As of 2016, a total of 36 million Americans had engaged in some form of yoga practice [1], with other countries demonstrating similar patterns of yoga practice [2-4]. Although healthy individuals incorporate yoga into their fitness routines to improve physical and mental health [1], yoga has also been used to manage symptoms of disease [2-5]. Even before the COVID-19 pandemic in 2020, individuals sought access to yoga in their home environments. In fact, results from a 2016 survey revealed that yoga was most commonly practiced at home [1]. During the pandemic, public health policies and individuals’ personal preference for social distancing and staying at home stopped or limited in-person practice. Under these circumstances, the availability of resources to practice yoga remotely increased, which could be beneficial for individuals with functional limitations.

However, when examining the available scientific literature, yoga interventions are most often delivered in person [6]. Although this may be feasible for constrained research projects, it greatly limits real-world applicability. First, access to qualified yoga instructors inevitably varies based on geographical location [7]. Second, individuals’ socioeconomic status and access to transportation limit their access [8]. Third, inherent physical limitations may hinder the patients’ ability to attend in-person yoga classes. Therefore, in-person yoga instruction presents several barriers regardless of the public health climate, such that alternative delivery methods (eg, videos, videoconferencing, and mobile apps) have been developed and are being used extensively. However, there seems to be a gap in the literature regarding interventions conducted in research studies and real-world practices.

Little is known about the evidence regarding the practice and outcomes of yoga using remote delivery methods. A previous scoping review of the yoga literature [6] included studies in which yoga was conducted at sites other than yoga studios, such as at home, where yoga instructors may not be present. However, details regarding intervention delivery were not provided, although there are multiple ways of delivering yoga remotely [6]. For example, information about yoga can be provided through hard copy resources, such as pamphlets and card decks, or through more engaging technological means. Specifically, technologies such as videoconferencing, DVDs, websites, and mobile apps can be used to provide yoga interventions. Furthermore, content can be delivered either (1) synchronously (an instructor interacts with participants in real time, eg, through videoconference) or (2) asynchronously (material is prerecorded for participants to use without real-time interaction). Despite these options and their potential impact, few studies have explored their use. Thus, little is known about how yoga is delivered remotely, whether remote delivery is used for some specific populations more than others, and what types of yoga are being delivered remotely.

With this, it is important to know that yoga is a multilayered ancient philosophy and practice intended to facilitate well-being through the cultivation of awareness by integrating mind and body, with an emphasis on self-realization [9]. There are various forms, often referred to as branches of yoga, which facilitate one’s ability to reach a greater state of being. For instance, Karma yoga is the branch that focuses on devotion to service, Jnana yoga focuses on the development of knowledge, and Hatha yoga involves the practice of physical postures. Hatha yoga is typically what comes to mind when one thinks of yoga in the Western world. The term Hatha yoga refers to the branch of the physical practice of yoga and is also usually used to refer to a broad style of yoga that incorporates postures and breathing. There are other styles of yoga that fit under the umbrella of the physical practice of yoga—Hatha yoga—but are often specified further. Examples of these styles include Iyengar yoga and Vinyasa yoga. For example, Iyengar yoga focuses on body alignment, sequencing, and timing. By contrast, Vinyasa yoga is generally energetic and involves flowing through sequences of postures with breath integration. In addition to the branches of yoga and specific styles of the physical practice of yoga, there are 8 limbs of yoga. These 8 limbs are described in the Yoga Sutras of Patanjali, a foundational text, and include (1) yama (abstinences), (2) niyama (observances), (3) asanas (postures), (4) pranayama (breath control), (5) pratyahara (withdrawal of the senses), (6) dharana (concentration), (7) dhyana (meditation), and (8) samadhi (bliss) [9]. These limbs help to inform the practice of yoga in general, and some of them are incorporated in Hatha yoga. Most commonly, asana, pranayama, and dhyana are incorporated into Hatha yoga; however, in some cases, different combinations of the limbs of yoga are used. A previous scoping review of the yoga literature showed that most yoga intervention studies did not specify the style of yoga used however, of those that did, the most commonly reported styles were Hatha yoga (129/456, 28.3% of studies) and Iyengar yoga (41/456, 8.9% of studies) [6]. It is unknown whether similar patterns occur when yoga is delivered remotely.

