conducted in 22 countries distributed across fve continents.Each country has diferent levels of digitalization and diferent levels of informatization, and these characteristics may afect the themes or efectiveness of interventions due to the level of development of the information society [75].Terefore, the fndings of this study should be interpreted carefully.We also suggest analyzing whether diferences exist in the type and efect of ICT intervention depending on the level of digitalization.

Conclusion
Tis study systematically reviewed research on the efects of ICT-based interventions on physical mobility among community-dwelling older adults and conducted a metaanalysis to determine the efect size of TUG, SPPB, and BBS, a variable for measuring physical mobility in older adults.Te outcomes demonstrated that ICT interventions using exergames, e-health, information applications, and robots were efective in enhancing senior physical mobility, as well as telecommunication interventions (as the most traditional ICT intervention).Moreover, ICT interventions were efective in enhancing physical mobility.In the future, more diverse ICTbased interventions should be developed and provided to older adults in the community to prevent impairments of mobility and cognitive function and to encourage older adults to live more independently, with a higher quality of life, based on extensive research on ICT-based interventions.

20
International Journal of Clinical Practice

Figure 2 :
Figure 2: Distribution of types of interventions on physical mobility of older adults.

Figure 4 :Figure 5 :
Figure 4: Forest plot displaying the results of a meta-analysis of the outcome (TUG) of ICT-based intervention in community-dwelling older adults.Te standardized mean diference (SMD) was used to quantify the efect size of outcome variables reported with diferent measurement tools or units, and the pooled results signifcantly showed that ICT-based interventions could improve TUG of older adults signifcantly (SMD � −0.33, 95% CI: −0.57 to −0.10, p � 0.005).

StudyFigure 6 :Figure 7 :
Figure 6: Forest plot displaying the results of a meta-analysis of the outcome (SPPB) of ICT-based intervention in community-dwelling older adults.Te standardized mean diference (SMD) was used to quantify the efect size of outcome variables reported with diferent measurement tools or units.Te pooled results were not signifcant; there is insufcient evidence that ICT-based interventions could improve SPPB in community-dwelling older adults (SMD � −0.19, 95% CI: −0.61 to 0.23, p � 0.375).

Figure 8 :
Figure 8: Funnel plots with TUG (a), SPPB (b), and BBS (c) as the outcome variable of ICT-based intervention for the community-dwelling older adults.(a) Funnel plots with Timed Up and Go (TUG).(b) Funnel plots with Short Physical Performance Battery (SPPB).(c) Funnel plots with Berg Balance Scale (BBS).

Table 1 :
Measurement variables according to the classifcation of physical mobility.MWT � 6 minute walk test; CS-30 � 30-sec chair stand test; MVC � maximum voluntary contraction; RFD � rate of force development; KES � knee extensor strength; COP � center of pressure; POMA � Performance Oriented Mobility Assessment; COP-VM � center of pressure velocity moment; BBS � Berg Balance Scale; UST �unipedal stance test; SLS � single leg stance; ML-COP � medial-lateral foot center of pressure; FAB � Fullerton Advanced Balance Scale.
dynamic/static balance (COP), single-leg stance test on frm and foam surfaces, tandem stance test, ML-COP, SLS, FAB Gait performance Single-task walking (m/s), Digiwalkers, maximum anteroposterior leaning range, coordinated lean score, 10 m walk, stepping reaction time, gait speed, step count, step counts from accelerometers, 6-minute walk test HPAS � Houston Physical Activity Scale; QAPPA � quantization autotuner for precision programmable accelerators; IPAQ � International Physical Activity Questionnaire; SF-36 � 36-item short form survey; CHAMPS � Community Healthy Activities Model Program for Seniors questionnaire; MVPA � moderate to vigorous physical activities; SQUASH � Short Questionnaire to Assess Health Enhancing Physical Activity; SPPB � Short Physical Performance Battery; FRT � Functional Reach Test; TUG � Timed Up and Go; FTSTS � fve times sit-to-stand test; 6-

Table 3 :
Details of ICT-based interventions on physical mobility of older adults.