Acceptability, Feasibility, and Effectiveness of Immersive Virtual Technologies to Promote Exercise in Older Adults: A Systematic Review and Meta-Analysis

Context: This review aimed to synthesize the literature on the acceptability, feasibility, and effectiveness of immersive virtual technologies to promote physical exercise in older people. Method: We performed a literature review, based on four databases (PubMed, CINAHL, Embase, and Scopus; last search: 30 January 2023). Eligible studies had to use immersive technology with participants aged 60 years and over. The results regarding acceptability, feasibility, and effectiveness of immersive technology-based interventions in older people were extracted. The standardized mean differences were then computed using a random model effect. Results: In total, 54 relevant studies (1853 participants) were identified through search strategies. Concerning the acceptability, most participants reported a pleasant experience and a desire to use the technology again. The average increase in the pre/post Simulator Sickness Questionnaire score was 0.43 in healthy subjects and 3.23 in subjects with neurological disorders, demonstrating this technology’s feasibility. Regarding the effectiveness, our meta-analysis showed a positive effect of the use of virtual reality technology on balance (SMD = 1.05; 95% CI: 0.75–1.36; p < 0.001) and gait outcomes (SMD = 0.7; 95% CI: 0.14–0.80; p < 0.001). However, these results suffered from inconsistency and the number of trials dealing with these outcomes remains low, calling for further studies. Conclusions: Virtual reality seems to be well accepted by older people and its use with this population is feasible. However, more studies are needed to conclude its effectiveness in promoting exercise in older people.


Introduction
According to the World Health Organization, the number of people aged 60 years and over will reach 2 billion by 2050, while those aged 80 years and above are expected to grow from 125 million (in 2018) to 434 million in 2050 [1]. This accelerated aging currently observed in most industrialized countries is causing an increase in the prevalence of people with functional limitations related to mobility and fall risks. Indeed, aging results in a progressive decline of different body functions, leading to a higher risk of morbidity [2] and recurrent balance and walking disorders. Over the age of 65 years, more than a third of people fall at least once a year [3] as gait and balance disorders increase with age [4]. Given their prevalence and the physical, physiological, and psychological impact in older people, falls are a significant concern for health systems. Falls are predictors of decreased social participation [5]. It is therefore critical to find effective avenues for helping older people to prevent falls and rehabilitate balance disorders in order to maintain their independence in daily activities.

Results
The electronic search strategy in the MEDLINE, CINAHL, Embase, and Scopus databases yielded 2542 records. Handsearching led to 41 additional articles ( Figure 1). As a result, a total of 2583 articles were exported into the Covidence software [22]. After removing the duplicates, 2070 titles and abstracts were screened. A total of 54 different studies (1853 participants) were finally selected. These studies were issued from 23 different countries (Table 1). The years of publication ranged from 2006 to 2022. In total, 91% of the included studies were published after 2015 and 67% were published in 2020 or later. In the next subsections, we report the most important findings.

