Gait is one of the most frequently assessed attributes in clinical settings because gait impairment is the hallmark of several different health conditions [1,2]. Prevalence of poor gait and gait-associated impairments is on the rise because people are living longer, with multimorbidity of chronic conditions such as obesity, diabetes, and arthritis, and because of a global increase in the proportion of neurological conditions [3-5]. Gait impairments and walking limitations from aging, disease, or injury increase the risk of falls, joint damage, and a sedentary lifestyle, leading to a vicious cycle toward further deterioration [6,7]. To illustrate, gait and balance impairments have been shown to increase with age—from 10% among individuals aged between 60 and 69 years to >60% in individuals aged >80 years [8]. Gait and mobility challenges are among the main concerns for older adults and people with neurological conditions and are a major reason for seeking rehabilitation services. There is considerable evidence to support the effectiveness of gait training in older adults and people with neurological conditions [9-15]. Although traditional therapist-led gait training strategies are well-accepted and effective in improving gait patterns, the effects abate with cessation of training [16,17]. Hence, traditional therapist-led gait training alone will not translate into sustained neuromuscular change nor lead to the behavioral change needed for physical activity guidelines to be met.

Effective and accessible treatments for gait impairments will increasingly be needed with the aging of the population and as people with health conditions live longer. Skilled therapy professionals are a limited resource, and therapy is rationed; furthermore, rehabilitation is a global target of the World Health Organization’s 2030 strategy, with key areas for action to increase affordable services and use technology to address this need to assess and reassess how individuals mobilize and move and implement long-term training programs [18]. Increasingly, people with gait vulnerabilities and their family members will turn to technological solutions to supplement and extend rehabilitation services [19,20]. Technological innovations are poised to close the gap between demand and supply [21]. There is no doubt that older adults and people living with health conditions would benefit from focused gait training beyond what is offered during a clinical visit [22,23]. Technology can provide people with opportunities to practice gait-related skills outside the clinical environment and gain ownership over their therapy [24]. There is evidence to support that technology alone can influence positive behavior and that smartphone apps have been shown to reduce sedentary time by 41 minutes per day [25,26]. These effects are thought to be a result of the user’s ability to self-monitor and self-correct, thus providing them with more control and responsibility for their own therapy [26]. Given the unmet need for access to rehabilitation services and the need to continue therapy outside clinical settings, the commercialization of technology is timely and necessary.

Available devices range in sophistication from nonelectronic shoe insoles and walking aids to inertial or pressure sensors. Most of the technologies used have gait assessment functionality, but there is now increasing interest in harnessing the capacity of wearable sensors for providing biofeedback. The literature is rich in supporting the effectiveness of biofeedback in improving gait patterns in healthy and clinical populations [27-30].

There is an increasing number of devices that claim to improve gait impairments through biofeedback. However, it is still rare for these devices to be available to the consumer; most are still tied to a laboratory setting. There is an urgent need to move technological innovations from research laboratories to the people who would benefit the most—those with gait impairments. The COVID-19 pandemic has alerted us to the vulnerability of seniors and people living with chronic health conditions when they were no longer able to access clinical and community resources [31-37]. In addition, the growing size of the older population means that one-on-one treatment will no longer be feasible, and a self-management strategy facilitated by technologies will be needed [31-33,38].

The market of people needing gait training technologies is huge. As a result of direct access of the general public to several technologies, the impact of evidence presented on the websites could affect purchasing behavior. A study on the purchasing intention of consumers who shop on the web found that “high involvement” consumers, defined as people living with health conditions who need to improve their gait to meet functional demands or mobility needs and are intently looking to purchase something specific, were more likely to purchase a product if the number of quality reviews was high [39]. Individuals with gait impairments may be considered “high involvement” consumers and, therefore, may purchase related products based solely on available reviews that may or may not have evidenced research quality.

Gait training technologies must be appealing with features such as product attractiveness, functionality, and price, as well as be supported by robust research demonstrating usability, reliability, efficacy, and effectiveness. All these features are equally important; an attractive product that does not work or a product that does work but is expensive would not be appealing. Although the attractiveness of technologies is often featured on websites, the evidence for efficacy could be hard to locate. Furthermore, the public is not likely to have access, time, or training to find the scientific literature and critically appraise the content to guide the decision to purchase such products.

The objectives of this customer-centered review were to create an inventory of devices targeting improvement in gait pattern and walking behavior and identify the strength of the evidence underlying the claims of effectiveness for devices that are commercially available to the general public [40].