Yoga can be delivered both in person and remotely, and there has been an increase in the availability and use of remotely delivered interventions in real-world applications. However, it is not known whether the scientific literature reflects this or reflects what types of populations have been enrolled in studies that have investigated the remote delivery of yoga, what the characteristics of these interventions are (including how remote delivery occurs), and what components of yoga are incorporated. Furthermore, the general feasibility and safety of these interventions or their preliminary efficacy are yet to be clearly defined. Therefore, the purpose of this scoping review was to examine the existing literature regarding the practice of yoga through remote delivery methods and identify current gaps related to (1) the populations studied, (2) the intervention characteristics (delivery methods and intervention components implemented), (3) the safety and feasibility of the interventions, and (4) the preliminary efficacy of the interventions.

Scientific databases, including PubMed, Scopus, Web of Science, PEDro, CINAHL, PsycINFO, IEEE Xplore Digital Library, and Cochrane Library, were searched during April 2021. The search strategies were modified for each database, with the Medical Subject Headings terms used as applicable. The following terms were used in various combinations: yoga, Iyengar, ashtanga, hatha, asana, telerehabilitation, tele-rehabilitation, telemedicine, videoconferencing, video, telenursing, DVD, remote delivery, eHealth, video games, online, and virtual reality (see Multimedia Appendix 1 for the specific search terms and search methods used for each database).

Articles were included if they investigated (1) adults aged ≥18 years; (2) an experimental intervention study with pre- and posttesting for at least one group; (3) a yoga intervention using physical yoga postures (asanas) and at least one other of the [9] 8 limbs of yoga, as described in the Yoga Sutras of Patanjali, including yama (abstinences), niyama (observances), pranayama (breath control), pratyahara (withdrawal of the senses), dharana (concentration), dhyana (meditation), and samadhi (bliss); and (4) a method of remote yoga delivery through technology for at least one group (eg, video, mobile app, and videoconferencing platform) as the primary intervention.

Yoga interventions can be heterogeneous and are often poorly described in the literature [6]. In some cases, they do not reflect the integration of the mind and body as intended in traditional practice [9]. To address this, we stipulated that selected articles should explicitly include physical postures (body) and another of the [9] 8 limbs of yoga (mind) using either English or Sanskrit words. This was intended to limit poorly described interventions or interventions in which yoga was not the primary intervention. For example, although mindfulness-based stress reduction includes components of yoga, there are other significant intervention components that make these interventions different from typical yoga interventions, which, if included, introduce even more heterogeneity across studies. Furthermore, it also allowed us to exclude studies that used only physical postures that could be more akin to exercise interventions.

All the data management processes discussed in this section were completed by the 3 authors (AJP, EZA, and JFD) in the following manner. The articles were divided into thirds. Then, 2 authors, screened, reviewed, and appraised, (two-thirds or 66.7% of the articles) with the third author available to resolve conflicts, such that each author (AJP, EZA, and JFD) was able to perform each task. The abovementioned criteria guided eligibility screening (completed via Rayyan QCRI [11]) and full-text article assessment. For articles that were included based on the full-text assessment, data were extracted simultaneously with the full-article review, and then a quality assessment was completed. The Quality Assessment Tool for Quantitative Studies Effective Public Health Practice Project was used to assess the level of bias and methodological quality of each study [12]. Data were extracted into a Microsoft Excel sheet designed by the scoping review authors (initial design by AJP, edited and approved by EZA and JFD) before the article review. The data extraction categories were determined by the authors (initial selection by AJP, edited and approved by EZA and JFD) based on the objective of the review (see Table 1 for descriptions and examples of data extracted from each study).

The search resulted in 1978 articles, with 1728 (87.36%) remaining after duplicates were removed. The title and abstract screening resulted in the inclusion of 2.03% (35/1728) of articles. Following a full-article review, of the 35 articles, 12 (34%) articles [13-24] published between 2003 and 2021 were included in the final qualitative analysis (see Figure 1 for the PRISMA [Preferred Reporting Items for Systematic Reviews and Meta-Analyses] flowchart).

Of the 12 studies, 8 (67%) were randomized controlled trials (RCTs) [13,14,18-23], 2 (17%) were quasi-experimental nonrandomized studies [16,17], 1 (8%) was an open-label single-group study [24], and 1 (8%) was a case series [15]. The comparison groups in the RCTs included regular activities [14,21], wait-list control [18], DVD program for strengthening [13], DVD program for walking [22], web-based stretching or toning program [23], and informational handouts about yoga [20]. One of the RCTs compared an in-person yoga intervention with a yoga DVD [19]. All active comparison groups were comparable in dose with the yoga interventions [13,19,22,23]. One of the quasi-experimental studies used health education [16], whereas the other [17] did not describe the comparison group. The methodological quality of the studies ranged from weak (7/12, 58% of studies [14-17,20,21,24]) to moderate (5/12, 42% of studies [13,18,19,22,23]; Table 2).