Characteristics of the Experiment Designed in the Selected Studies
As shown in Table 1 [36,43,69], and 3 studies [30,33,45] used the Samsung Gear VR. The remaining studies used the following headsets: Revelation 3D VR Headset with a Lumia 930 phone [53], University of Ulster's Virtual Reality Rehabilitation (UUVRR) System [40], Valve Index [48], VR GLASS [70], and Balance Rehabilitation Unit (BRU) [57]. Jung et al. [44] did not mention the type of VR headset used in their study. The following AR headsets were also used in different studies: AIRO II [72], Glasstron PLM-5700 [76], Laster WAV∃ [73], Microsoft Kinect [74,78], NEURO RAR [79], Portable Exergame Platform for Elderly (PEPE) [75], Microsoft HoloLens [67], UNICARE HEALTH [77], and i-visor FX601 [81]. Table 1 also reports the different types of populations groups, as well as the average age, the time since diagnosis, and the severity of illness when available in the selected paper. In total, 1 study [30] (using VR technology) included subjects with mild to moderate dementia, 6 studies included participants with Parkinson's disease (5 VR and 1 AR [72]), 29 articles (22 VR, 7 AR, and 1 CAVE [82]) included healthy older people, 7 studies (3 VR and 4 AR) included people with stroke, 3 VR studies included patients with pain affecting their daily activities, 2 VR studies included patients with vestibular impairments, 2 VR studies included subjects at risk of falling, 3 VR studies included patients with cognitive impairments, 1 VR study [55] included a patient with a total knee replacement, 1 VR study [71] included a patient suffering from functional incapacities, 1 VR study [62] included a patient with hypertension, and 1 VR study [52] included a patient with a distal radius fracture.
In total, 10 studies (9 VR and 1 CAVE [82]) exposed their participants for no more than 15 min per session. In 18 studies using an immersive technology, the participants were exposed to a maximum of 30 min per session, whereas 10 studies (8 VR and 2 AR) exposed their participants to more than 30 min per session. Furthermore, 13 studies (9 VR and 5 AR) did not mention the exposure duration. As reported in Table 1, 30 studies (24 VR and 6 AR) exposed participants to several VR sessions. In 45 studies (36 VR, 10 AR, and 1 CAVE [82]), the participants were supervised during their experimentation. Finally, 9 studies (7 VR and 2 AR) did not report whether supervision was provided to participants while exposed to the virtual environment.

Methodological Quality Assessment
Tables 2-4 present the methodological quality assessment of the different studies included in this review. According to the PEDro scale (Table 2), 16 studies (14 VR and 2 AR) showed good quality and 14 studies (10 VR and 4 AR) showed acceptable quality. Regarding the non-experimental studies, the results are presented in Table 3. Based on the CEBM scale (Table 4), four qualitative studies (three VR and one CAVE) could be classified as of good methodological quality. However, the small sample sizes of these studies limit the generalizability of their respective findings.
The listed criteria below is Present (+) or Absent (−): a = the eligibility criteria have been specified; b = participants were randomly assigned to the groups; c = the assignment of participants to a group was concealed; d = at the beginning of the study, the groups were similar; e = evaluators who measured at least one key outcome did not know which group the participants were assigned to; f = measures of at least one key outcome were obtained in more than 85% of participants initially assigned to the groups; g = all participants for whom outcome measures were available received the assigned intervention; h = results of inter-group statistical comparisons are provided for at least one key outcome; i = the study provides both an effect size measure and a measure of dispersion for at least one key outcome.
Benham et al. [32] 2019 [36] 2022 Janeh et al. [42] 2019 Kim et al. [46] 2017 ;~= Uncertain; a = Was the research question or objective in this paper clearly stated?; b = Was the study population clearly specified and defined?; c = Was the participation rate of eligible persons at least 50%?; d = Were all the subjects selected or recruited from the same or similar populations (including the same time period)? Were inclusion and exclusion criteria for being in the study prespecified and applied uniformly to all participants?; e = Was a sample size justification, power description, or variance and effect estimates provided?; f = For the analyses in this paper, were the exposure(s) of interest measured prior to the outcome(s) being measured?; g = Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed?; h = For exposures that can vary in amount or level, did the study examine different levels of the exposure as related to the outcome (e.g., categories of exposure or exposure measured as continuous variable)?; i = Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?; j = Was the exposure(s) assessed more than once over time?; k = Were the outcome measures (dependent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?; l = Were the outcome assessors blinded to the exposure status of participants?; m = Was loss to follow-up after baseline 20% or less?; n = Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure(s) and outcome(s)?
3.3. Findings on the Acceptability, Feasibility, and Effectiveness 3.3.1. Acceptability Twenty-one articles (Table 5) have addressed the acceptability of VR [16,30,32,33,36,39,41,42,[48][49][50]52,57,62,64,66,67,80,82]). Syed-Abdul et al. [64] indicated that the headset (HTC Vive) was comfortable for the participants. Appel et al. [30] and Benham et al. [32] indicated that the participants found the immersive VR experience enjoyable (via a home questionnaire showing a high satisfaction rate). Brown [33], De Keersmaecker et al. [16], and Syed-Abdul et al. [64] also reported that their participants enjoyed the experience. In Appel et al. [30] and Brown [33], the participants reported that they would be willing to repeat the experience in the future if they had the opportunity. Benham et al. [32] showed that older people were very keen to try this new technology and Phu et al. [57] observed a similar rate of treatment adherence between the conventional exercise group and the immersive VR group, contrary to Cikajlo and Peterlin Potisk [39] and Syed-Abdul et al. [64] who reported a higher motivation towards the treatment in the VR groups compared to the conventional treatment groups. Janeh et al. [42] highlighted a moderate level of immersion and low fear of physical contact with the real environment during immersion. Syed-Abdul et al. [64] concluded that older people consider using a technology based on its ease and usefulness. Indeed, the enjoyment obtained during the experiences, as well as the perception of their participants, provided positive attitudes concerning the use of this new technology. No study has evaluated the acceptability of AR-based interventions and only one study addressed the acceptability of the CAVE system. Pedroli et al. [82] found that their participants were highly engaged when immersed in the CAVE environment. Accordingly, it appears that most participants reported that they forgot the training context, which could be responsible for their increasing implication in rehabilitation. Brown, 2019 [33] Interviews and focus groups Older people with experience with digital platforms needed less guidance. Experience would have been more enjoyable with music. Enjoyed seeing places in the present, but would also enjoy seeing places in the past or visiting places that they would not have the capacity to do so today. Should be able to share this experience with others and not just do it alone for storytelling and socialization. Could help those with cognitive or physical limitations. Could have 3D meetings with family members or friends. Could increase feelings of isolation, anxiety, and depression in some people who are physically limited.