As there is no systematic or reproducible way to identify technologies available to the public to help improve gait, we used a pragmatic and iterative approach. Our search strategy involved a search of gray literature as well as published literature. Figure 1 shows the 4 approaches used to identify a list of commercially available biofeedback devices. We used simple words, including those suggested by laypersons. Our focus was on devices that provided feedback, but this would not be thought of by the consumer. Therefore, we supplemented this strategy by searching for condition-specific organizations or charities because they might endorse such devices. This search yielded 2 feedback devices. We also performed a search using clinical impairment–specific search terms, and this yielded another 15 feedback devices. Finally, we searched for systematic reviews covering gait but found no new devices [41-45]. The search was first conducted in October 2021 and repeated in December 2022. Once we had a list of devices, we searched for evidence of efficacy published on the device web page as well as on PubMed and Google Scholar using the device name to search.

A findable list of technological devices targeted to health conditions was extracted separately by 3 authors and compared for completeness. For each device identified, the evidence for efficacy or effectiveness was extracted from material displayed on the websites, and full-text articles were obtained from the scientific databases PubMed, Ovid MEDLINE, Scopus, or Google Scholar. This step was carried out by MM, AA, MW, OS, SG, DG, and HD; any conflicts were resolved in consultation with KM and AA-S. Finally, KM and NEM organized the results into tables and reverified all data and assigned levels of evidence. A level of evidence was assigned to each study involving the device using the Oxford Centre for Evidence-Based Medicine scale [46,47].

The levels of evidence rating system is a method of quantifying the best clinical evidence that is available about the efficacy and safety of treatment approaches that are destined to be implemented in clinical care [46,47]. The Oxford Centre for Evidence-Based Medicine scale was used because it provides the best granularity of evidence arising from the majority of trials of new technologies that usually are not included in meta-analyses and do not have randomized clinical trials with large sample sizes producing narrow CIs. Multimedia Appendix 1 shows the Oxford Center for Evidence-Based Medicine levels of evidence.

Information on the target population, mechanism of feedback, evidence for efficacy or effectiveness, and commercial availability was obtained from the published material or websites. Only devices that claimed gait rehabilitation or gait quality improvement through any sensory feedback mode—visual, auditory, haptic (tactile or kinesthetic), or vibration—were included. The devices were excluded if the technology was not targeted to any health condition or if it targeted high-functioning populations such as athletes or healthy individuals. The term feedback is defined as a physiological or performance signal arising as a result of human movement that, in turn, generates an output (error or correct performance) that is relayed back to the user and that has the potential to modulate (enhance or diminish) subsequent movement.

The search yielded 17 wearable devices that claimed to target improvement in gait quality through various types of feedback: 11 (65%) were commercially available, and 6 (35%) were at various stages of research and development. Of the 11 commercially available devices, 2 (18%) were sold under the trademark WalkWithPath: Path Finder Laser Shoes and Pathfeel. The inclusion of the devices was appraised by KM and NEM. Table 1 presents a brief description of the devices (grouped into insoles, wearable sensors, and vests or walking aids), feedback type, target condition, and components. The devices are organized according to availability: directly available to clients or only for research purposes and thus not commercially available. Among the 11 commercially available devices, there were 6 (55%) insoles, 4 (36%) wearable sensors, 1 (9%) vest, and 1 (9%) walking aid; for example, BalancePro insoles, which provide only passive sensory feedback, and FeetMe insoles, which have embedded sensors to provide different types of feedback, including electrical stimulation or an auditory signal. All technologies used a variety of biofeedback (positive, negative, and continuous) and offered options for choosing or providing a single preset sensory stimulus—auditory, haptic, visual, or vibration—enabling users to set individual preferences. A few of the devices offered practitioners and consumers a choice to select the feedback frequency and type of stimuli. Regarding choosing the type of sensory stimulus, there is no information available on the efficacy of one sensory stimulus compared with that of another. Most devices target gait improvement for people with neurological conditions, specifically people with Parkinson disease.

Textbox 1 presents information on population, intervention, control, outcomes, time, training, results, usability, and level of evidence with study design. Of the 11 commercially available devices, 4 (36%) have published evidence of efficacy reported in 10 studies with sample sizes ranging from 6 to 40: CuPiD/Gait Tutor, BalancePRO, Heel2Toe, and WalkWithPath [58,65-67]; for example, the BalancePro insoles are plastic insoles with a raised ridge around the perimeter that provide continuous haptic feedback and are targeted to people with Parkinson disease and older adults. The insoles are available for direct purchase on the company website, and the design patent application is under review. The evidence supporting the BalancePro technology comes from 2 crossover study designs and 1 randomized controlled trial, all at level 2b of evidence using the Oxford Center for Evidence-Based Medicine scale.

