The aim of this review was to evaluate and guide the clinical applicability of eHealth devices in the screening, assessment, and treatment of elderly people with balance disorders but without neurological disease.

This systematic review was performed in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) statement [48] with the following steps: development of research questions, development of a search strategy with eligibility criteria, data selection, and qualitative analysis.

The protocol for this systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD42019120774) and the complete protocol is available on the National Institute for Health Research program website.

This review focused on answering the following specific research questions, according to the PICO (Population, patient, or problem; Intervention; Control, Comparison, or Comparator; Outcome) strategy [49] (Textbox 1): (1) What are the main contributions of eHealth to elderly people with balance disorders with risk of falling? and (2) Is there any evidence that eHealth improves the quality of health care services in this context? If not, what are the reasons?

In this study, we defined elderly people as those over 60 years of age [50]. Knudson [51] defined balance as a “person’s ability to control their body position relative to some base of support.” According to Agrawal et al [6], vertigo and dizziness are defined as “sensation of self-motion when no self-motion is occurring or the sensation of distorted self-motion during an otherwise normal head movement” and “sensation of disturbed or impaired spatial orientation without a false or distorted sense of motion,” respectively. Falls refer to “inadvertently landing on the ground, floor or other lower level” [10]. Gait is defined as “the pattern of movement of the body during locomotion” [52].

Telemedicine is defined according to the World Health Organization Group Consultation on Health Telematics [53] as “delivery of health care services using ICT for the exchange of valid information for diagnosis, treatment and prevention of disease and injuries, research and evaluation, and for the continuing education of health care providers.” eHealth is defined according to Eysenbach [54] as:

Teleconsultation is defined as “synchronous or asynchronous consultation using ICT to omit geographical and functional distance” [55]. Finally, a sensor is defined as a “device that responds to a physical input of interest with a recordable functionally related output that is usually electrical or optical” [56].

Articles were retrieved through searching the PubMed, Google Scholar, Embase, and SciELO databases. The search algorithm included multiple group combinations, as shown in Table 1.

The inclusion criteria were randomized controlled trials (RCTs) or quasiexperimental studies (QESs) published in English between January 2015 and December 2019, studies related to use of eHealth in the context of balance and falls, and the sample was restricted to an elderly population (60 years old and above).

The exclusion criteria were: (1) review articles, brief reports, protocols, proof-of-concepts, pilot studies, conference papers, and letters to the editor; (2) studies including elderly people with a reported functional limitation due to a neurological disease; and (3) articles without an age sample reference or with participants aged below 60 years.

First, both authors screened the papers independently, looking at titles, abstracts, and methods, and agreed about their inclusion or exclusion according to the eligibility criteria. Second, the potentially relevant papers were retrieved for full-text evaluation against the eligibility criteria. Any articles that were deemed to be questionable in the first stage were included for further evaluation in the second stage. The selection of papers was performed by checking the extracted data and risk of bias.

The main outcomes included population characteristics, balance disorder, identification of faller participants, eHealth platform and services, health benefits, and fall prevention literacy.

The quality of the evidence to respond to the research questions was independently assessed using the Joanna Briggs Institute (JBI) Critical Appraisal for Experimental Studies and JBI Critical Appraisal Checklist for Quasi-Experimental Studies tools [57]. The two researchers discussed the results of the quality appraisal, reaching a consensus in case of any divergence. The included RCTs were assessed using the Cochrane risk of bias tool [58] to evaluate the risk of internal bias for a series of domains: selection bias, performance bias, detection bias, attrition bias, and reporting bias. Disagreements were solved by consensus between the two researchers.

We provide a narrative synthesis of the main outcomes from the included studies. First, the articles were categorized according to the study design. Second, the articles were categorized based on the focus of eHealth services (screening/assessment and treatment/rehabilitation) for comparison of clinical use and applicability according to digital devices.

A total of 1030 unduplicated articles were identified, 984 of which were excluded after title and abstract screening. Among the 46 full-text publications assessed for eligibility, 25 articles were excluded owing to functional limitations due to neurological disease (n=4), age of participants (n=5), focus on technological implementation (n=2) or model/algorithm (n=8), and specific descriptions of elderly gait parameters (n=6).

Twenty-one articles [59-79] were ultimately included in the review (Figure 1).

The 12 QESs were focused on screening and assessment. These studies compared the use of sensors in participants with a history of falling, or with clinical tests and functional scales [66,68-77,79] (Table 5). No instrumental test of the inner ear was performed as a gold-standard test. All RCTs [59-65] and two QESs [67,78] were focused on balance treatment or rehabilitation (Table 6). Again, no instrumental test of the inner ear was used in pre and postrehabilitation assessments.