N/A
Participants reported that they enjoyed the experience and would consider using VR again if given the opportunity. Good option for reliving certain experiences and for entertainment, exploration, education, and socialization. People with more experience with new technologies would find it easier to use virtual reality. Participants reported feeling safe at all times. VR can promote socialization if it allows for the incorporation of family and friends. May increase some feelings of isolation, anxiety, or depression in some people. Would benefit to discuss these concerns prior to use.  The acceptance did not differ between the guided instruction VR-SET and the gamified VR-ET exergame. Interview with open questions "I felt like I was in a real park" "I was focused on the task "The environment was realistic" "I think it is easier to train with this tool" "I felt like the animals were touching me" "I felt passive and not active in the environment" N/A Participants were very involved in the environment and in the task. Participants forgot the context in which they were training. This may encourage patients to participate in their rehabilitation sessions.

Feasibility
Twelve studies [16,[35][36][37][38]41,46,48,53,56,61,64] used the Simulator Sickness Questionnaire (SSQ) [83] to assess the feasibility of immersive technology (Table 6). This questionnaire was administered before and after VR exposure. Saldana et al. [61] administered the SSQ questionnaire over two sessions and observed a decrease in the total score among the group using VR at the second assessment session. Indeed, the difference before and after exposure to the technology were −1.38 ± 2.29 at the first session and −0.25 ± 1.91 at the second session, indicating that fewer symptoms were present at the second visit. However, other studies [16,46,53,55] showed that, for a healthy population, the score averaged from 7.78 (2.39-16.45) before exposure to VR to 10.23 (1.36-15.21) after exposure to the technology, which, compared to populations with various health conditions, indicates an increase in the symptoms of discomfort related to the simulation. Across the papers addressing feasibility, while other works reported a decrease in the experienced side effects [42], opposite trends (an increase after immersion) were observed in [46].
Appel et al. [30] carried out VR testing in which the data were collected during pre/post-intervention. They concluded that there were no negative side effects to using the VR technology in the neurologically impaired population. Most of the participants had positive feedback and felt more relaxed, with a decrease in anxiety (1. With a home-built questionnaire evaluating the usability and engagement in AR, Bank et al. [72] reported a mean score of 69.3 ± 13.7 out of 100 for the usability section, indicating that the use of such technology was possible, and a mean score of 3.8 ± 0.5 out of 5 for engagement, which could be considered as moderate engagement. The ease of use and realism in manipulating objects are elements that may affect this sense of engagement. They concluded that AR is well tolerated and participants' augmented experiences were close to real experiences. Crosbie et al. [40] assessed the physical demands of using VR with the Borg scale [84], ranging from 0 to 10. The perceived exertion score in the virtual environment was 5.6 ± 2.22 in the stroke group and 1.6 ± 1.24 in the healthy group, indicating that performing tasks in the virtual environment appeared to be more difficult for people with a stroke than healthy people. However, both groups reported favorable experiences with VR, even though the stroke group faced greater physical demands with completing the same tasks.