Textbox 2 outlines some important areas that would help judge the usefulness of technologies targeting gait from the perspective of consumers. These areas emerged from this review because the needed information was either absent from the papers or inconsistently presented. The list of technology-relevant items presented in Textbox 2 would be applicable for inventors publishing in the scientific literature.

This review identified a total of 17 wearable biofeedback devices targeting gait patterns and walking behavior. Of these 17 devices, 11 (65%) are commercially available to the public and have a dedicated website for direct purchase. Of these 11 devices, 4 (36%) had published evidence on effectiveness at level 2b according to the Oxford Centre for Evidence-Based Medicine scale (Textbox 1). There was no searchable evidence available for the efficacy or effectiveness of the feedback from the remaining gait training technologies (7/11, 64%). Evidence is primarily generated for 1 health condition, but the claims are generalized to other health conditions with similar gait impairments. There was limited to no data available on accuracy, reliability, usability, and safety. Almost all websites presented user reviews or testimonials, which are likely to be selective in favor of supporting the technology. It is important for clinicians to be aware that some scientific evidence supporting the technology may exist, but a consumer is most likely unable to access the published material. A consumer may be driven to purchase a device or not merely by reading reviews or testimonials.

This review provides a summary of commercial wearable gait training technologies that are currently available in the market or the development phase. A unique feature of this review is that it was conducted from a consumer’s perspective and then augmented by summarizing the evidence from scientific publications. Although the strength of the evidence supporting the effectiveness of these technologies is low or moderate at best, the claims on the website often outweigh the evidence. The results of our review can also be used by professionals involved in gait rehabilitation to direct their clients to promising technologies based on available evidence. These technologies can also be incorporated into treatment plans.

Several of the papers (9/10, 90%) that contributed evidence toward the efficacy and effectiveness of the wearable sensors failed to capture or report patient-centered outcomes or declared level of evidence. In summary, the quality of evidence was low. Only 36% (4/11) of the devices had searchable evidence for efficacy potential, with all studies being small-sample sized (Textbox 1). This calls into question the strength of the evidence and the generalizability of the findings outside the study population. Although the mechanism of action and information on spatiotemporal gait parameters have been reported for all devices, it is important to provide information on walking speed, distance, physical function, and walking behaviors such as step count or walking bouts. Overall, the approach to statistical analysis is rudimentary, and inference is mainly based on within-group P values rather than CIs. Lack of raw data in the published manuscripts, such as mean, median, SD, and range, prevented a calculation of between-group effects, effect sizes, and reliable change among other metrics that can potentially provide more interpretable information. Sample sizes are typically small, leading to a high degree of uncertainty in the results. Very few papers (4/10, 40%) reported information on missing data or steps taken to account for missing data and the potential impact on the conclusions.

There is a challenge in searching for information on technology. A 2022 review evaluated the type and quality of information available on the web for aquatic physiotherapy targeted to people with Parkinson diseases [59]. The authors used a commercial social listening service Awario that searches social media platforms (eg, Twitter and Instagram) and the web for investigator-selected keywords [59]. The strategy used here was a form of snowball sampling where systematic reviews served as the source, and the web was searched for any devices named in these reviews.

Many commercial technology companies reported ≥1 clinical trials that are underway; yet, there is a lack of trial-specific information. A potential consumer is unlikely to track these details. It is important to consider the transparency and accessibility of scientific evidence when making evidence-based recommendations to consumers. There is limited research in this area, specifically from a consumer’s perspective. Although there is a need to provide therapy outside clinical settings, it is critical that companies marketing technologies do not scam people into buying products that are possibly noneffective or even harmful and that clear reporting standards for consumers are made mandatory for these technologies, similar to those now standard for food.

The approach taken here may not have yielded complete results, and, because new technologies are continually developed and added to or removed from the market, the results can quickly become out of date. Many technologies are developed in research settings and are not given a proprietary name until there is evidence to support commercialization. Hence, searching for earlier information is impossible. In addition, the inventors, the authors of the papers, and the entrepreneurs commercializing the technology may not be the same people; hence, an author search will also be fruitless. CuPiD/Gait Tutor is an example of a name change [61,65]. Finally, there is no gold standard for rigorous, systematic gray literature search methods, and there are few resources on how to conduct this type of search; for example, the Cochrane Handbook, often cited as the gold standard for conducting systematic reviews, provides limited guidance and specificity for gray literature search methods [62]. In addition, the reporting of gray literature search methods in systematic reviews is often not held to the same high standards in transparency and reproducibility as the academic database search methods.

