Non-Immersive Virtual Reality for Rehabilitation of the Older People: A Systematic Review into Efficacy and Effectiveness

Objective: the objective of this review is to analyze the advances in the field of rehabilitation through virtual reality, while taking into account non-immersive systems, as evidence have them shown to be highly accepted by older people, due to the lowest “cibersikness” symptomatology. Data sources: a systematic review of the literature was conducted in June 2019. The data were collected from Cochrane, Embase, Scopus, and PubMed databases, analyzing manuscripts and articles of the last 10 years. Study selection: we only included randomized controlled trials written in English aimed to study the use of the virtual reality in rehabilitation. We selected 10 studies, which were characterized by clinical heterogeneity. Data extraction: quality evaluation was performed based on the Physioterapy Evidence Database (PEDro) scale, suggested for evidence based review of stroke rehabilitation. Of 10 studies considered, eight were randomized controlled trials and the PEDro score ranged from four to a maximum of nine. Data synthesis: VR (Virtual Reality) creates artificial environments with the possibility of a patient interaction. This kind of experience leads to the development of cognitive and motor abilities, which usually positively affect the emotional state of the patient, increasing collaboration and compliance. Some recent studies have suggested that rehabilitation treatment interventions might be useful and effective in treating motor and cognitive symptoms in different neurological disorders, including traumatic brain injury, multiple sclerosis, and progressive supranuclear palsy. Conclusions: as it is shown by the numerous studies in the field, the application of VR has a positive impact on the rehabilitation of the most predominant geriatric syndromes. The level of realism of the virtual stimuli seems to have a crucial role in the training of cognitive abilities. Future research needs to improve study design by including larger samples, longitudinal designs, long term follow-ups, and different outcome measures, including functional and quality of life indexes, to better evaluate the clinical impact of this promising technology in healthy old subjects and in neurological patients.


Introduction
Virtual reality (VR) is a trending, widely accessible, contemporary technology of increasing utility to biomedical and health applications [1]. VR is the technological experience that allows for a full immersion in virtual spaces with which you can interact via specific wearable or using only your hand.
PubMed, and Science Direct databases, analyzing manuscripts and articles of the last 10 years (from June 2009 to June 2019), in order to obtain the latest evidence in the field.
Based on consultation with the multidisciplinary team, non-immersive VR studies and applications related to rehabilitation intervention were searched while using the following search terms, and the combination thereof: non-immersive, virtual reality, virtual game, rehabilitation, motor impairment, and cognitive impairment.
After the preliminary search, 26 articles resulted from PubMed, 19 from Scopus, 283 from Science Direct, 10 from Embase, and 11 from Cochrane.
The findings were analyzed and screened by four experts of the team, a bioengineer, a clinical neuropsychologist, a statistician, and a neurologist. In particular, three review authors independently reviewed titles and abstracts that were retrieved from the search in order to determine whether they met the predefined inclusion criteria. A fourth review author (a statistician) moderate any disagreement. The full text articles were subsequently analyzed.
The first screening was based on the analysis of the title and of the abstract, as well as deduplication of the findings. Another researcher confirmed the accuracy of the papers selection and screened for any possible omission. After the first step, 11 articles resulted from PubMed, two from Scopus, and 0 from Science Direct, Embase, and Cochrane.

Study Selection
We included RCTs and reviews written in English that aimed to study the use of non-immersive virtual reality in rehabilitation. Thus, we selected studies meeting the following criteria: 1.
Studies conducted on adult patients aged ≥65 years 2.
Studies devoted to use a non-immersive virtual reality in rehabilitation 3.
Studies including upper limb rehabilitation, lower limb rehabilitation, or cognitive rehabilitation 4.
Randomized clinical trials, with control group that received conventional rehabilitation therapy 5.
Before-after comparison of a single group 6.
Review articles On the contrary, we excluded studies that met the following criteria: 1. Conference proceedings 2.
Studies for which the full text was not found 3.
Studies written in languages other than English 4. Technical papers 5.
Qualitative studies All case-report studies and case-control studies were excluded for a lack of sustainability of results, as well as works concerning the development of new technologies.