N/A
The headset is suitable for most people, Precautions should be taken for people with head and neck pain. The helmet may be heavy for some users and cervical movements may create pain for those with cervical restrictions or those who are confined to a bed/chair. Vision problems may be an issue for some.

Handheld controller
Controller in the virtual environment was not aligned in the same direction as the one in reality.

N/A
The handheld controller may be difficult to use due to non-alignment with its position in real space.

Balance
Stability was an issue when moving, some with the sensation of head spinning or being too high. For some, feeling that there was too much movement around them in the virtual environment, as the helmet moved when actually moving. 2 participants experienced slight loss of balance.

N/A
Balance problems are possible even in people who do not have this problem. It is therefore an even greater issue for people with balance problems. The findings show that the IVR intervention is a feasible method to approach a personalized exercise program and an effective way by which to improve physical function in the target population.  The results of the presence questionnaire total score indicated a higher perception of presence in the strength endurance training than in the endurance training exergame. 5 Questionnaire "Motor and cognitive tasks were easy." "The 3D glasses were not uncomfortable." "The environment was beautiful." "The ergo-cycle was manageable." No nausea or discomfort related to the simulation. "It is difficult to recognize small animals" or "when they are from behind". The sound of the bike can be confused with auditory cues. One patient was tired before the end of the task.

N/A
The system has good usability. Several patients reported difficulty in recognizing animals that were too small or not facing the subject. Some confused similar animals. Some also had difficulty discriminating auditory cues from bicycle noise. A practice session prior to using the system would familiarize the participants with the environment and address these issues.

Effectiveness Meta-Analysis
There was sufficient data (n ≥ 4) to quantify the effect of VR and AR on older adults' balance and gait functions. Thirteen studies used instrumental measures to assess balance outcomes. These were the ABC scale, the Mini-BESTest, the Berg Balance Scale, the Tinetti Balance Test, the One-Leg Standing Balance Test, and the limits of stability (a posturography index). As underlined in Figure 2, AR and VR led to significant improvements in the balance function (SMD = 1.05; 95% CI: 0.75-1.36; p < 0.001) with a large magnitude of effect (SMD > 0.8). However, the heterogeneity between the studies was found to be moderate (I 2 = 42%). According to the GRADE approach, owing to the limited number of studies and sample size, the potential risk of bias, and the significant heterogeneity of these results, the strength of the body of evidence was decreased by two and therefore considered as low.

Studies Results
Kim et al. [46] used the Mini-BESTest to assess the balance in healthy participants and people with Parkinson's disease. For the healthy population, the pre-intervention balance score of 23 ± 4 changed to 25 ± 3 after experimentation with VR, while in participants with Parkinson's, the score increased from 21 ± 4 to 23 ± 4, with the change being statistically significant in each group (F (2,30) = 5.33, p < 0.05). They also reported a gait speed improvement after VR exposure. The participants walked significantly faster after expo- Fourteen studies used instrumental measures of gait outcomes. These were the gait speed, six-minute walk test, and Timed Up and Go test. As underlined in Figure 3, immersive technologies were found to significantly improve the outcome (SMD = 0.47; 95% CI: 0.14-0.80; p < 0.006). The effect size was considered as moderate (0.2 < SMD < 0.8) and the heterogeneity between the studies was found to be substantial (I 2 = 61%). Given the low number of included studies, the potential risk of bias, the moderate effect size, and the substantial heterogeneity, the certainty of evidence was considered as very low.