The included studies had large differences in sample sizes, ranging from 23 to 153 participants [59-65] among RCTs and from 16 to 196 participants [66-79] among QESs. As shown in Tables 5 and 6, many studies included a small sample size that was described as a limitation.

The age range was 60-91 years for the RCTs and 60-92 years for the QESs. Most of the studies included more women than men. In four studies, only women participated [65,67,71,76]. The decision to only recruit women was explained in one study as “to avoid the influence of gender differences on risk of falling” [76]. Two studies excluded the few male participants [67,71] and the remaining article did not describe the reason for the exclusive participation of women [65]. One study did not describe the age range or the gender distribution of the participants [79].

The participants (≥60 years old) were recruited from the community [60-62,64,65,67-70,72-74,76,77], gerontology services [71,75,78], both [59,79], or at a hospital [66]. One study included participants who were referred for geriatric inpatient rehabilitation [63].

Most of the studies did not describe the characteristics of health conditions of the sample [59,62,64-66,68-72,74-77,79]. Only some authors provided quantitative data about the participants’ medication use [60,61,73,78] and their comorbidities [60,61,63,67,73,78]. One study [60] highlighted the following comorbidities of the participants: heart problems, high blood pressure, osteoporosis, lower back pain, hip pain, knee and/or leg pain, and foot pain. Two studies excluded participants with self-reported balance disorders [62,63]. Two other studies included frail or prefrail elderly adults [64,75].

The included studies used functional balance tests, with or without sensors, to evaluate balance and risk of falling. An objective identification via exploration and quantification of the function of the inner ear by instrumental tests was not employed in any of the considered studies, as mentioned above. Therefore, the presence of prebyvestibulopathy or other balance disorders was not known.

Some authors highlighted the potential of sensor-based tests in identifying early balance deficits [71] and in evaluating the risk of falling [66,72,73,76,77]. Improved balance and gait with technology-based training were mentioned in some studies [59,62,64].

The identification of faller participants based on retrospective [68,70,73,76] or prospective occurrence of falling [74] was employed to compare the technology results. The benefits of virtual training in reducing the risk of falling was also described [60,61,65,78].

No study focusing on detection of falling fully complied with the inclusion criteria of this review (RCTs or QESs, published in English between January 2015 and December 2019, restricted to the population 60 years or older).

Different platforms were used for the provision of eHealth services. The main platforms identified were computer-based apps, either via the internet or mobile based platforms (Tables 5 and 6).

As mentioned above, 12 studies focused on screening or assessment [66,68-77,79] using different types (wearable inertial, wearable pressure, pendant, smartphone), quantities (range 1-5), and locations (head, sternum, lumbar vertebra, pelvis, hip, leg, shanks, foot) of sensors. Single or combined sensor-based tasks were employed. Only two studies [73,75] evaluated activities in a real-life environment (Table 5).

Nine studies explored balance rehabilitation [59-65,67,78] with different exercises and duration of training. The follow-up time was short (less than 6 months) in most of the studies, with the longest follow-up of 1 year [59]. The development of eHealth services was explored both inside and outside the laboratory environment (Table 6).

One study used a robot to provide information about training and fall prevention. However, the authors pointed out that the screen and the volume speaker were not adequate for use by elderly people [78].

The use of technical language and the presence of disabilities such as visual and hearing impairment were highlighted as the main barriers in using eHealth [78].

Only one study did not report better adherence, enjoyment, motivation, and balance performance with virtual training. This was explained by the possible fragility of the sample included in the study and by the short duration of the training intervention [63].

The remaining papers emphasized the potential contribution of digital solutions to improve balance performance and risk of falling. The sensors used during balance tests improved the evaluation of balance and gait [66,68,69,71,72,79] and improved the identification of potential faller participants [70,73-77]. In addition, the use of eHealth devices for balance rehabilitation increased balance and gait performance [59,60,62,64,65,78], and reduced the risk of falling [60,61]. However, no long-term follow-up was reported. Virtual programs of falls prevention seemed to increase knowledge on the subject [78] (Table 7).

None of the studies explored the previous health literacy of the participants. Only two papers described the educational level of the participants [60,78].

One study compared use of a fall prevention software to a conventional handbook to evaluate the improvement of knowledge on fall prevention. Both the intervention and control groups showed improvement in knowledge, without a significant difference [78].

Population aging, and the associated vulnerability to the development of multiple chronic pathologies and balance disorders, have motivated research and the implementation of new strategies for the provision of health care. eHealth devices have been studied to help assess balance and gait performance, risk of falling in and outside a laboratory setting, and to perform in-home balance rehabilitation. In this review, we confirmed the potential of eHealth to complement the health care of elderly people. However, most of these studies were not designed to provide clinical guidelines.