Data Collection
After the screening based on the inclusion/exclusion criteria, conducted on the full text articles, the studies were selected as follows: 0 from Scopus, 10 from PubMed, and no one from Cochrane, Science Direct and Embase, and one from other sources. The countries of the selected studies are: Spain (2), France (2), Italy (2), Israel (1), United State of America (1), Canada (2), and Brazil (1). The fact that the studies have been performed in different countries shows that the topic is of general interest. Figure 1 shows the flowchart search strategy applied.
fact that the studies have been performed in different countries shows that the topic is of general interest. Figure 1 shows the flowchart search strategy applied.

Study Quality Evaluation
Quality evaluation was performed based on the PEDro scale and on the Cochrane's Risk of Bias (RoB) tool, suggested for evidence based review of rehabilitation while using non-immersive virtual reality [32,33]. The final score was settled when three authors reached agreement after repeated review and analysis. Of eight studies considered, five were randomized controlled trials and the PEDro score ranged from four to a maximum of nine, and the RoB score ranges from one to five (Table 1).

Study Quality Evaluation
Quality evaluation was performed based on the PEDro scale and on the Cochrane's Risk of Bias (RoB) tool, suggested for evidence based review of rehabilitation while using non-immersive virtual reality [32,33]. The final score was settled when three authors reached agreement after repeated review and analysis. Of eight studies considered, five were randomized controlled trials and the PEDro score ranged from four to a maximum of nine, and the RoB score ranges from one to five (Table 1).

General Characteristics of the Study Population
All of the studies were focused on older people with a mean age of 65.2 (±9.4) years for the experimental group and 69.2 (±9.4) years in the control group. The number of participants that were involved in all the studies is 1008 ranged from 6 to 376.
To our knowledge, of the older people involved in the trials, 586 were males and 392 females. The majority of the patients suffered from stroke (n = 593), followed by older people at high risk of falls, with more than two falls in six months (n = 182), patients with amyotrophic lateral sclerosis (n = 30), AD (n = 24), or PD (n = 24). Table 2 shows the characteristics of the studies. The outcome could not be pooled into meta-analysis due to the following reasons. Clinical heterogeneity ( Table 2) can be clearly observed from the participant, intervention, exercise mode, and outcome measures of the included studies. Diversity is seen in patient conditions and pathology, frequency, and duration of VR intervention, whether the impairment concerns the upper or the lower limb, whether the experiment conducted was pure VR (only VR) or VR mixed with traditional physical therapy or with exercise therapy, and whether the outcome measure contains follow-up.  Cochrane's Risk of bias tool [33] Sequence generation  -A partial body weight-support treadmill in conjunction with a television mounted on a stand in front of the treadmill to display the VR walkthrough environment.

or 3 training sessions per
week with partial body weight-supported tredmill with virtual reality system (total 12 training sessions). Initial training duration is 10 minutes; duration was progressed as tolerated.
-(1) FGA scores increased by 30% (2) BBS scores improved by 10% (3) Overground walking speed increased by 38% The virtual environment simulated a fully texture, medium-size kitchen. In the foreground, there was a work plane with all the objects needed to prepare a cup of coffee with milk and sugar.
Patients controlled the 2D cursor using a computer mouse.  To walk on a treadmill without virtual reality.
Experimental group increased SAI, reduced the number of falls, improved obstacle negotiation performance.