Studies Results
Kim et al. [46] used the Mini-BESTest to assess the balance in healthy participants and people with Parkinson's disease. For the healthy population, the pre-intervention balance score of 23 ± 4 changed to 25 ± 3 after experimentation with VR, while in participants with Parkinson's, the score increased from 21 ± 4 to 23 ± 4, with the change being statistically significant in each group (F (2,30) = 5.33, p < 0.05). They also reported a gait speed improvement after VR exposure. The participants walked significantly faster after exposure, from 1.08 ± 0.34 m/s to 1.12 ± 0.27 m/s and from 1.16 ± 0.18 m/s to 1.20 ± 0.18 m/s, respectively, for healthy adults and people with Parkinson's disease. Yoo et al. [81] also reported that AR contributed to a positive change in gait parameters, balance, and fall risk in older people after AR exposure.
Phu et al. [81] also investigated the gait speed in relation to the use of the BRU. This VR platform resulted in a significant 12% improvement in walking speed. The authors also observed a significant decrease in the risk of falling after the use of BRU. Indeed, the Falls Efficacy Scale-International (FES-I) post-exposure score decreased by 11.3 points,

Studies Results
Kim et al. [46] used the Mini-BESTest to assess the balance in healthy participants and people with Parkinson's disease. For the healthy population, the pre-intervention balance score of 23 ± 4 changed to 25 ± 3 after experimentation with VR, while in participants with Parkinson's, the score increased from 21 ± 4 to 23 ± 4, with the change being statistically significant in each group (F (2,30) = 5.33, p < 0.05). They also reported a gait speed improvement after VR exposure. The participants walked significantly faster after exposure, from 1.08 ± 0.34 m/s to 1.12 ± 0.27 m/s and from 1.16 ± 0.18 m/s to 1.20 ± 0.18 m/s, respectively, for healthy adults and people with Parkinson's disease. Yoo et al. [81] also reported that AR contributed to a positive change in gait parameters, balance, and fall risk in older people after AR exposure.
Phu et al. [81] also investigated the gait speed in relation to the use of the BRU. This VR platform resulted in a significant 12% improvement in walking speed. The authors also observed a significant decrease in the risk of falling after the use of BRU. Indeed, the Falls Efficacy Scale-International (FES-I) post-exposure score decreased by 11.3 points, while the Five Times Sit-to-Stand (FTSTS) showed a significant decrease of 26.69% in the time required to complete the five repetitions. These two results led to the conclusion that BRU might be effective at reducing the risk of falls in older people. Two other studies [44,53] used the Activities-specific Balance Confidence scale (ABC) to quantify balance in people with vestibular impairment [53] and stroke [44] and observed a significant improvement in the performance with the ABC mean scores changing from 62.54 ± 4.8 to 71.36 ± 4.24 [53]. Jung et al. [44] reported an improvement of 9.5% ± 6.0%. It appears that VR training can improve the perception of balance in people with health problems. The researchers also used the Timed Up and Go (TUG) test to evaluate the potential effects of VR. They reported a mean decrease of 2.7 ± 1.9 sec in the time to complete the test after exposure to VR, showing an improvement in gait balance.
Janeh et al. [42] used the GAIT-Rite system to analyze different walking parameters before and after the use of a VR device. The length of the shortest step increased from 58.34 ± 8.27 cm to 60.45 ± 8.16 cm after exposure, while the walking symmetry varied from 1.05 ± 0.04% to 1.01 ± 0.06%. In this study, the cadence before exposure to VR was 102.81 ± 8.19 steps/min and this changed to 97.41 ± 9.9 steps/min after exposure to VR. The cadence parameter was also used by Yoo et al. [81] to document the effects of AR. The cadence before exposure to AR was 100.79 ± 9.92 steps/min and this increased to 116.73 ± 8.81 steps/min after exposure, indicating an increase in the walking cadence. It can therefore be suggested that, unlike the immersive VR used by Janeh et al. [42], AR leads to an increase in the walking cadence. Yoo et al.'s study [81] also found a significant increase in the Berg Balance Scale scores (47.60 ± 5.36 before and 53.50 ± 2.30 after exposure to AR).
Benham et al. [32] used VR to address pain. The Numeric Pain Rating Scale (NPRS) score showed a significant decrease, with pain scores changing from 3.5 ± 1.73 to 0.9 ± 1.62 after exposure to VR. Their outcomes also included the World Health Organization Quality of Life Scale Brief Version (WHOQOL-BREF), where no effect was reported. In conclusion, we can note that there is a significant improvement in pain via the distraction provided by VR.
Furthermore, as presented in Table 7, some papers (VR [39,57], AR [72,76]) focused on the upper limbs. Phu et al. [57] investigated the grip strength and found that there was a significant improvement in the grip strength in the immersive VR users. Indeed, the BRU group reported a significant increase (p = 0.027) of 6.82% over the initial score [57]. Fischer et al. [76] used AR coupled with a pneumatic orthosis for the upper limb. This study reported a significant increase in the task performance on the Wolf Motor Function Test (WMFT), which was illustrated by a 12.9-point decrease (p = 0.02). However, they did not report a significant change in the biomechanical measures of hand or grip strength (p > 0.20) but reported that the AR would allow faster transitions between tasks and more opportunities to practice gripping objects that would not be available in the conventional clinical environment.
The decrease in the visual field has no impact on the gait pattern. The manipulated foot condition with visuo-proprioceptive dissociation was the most effective method to decrease the asymmetry of the gait pattern and to adjust the step length of both legs.