Therefore, the findings of this review are only valid based on the search conducted at the time. Given the difficulty in searching for gait training technologies, the search method reported here may be difficult to reproduce. Nevertheless, the information presented on the technologies discovered in this search uncovers existing gaps in the evidence and the reporting.

As newer technologies for gait training are continually developed, it is important that the evidence supporting their efficacy and effectiveness is quickly made available to people to make an informed choice. Often, the published literature is unavailable to the general population because of journal paywalls. There was also a lack of consistency in reporting information related to usability, safety, or user feedback. Standards for reporting on research involving technological devices, in the form of reporting guidelines, seem to be a critical need to ensure that the data needed by the potential consumer are communicated.

There are several reviews on the efficacy and effectiveness of gait training technologies. One objective of the research is to build capacity and empower patients who wish to take charge of their health. By equipping people living with gait impairments with the opportunity to improve walking outside clinical settings through biofeedback is a step in the right direction, given the limited access to rehabilitation services. A few technological innovations were initiated along the commercialization path but were abandoned at different stages. Despite the many benefits of at-home therapy, some challenges exist, including device maintenance, battery life, and technological literacy. One study suggests that the most effective devices are those that have a “user-centered design,” meaning patients or practitioners are involved in the design process [30]. Many of the devices included in this review use this approach by consulting patients for feedback on comfort, ease of use, and preferred feedback modes, when applicable, during pilot studies.

Although there is strong interest from academic institutions and government agencies to transfer technologies from laboratories to clients, there is a need for due diligence on the part of both the institutions and industry to accurately report all the findings that not only support the science but will also influence a client’s or an organization’s decision to purchase the technology.

Research in the field of technology development seems to lack the rigorous research method standards required for drug testing, allowing some devices to enter the market based mainly on safety rather than efficacy. Almost all commercial devices overclaim the efficacy of the technology to other populations not supported by their research; for instance, a website will claim effectiveness for people with gait impairments when the device was tested only in people with Parkinson disease. Finally, it would truly benefit the general public to have a summary of the research in layperson’s terms similar to food standards.

The field of technology evaluation would benefit from reporting guidelines to extend the guidelines for reporting on randomized clinical trials (eg, Consolidated Standards of Reporting Trials [CONSORT]), such as are available for many different types of experimental studies, including pilot and feasibility studies and crossover designs, all of which can be found on the Enhancing the Quality and Transparency of Health Research (EQUATOR) website [54]. There are also guidelines for reporting on the features of the intervention (Template for Intervention Description and Replication [TIDieR]), which would be helpful to fully understand the intervention protocol and encourage replication [56]. For technology, it would be useful to provide additional information on user experience using both closed- and open-ended formats to identify challenges that users encounter with the technology.

This review is the first of its kind from a consumer’s perspective that critically appraises wearable biofeedback gait devices found on the internet, the literature available on the respective websites, and the strength of evidence supporting the claims. The review highlights the need for providing standardized reporting of device capabilities as information accessible to the public when marketing commercialized devices. This review provides the public and health care practitioners with a summary of information that can be used to choose wearable biofeedback gait technologies or decide not to adopt them. The review covers 17 wearable devices that provide 1 form of feedback to improve gait and outlines the mechanisms claimed to underlie gait improvement. There was no predominance for biofeedback type (positive, negative, or continuous). A variety of biofeedback modes have been used (auditory, visual, or haptic), with auditory and vibratory haptic being the most common. The strength of the evidence supporting these devices from scientific sources was at 2a (lower randomized controlled trial) or 2b (prospective controlled trial—not randomized) level. Gaps in reporting all needed information for the consumer were uncovered. The propensity of small trials and heterogeneity of studies and conditions highlight the requirement for standardizing reporting of feedback intervention measures and doses to enable meta-analyses to move gait technological rehabilitation forward. Of note, there is a lack of evidence for motor learning interventions even in the field of sport, with a need for current evidence to be extended by theory-driven, high-quality studies to allow for more consolidated and evidence-based recommendations. Technology has the potential to advance the rehabilitation space and enable a better understanding of optimal interventions for learning and maintaining skills. Taken together, our findings target the need for clear reporting standards for gait interventions.