Despite growing interest about this subject in the last 20 years, we decided to focus on studies published in the last 5 years (RCTs and QESs), taking into consideration both continuous advances in technological innovation and the opportunity to apply new clinical applications in balance disorder and risk of falling for the elderly population.

Unlike other reviews, our eligibility criteria ruled out many initially retrieved articles, especially studies with participants under 60 years old, those without reporting the age of participants, or with participants having a functional limitation due to neurologic disease. Therefore, only 21 articles fully complied with the requirements of this review [59-79].

Except for one study [63], the others showed the potential of eHealth to evaluate balance assessment and risk of falling of elderly people and to promote balance training. The eHealth devices allowed collecting additional information about the balance, gait, and risk of falling of elderly people, and to monitor their daily activities.

In particular, eHealth seems to provide an opportunity for increasing medical-patient interactions and to reduce access inequities [30]. In 1996, Viierre et al [80] had already mentioned the potential of eHealth in this field: “remote medical diagnosis and treatment facilities could make the few vestibular disorder specialists much more available to patients.” However, as observed in other reviews [36,37,42-47], the differences in methodologies and of variables included in the studies did not allow for a proper comparison to guide clinical applicability.

First, there was a broad range of sample sizes, which were generally quite small (ranging from 16 to 196 participants). A small sample of participants is considered a limitation for extrapolating the results, especially for the exploration of risk of falling.

Second, there were missing data about the clinical conditions of the participants. Except for one study [60], several volunteers were recruited from the community and were defined as “healthy” elderly people only based on a self-reported assessment. There were also participants recruited from geriatric services without reference to their clinical conditions. Despite the exclusion of participants with self-reported balance disorders in two studies [62,63], we consider that the exclusion rules should be more rigorous and based on objective data such as instrumental inner ear tests. We have to take into consideration that elderly people can have instability due to many conditions, including the normative aging process, and therefore the outcomes from a balance rehabilitation intervention could be sensitive to these differences [38-41].

Third, different research methodologies were used for screening and assessment. We observed a wide range of clinical tests, functional scales, faller classifications, and sensor-based tasks among the included studies. The lack of homogeneity of these variables limited an appropriate comparison among the studies. Moreover, functional inner ear tests were not used as the gold-standard test. We consider this as a weakness common to all studies.

Fourth, different types of sensors were used for screening and assessment. Similar to the findings of other reviews [36,37,44-47], the studies employed mainly accelerometers, with variations in both number and body location.

Fifth, as observed previously [42,43], studies focusing on treatment and rehabilitation used different devices, training durations, and follow-up times. Some authors employed supervised training. In one study, this was used a telepresence-based exercise platform [65]. Others employed in-home self-regulated exercises training [61], thereby avoiding the need for participants to travel to the rehabilitation center. None of the studies described pre and postintervention data about the function of the inner ear. The studies did not verify the long-term effect of training, especially with respect to fall occurrence. Only two studies explored a sensor used in real-life activities [73,75], which is relevant since it allowed for a better evaluation of the remote interaction and monitoring of daily activities.

Additionally, we observed a constraint related to the use of devices that are not fully adequate to match the abilities of elderly people [78]. We also highlight the importance of providing a better definition of the eHealth user profile to improve adherence.

Future studies in this field should consider the above topics as a starting point, as well as for health policy implementations on eHealth apps for elderly people with balance disorders.

The use of eHealth can play an important role as a complementary method to provide health care services, encouraging health promotion and patient participation, as well as allowing for the remote management of balance disorders.

Based on this review, we can provide the following recommendations to improve studies and applications of eHealth for preventing fall risk in the elderly population.

First, this review highlights the need for further research on the use of eHealth devices in proper clinical settings. This represents an opportunity to be explored, reaching out to elderly people with balance and risk of falling.

Second, despite several efforts to explore balance among the elderly, there is still a need for better characterization and description of the health condition of the population under study. In particular, we recommend future studies to include the results of functional tests of the inner ear as a gold-standard test or for comparison of the outcome before and after remote balance rehabilitation. Most of the interventions were developed with only functional balance tests. Future studies should also focus on the real-life environment, allowing for additional information of the daily activities among elderly participants.

Third, a longer follow-up time is important to evaluate the long-term benefits of eHealth tools on the balance performance and risk of falling of elderly people.

Finally, the eHealth devices should be user-friendly to improve adherence among elderly people.

This review was limited to articles written in the English language and available on the PubMed, Google Scholar, Embase, and SciELO databases for the last 5 years; therefore, it is possible that relevant studies were missed.

The inclusion of eHealth services can play a critical role for the better provision of health care to elderly people with a balance disorder and risk of falling. The differences in populations, methodologies, eHealth devices, and follow-up times of the included studies did not allow for a clear comparison between results, therefore limiting the possibility of obtaining valid guidance for clinical applicability. More rigorous studies are recommended.