Discussion
This review summarized what is currently known about the use of immersive VR and AR technologies in older people. The following subsections discuss the results regarding the main research purpose, such as the acceptability, the feasibility, and the effectiveness of VR. We also highlight the limitations of the present study.

Acceptability
Our review identified 21 articles addressing the acceptability of immersive technology in older adults ( Table 5). The results emphasize that, when compared to conventional repetitive treatment, immersive technology allows for greater interest, enjoyment, and motivation [16,39,42,57,64]. In addition, different authors [30,39,64] reported that participants, namely older people with mild to moderate dementia, people with Parkinson's disease, or healthy older people had a pleasant experience with the VR. This can be explained by the feelings of relaxation and adventure that were present, as well as the reduction in anxiety, stress, and pain that was observed after exposure. This hypothesis is supported by recent studies demonstrating a stress and anxiety reduction among adults immersed into the VR environment [85,86]. Furthermore, a high level of interest and excitement about the VR technology before trying may have also contributed to these positive feelings reported after immersion. For instance, Appel et al. [30] and Brown [33] found that participants in their study wanted to use the immersive technology again in the future and would recommend it to a friend (Table 5). In most cases, the participants said that the headset they used (e.g., HTC Vive) was comfortable [64]. However, further studies are needed to confirm the acceptability of different types of immersive technology devices.
During the immersive experiences, some studies have focused on the environments that older people preferred to visit. Appel et al. [30] and Brown [33] showed that older people were interested in dynamic, social, and familiar real-world scenes (e.g., real places in the world, past or present). The authors suggested that the geriatric population would like to share these experiences with loved ones such as their grandchildren for narrative purposes or in order to explore places they no longer have the physical or psychological capacity to visit [33]. In addition to exploration and tourism, including mental relaxation, it should be noted that older people would also be open to other experiences with VR [64]. However, the environment in which a user is navigating significantly influences his or her desire to use VR [32].
Contrastingly, most commercial applications could be too complex and difficult to be used by older adults [39], especially for those with less experience with new technologies [33]. This may have decreased the acceptability of such devices in this population [64] and therefore may make further experiences less enjoyable. Moreover, there could be an increased feeling of isolation and loneliness for some people with physical or cognitive limitations [33]. Those feelings could subsequently promote depressive or anxious feelings and thus produce the opposite of the desired effect. Nevertheless, as suggested by Brown in their study [33], these concerns can be addressed with users prior to experimentation in an immersive environment.