Intervention Effects
Eight papers report the results of clinical trials involving a group of patients that performed a training with VR system versus a control group that performed a traditional physiotherapy training [25][26][27][28][29]31], or a comparison within the same group performing a VR training while using different exercise control modalities [30].
The period of VR training ranged from four to six weeks, each day of the week or three training sessions per week, while the duration of each single session with VR system ranged from 15 to 60 minutes. Only in one study [27] the duration of intervention was two weeks. Generally, all the experimental groups (EGs) in the studies have received both therapies with VR and traditional physiotherapy, while the control groups (CGs) have only received traditional physiotherapy. Two studies have a follow up after eight and 12 weeks [29] or six months [28].
The study of Walker et al. [24] reported the lowest number of subjects (n = 6) within one-year post-stroke. It is a before-after study and all the subjects performed training with a treadmill equipped with a VR system. All the participants made significant improvements in their ability to walk, increasing the over ground walking speed and the Berg Balance Scale (BBS) scores.
The study of Turolla et al. [25] involves 376 post-stroke patients randomized into two groups, receiving combined VR and upper limb conventional therapy or traditional therapy alone. VR rehabilitation seems more effective than conventional interventions in restoring upper limb motor impairments and motor related functional abilities.
The study of Allain et al. [26] involves 24 Alzheimer's disease patients as compared with 32 healthy elderly controls on a task designed to assess their ability to prepare a virtual cup of coffee, comparing the performance with an identical daily living task. Significant relations are found between virtual and real coffee-making scores, and between virtual score and Instrumental activities of daily living (IADL) scale, which supports the validity of the virtual reality training.
In a study of Saposnik et al. [27], 141 post-stroke patients were randomized into two groups: the first received the VR therapy and the second received recreational therapy. The results show that within each group the performance time improves from baseline to the end of treatment, whereas no differences are found between groups.
The objective of Mirelman et al. [28] was to verify whether an intervention combining treadmill training with non-immersive virtual reality (to target both cognitive aspects and mobility) would lead to fewer falls than treadmill training alone would. To do this, the authors recruited 282 older people at high risk of falls and randomized them into two groups to receive treadmill training plus VR or treadmill training alone. In the six months after training, the incident rate was significantly lower in the experimental group.
The study of Segura-Ortì et al. [29] involves 18 patients on hemodialysis: nine performed 30 minutes of non-immersive virtual reality training and nine performed 30 minutes of aerobic training. Both interventions improved physical function, such as gait speed and no significant differences, were found between groups.
In the study of Trevizan et al. [30], the performance on a computer task in patients with amyotrophic lateral sclerosis while using three (motion tracking, finger motion control, or touch screen) different commonly used non-immersive devices was evaluated. The control and experimental group both showed better performance on the computer task when using the touch screen device.
Pelosin et al. [31] analyzed 39 patients with Parkinson's disease, assigned to treadmill training group or treadmill training with non-immersive virtual reality intervention group to assess cholinergic activity. The results showed that the experimental group improved obstacle negotiation performance, and reduced the number of falls as compared with control group.