Feasibility
The most commonly reported measure for determining the feasibility of a VR technology was the use of the Simulator Sickness Questionnaire (SSQ) [83]. The SSQ was developed to measure sickness that can occur when using VR technology (Table 6). It consists of side effects similar to those of motion-induced sickness [87]. These side effects may be caused by the visual conflict created by the immersive headset [32]. For example, after VR exposure, it has been reported that a side effect such as postural instability could significantly affect the Mini-BESTest score [46]. Moreover, owing to their medication or non-motor symptoms related to their condition [46], people with Parkinson's disease may have a higher score on the SSQ questionnaire even before immersion [42]. Thereby, the use of immersive technology can generate increasing variation in the participants' scores. However, some studies [46,55] showed that these changes were generally mild (with transient symptoms such as nausea, eye discomfort, disorientation, etc.) or not significant; although, in young and healthy adults, a longer duration of exposure seems to lead to more intense symptoms [88]. Further research would therefore be needed to generalize these observations to the geriatric population.
People with physical limitations may have to make more effort to succeed in virtual task completion, potentially limiting the treatment adherence and inducing some stress [40]. In addition, although dynamic activities such as walking seem to reduce the symptoms among healthy young adults [88], it is worth noting that for an older person performing walking movements during VR immersion, there is a higher risk of feeling stress [33,42,55,56]. Consequently, a familiarization period with the VR or AR equipment might be recommended prior to the interventions. This would ensure the comfort and feasibility of the experience and limit the unpleasant effects [33]. Moreover, the decrease in vision loss that occurs with ageing might be another barrier to the use of VR. Nevertheless, studies have provided recommendations for its use in people with vision loss [89,90]. First, VR applications should offer their users the possibility to modify the virtual visual field and light intensity according to their vision possibilities. Second, visual cues can be provided during the game to direct users' attention towards important information that would be displayed in their affected field of view. Lastly, the use of prism in AR should be considered to optically shift objects from outside the vision field.
The types of headsets used (Table 1) are essential during an immersive experience since they can have a great impact on the occurrence of side effects. Indeed, modern headsets such as the Oculus Rift or the HTC Vive can decrease the occurrence and severity of the side effects due to a better refresh rate, larger field-of-view, and better head tracking compared to older or lower quality immersive headsets [16,46]. This may result in less intense and transient symptoms. However, the use of controllers is challenging for older people, especially if they are not familiar with the new technologies [33]. Additional difficulties that can negatively affect the use of immersive technology comprise the controller's calibration and connection with the headset [33]. To overcome these difficulties, several systems have now developed hand-tracking technology, which allows for the use of a VR headset without using controllers. Indeed, hand-tracking enables one to generate a virtual model of hands and fingers into the VR environment by recording and identifying the movements of these body parts using infrared cameras. These methods have already been used and validated among patients with stroke and healthy older adults [91].