Discussion
A review of the evidence on VR efficacy in patients affected by a neurological disease is mandatory due to the rapid development of VR programs in the last years and the increasing literature on VR application in neurological conditions, in order to enable clinicians to have an up-to-date understanding of the potential clinical beneficial effects of these techniques.
Therefore, the aim of this paper was to systematically evaluate the evidence of the effectiveness of VR compared to conventional therapies. It must be stressed that few studies summarize the current best evidence on the effectiveness, user compliance, feasibility, and safety of VR interventions for rehabilitation treatment in neurological disorders.
VR creates artificial environments with the possibility of a patient interaction. This kind of experience leads to the development of cognitive and motor abilities, which usually positively affect the emotional state of the patient, increasing collaboration and compliance.
Moreover, the VR rehabilitative treatment might be personalized according to the specific abilities and needs of the subject.
Parkinson disease is one of the most common age-related brain disorders with both dopamine-related motor symptoms and nonmotor symptoms due to other neurotransmitter circuits involvement, such as the cholinergic, noradrenergic, and serotonergic pathways.
The cognitive decline is among the most common and relevant nonmotor symptoms in PD and it affects different cognitive domains, in particular attentional, visuospatial and executive domains, and also memory. VR in the cognitive PD treatment could be useful in improving, in particular, the visuospatial and executive abilities, which represent the most compromised aspects of cognitive decline in PD patients [7].
Moreover, falls are frequent in ageing and PD patients, due to an impairment in the cholinergic-mediated gait pathway. A rehabilitation approach using treadmill training combined with non-immersive VR seems to induce changes in cortical cholinergic activity, which enables functional gait improvements and reduces the fall rate in comparison to a traditional rehabilitation method [31].
A reduction in static and dynamic balance is a major risk factor for falls also in stroke survivors [34]. In fact, the majority of the individuals with stroke who have fallen usually develop fear of falling again (88%). Fear of falling is related to balance and gait deficits [35], and it often leads to reduced physical activity and deconditioning. In fact, 44% of stroke fallers report restriction of activity after the fall. Given the very low physical activity and cardiovascular fitness levels already near the lower limit of those required for basic ADL, further activity reduction and deconditioning due to the fear of falling can easily lead to a loss of independence in individuals with stroke.
A recent review on post-stroke rehabilitation therapy [6] provided evidence for a moderate beneficial effect in balance improvement of VR combined with conventional therapy, as compared to conventional therapy alone.
More promising effects seem to be evident in the case of upper limb motor impairments in stroke rehabilitation [25], but further studies are needed on this subject. In fact, the trial described in [27] found that non-immersive VR as an add-on therapy to conventional rehabilitation was not superior to a recreational activity intervention in improving motor function, which suggested that the added intensity of training only induces early motor recovery of the upper limb, and that this can be achieved with VR or with other simple and inexpensive arm activities.
VR technology has considerable potential for detecting functional limitations in IADL performance in AD patients, beyond that of current neuropsychological measures, as shown by Allain et al. [26]. Moreover, studies were carried out to assess the effectiveness of a VR cognitive training program on cognition in mild cognitive impairment (MCI) and AD patients [36].
VR cognitive training for individuals with MCI and dementia has proven to result in improvements in the cognitive domains of attention, executive function, and visual and verbal memory. Moreover, significant reductions in depressive symptoms and anxiety were evident, with a delay in the progression of cognitive impairment [37].
Additionally, the VR format might help in training adherence, as individuals with MCI and dementia patients seem to prefer the VR format of a task over the paper version, as confirmed by a feasibility study with image-based rendered VR in patients with mild cognitive impairment and dementia [38].
Some recent studies have suggested that rehabilitation treatment interventions might be useful and effective in treating motor and cognitive symptoms in different neurological disorders, including traumatic brain injury [8], multiple sclerosis [9], and progressive supranuclear palsy [39].
Finally, VR represents an effective tool that could improve the traditional cognitive and motor rehabilitation in patients that are affected by a neurological disease. Moreover, home-based VR might offer a promising addition or alternative to existing rehabilitation programs, and a chance to provide and/or prolong the required therapy after discharge in a more accessible setting, potentially improving clinical outcomes.
Future research needs to improve the study design by including larger samples, longitudinal designs, long term follow-ups, and different outcome measures, including functional and quality of life indexes, to better evaluate the clinical impact of this promising technology in healthy old subjects and in neurological patients. In particular, the next challenge for the research on VR and rehabilitation can be summarized in the following questions: 1. Does an innovative intervention enriched with VR provide a significant improvement in mobility, compared to traditional physiotherapy? 2. Is the intervention cost-effective for the health management systems?
To answer to these questions, it is crucial to understand how to improve the rehabilitation path of older people, through multidisciplinary multicomponent and person-centered intervention, integrated with VR.
The evidence reported in the paper are in line with the aims that were expressed by the National Plan for Health Research, whose priorities are defined in accordance with the indications contained in the regulation of the European Parliament and of the Council on the establishment of the "Health for growth" program, which pursues as a goal the achievement of a strong potential for economic growth thanks to the improvement of the state of health, through the facilitation of innovation in health care, the improvement of skills and information on specific diseases, and the identification of good practices for effective prevention [40]. In line with what has been expressed, the role of health technology assessment is of crucial importance. In this perspective, the available services must necessarily be enriched with adequate equipment of proven efficacy, as the promising sector of VR, to be able to advance at both the methodological and assistance level.

Conclusions
As it is shown by the numerous studies in the field, the application of VR has a positive impact on the rehabilitation of the most predominant geriatric syndromes. The level of realism of the virtual stimuli seems to have a crucial role in the training of the cognitive abilities. Nevertheless, semi-immersive or non-immersive VR systems have the advantage of being more accepted by the users, as they experienced less cybersickness after the training. Moreover, the integration of these devices in the health management systems are still lacking despite the evidence and the peculiarity of VR technologies with different level of immersivity. A tentative explanation can be found, not only in the cost of technology that seems to be more affordable in the recent years, but, most of all, in the absence of a standardized protocol and procedure, to harmonize traditional rehabilitation therapies and innovative VR systems. For this reason, it will be necessary to improve the research in the field, adopting RCTs study design as well as indicators of health technology assessment, to understand the effectiveness and efficacy also in terms of optimization of the clinical pathways. In addition, as VR systems can be easily adopted at home, it can be considered to be useful for the continuity of care.