Effectiveness
The most important results of this systematic review also concern the effectiveness of the VR technology among the community-dwelling older adults ( Table 7). The results on the effectiveness can be summarized in three main aspects.
First, as shown in Table 7 and Figure 2, VR can be used to improve balance in older people and reduce the risk of falls. Indeed, this technology could achieve results similar to conventional exercises, but in half the time, with an intensity of 2 sessions of 30 min per week for 6 weeks in healthy subjects [57]. A significant improvement was also observed with the Mini-BESTest scores in people with Parkinson's disease [46], as well as the Activities-specific Balance Confidence (ABC) Scale and the TUG among participants with stroke [44]. VR can also promote a more personalized approach for the user, allowing for greater specificity in the treatment of balance deficits, thus improving gains and adherence [57]. In fact, the VR method proposed in [51,56] has been shown to significantly decrease anterior trunk rotations, keeping the center of mass within the base of stability and thus reducing the incidence of falls [92]. By using the Falls Efficacy Scale (FES-I) in a study involving healthy subjects, Phu et al. [57] showed that the use of VR led to a small but significant decrease in the fear of falling.
Second, VR can be used to correct the gait pattern. Indeed, thanks to the screen embedded in the headset, the participant's virtual foot appears to take a larger step in contrast to reality, exaggerating the decrease in the step length on the more affected side and thus forcing the user to take more symmetric steps on both sides [42]. Thus, the stance and swing times appeared to be more symmetrical after exposure to VR. This leads to the regularization of the cadence of the more affected side and a more symmetrical overall gait pattern [42]. However, larger randomized studies over a longer period are needed to confirm this effectiveness.
Third, Benham et al. [32] have shown a significant (p < 0.05) decrease in pain among participants after one session of VR. This decrease in the pain can be attributed to the distraction provided by the immersion. In fact, several studies have suggested that the immersive aspect of VR might be responsible for the reduced subjective experience of pain as the interaction with the real-world cues are being limited by the use of HMD [93]. This effect might be enhanced by the engaging, pleasant, and multisensorial feature of the immersive VR environment [93].
The effect of VR was also shown for other outcomes. VR resulted in fine motor skills improvements in a group of participants with Parkinson's and a slight improvement in the Unified Parkinson's Disease Rating Scale (UPDRS) [39]. An improvement in the grip strength was also observed in healthy subjects with the use of BRU [41]. Micarelli et al. [53] concluded that the addition of VR resulted in a significant improvement of the vestibuloocular reflex by increasing the frequency of visuo-vestibular conflicts. However, this study was conducted with a small sample of older people with mild cognitive impairments.

Perspectives
The results reported in this work provide several perspectives for the use of immersive technologies among older adults. While these technologies (VR and AR) are not yet implemented in our daily life, their increasing popularity, the decrease in their price, and their potential in terms of realism, interaction, and communication might reverse the situation. For instance, with the development of the metaverse, a parallel immersive virtual world intended to supplant the internet, it could be that rehabilitation is delivered remotely more often [94,95]. Such environments might also be of interest to improve older people's social participation as it allows for realistic multi-user interactions. Given such perspectives, we believe that, in the future, VR will be used as a mean to deliver effective remote rehabilitation to complement therapy and increase treatment adherence and intensity.
Moreover, VR devices (and metaverse development) also hold potential to deliver rehabilitation and promote activity among people who have no/few access to healthcare services, such as in low-income countries. Several studies have demonstrated the feasibility of implementing such interventions in developing countries [96,97].

Limitations
Despite the positive effects of VR reported in different studies and summarized here, the generalization of these results ais limited by the small number of articles available for each population and for each type of outcome, as discussed above. The overall low methodological quality of the articles included in this review potentially reduces the strength of the reported conclusions. Despite these limitations, this first systematic review shows encouraging results for further research and decisions in clinical settings.

Conclusions
Virtual reality is well accepted by older people and provides an enjoyable experience. The results suggest that the use among this population is feasible since few symptoms were reported and the increased SSQ scores were not significant in most cases. Currently, despite the several advantages described above, it is impossible to conclude on the effectiveness of VR in relation to different pathologies and deficiencies since few studies with good methodological quality and sufficiently large sample sizes are available. However, the beneficial effects have been observed regarding balance, risk of falling, and gait pattern in studies with acceptable methodological quality.

Data Availability Statement:
No new data were created in this review study. Data sharing is not applicable to this article.

Conflicts of Interest:
The authors declare no conflict of interest.