The studies included individuals with physical and mental health conditions, as well as generally healthy adults. For example, 17% (2/12) of studies involved individuals with cancer, such as early-stage breast cancer [13] and myeloproliferative neoplasms (blood cancers) [18]. Approximately 25% (3/12) of studies involved individuals with cardiopulmonary conditions, including pulmonary hypertension [15], neurocardiogenic syncope [17], and a combination of chronic obstructive pulmonary disease (COPD) and heart failure (HF) [16]. Other studies involved sedentary adults who were overweight [19], veterans with diabetes [20], women who had experienced stillbirths [23], and women with depression [22]. Approximately 25% (3/12) of studies involved adults without any specified health conditions, including adult women [14], female childcare workers [21], and the general public in India during the COVID-19 pandemic [24]. The mean age of the participants ranged from 21 to 71 years. Across all studies, most participants were female, and 50% (6/12) of studies included only female participants [13-15,21-23]. For more details related to the study populations, refer to Table 3.

In addition to extracting information about the enrolled populations, justification of the intervention delivery method based on the enrolled population was extracted when available. This was done to identify whether the study authors chose to implement remote delivery to address population-specific needs. Approximately 50% (6/12) of the studies reported some type of justification for the remote delivery method and related it to population-specific needs. Interestingly, one of these studies was designed during the COVID-19 pandemic and was conducted to provide easily accessible stress management strategies to the general public during social isolation [24]. Approximately 17% (2/12) of the studies [13,16], one for individuals with COPD and HF [16] and the other for women with breast cancer [13], justified their choice of implementing a remote intervention by noting that transportation-related barriers impeded individuals’ abilities to attend in-person interventions. Another study involving veterans with diabetes supported the choice of implementing a remote intervention by noting that physical impairments make it challenging for individuals to access and engage in physical activity [20]. Another study with older adults [14] reported low adherence to in-person exercise interventions in this population. Finally, a study with childcare workers [21] justified remote delivery using a yoga DVD as it was a convenient low-cost option. In contrast, 50% (6/12) of studies [15,17-19,22,23] did not provide any justification for the implementation of remote delivery.

The purpose of this scoping review was to examine the existing literature regarding the practice of yoga through remote delivery methods and identify current gaps related to (1) the populations studied, (2) the intervention characteristics (delivery methods and intervention components implemented), (3) the safety and feasibility of the interventions, and (4) the preliminary efficacy of the interventions. In summary, the studied populations included adults across their life spans, including individuals with physical and mental health conditions and some generally healthy adults. The review showed that, to date, most studies implementing remotely delivered yoga have implemented asynchronous delivery. In addition, the delivered interventions were primarily Hatha yoga interventions, including postures, breathing exercises, and meditation and relaxation exercises. The interventions were shown to be generally safe and feasible, with some feasibility challenges present in the study that implemented synchronous delivery. The heterogeneity of the included studies did not allow for an adequate evaluation of the preliminary efficacy of remotely delivered yoga interventions.

This scoping review showed that remotely delivered yoga has been successfully implemented in a heterogeneous sample of populations with and without chronic conditions. This is similar to the body of literature investigating the impact of in-person yoga [6]. In this review, studies included older adults; childcare workers; and individuals with chronic conditions such as COPD, HF, cancer, depression, diabetes, pulmonary hypertension, and cardiogenic syncope. Notably, most of these populations have also been implicated in in-person yoga studies. For instance, individuals with COPD [25] and HF [26], individuals with cancer [27,28], women with depression [29], older adults [30], adults with diabetes [31], and adults with cardiovascular conditions [32] have all been included in previous in-person yoga studies. Interestingly, no two studies reviewed here were conducted on the same population. Replication of studies for similar populations is seen in the in-person yoga literature [6] but is missing from the body of literature reviewed in this study. Therefore, replication using larger samples is needed. In addition, despite the variety of populations enrolled in the included studies, some populations that were commonly enrolled in in-person yoga studies were not included in the reviewed studies. For example, none of the included studies involved populations of individuals with neurological conditions or balance impairments. This differs from the body of literature investigating in-person yoga, which shows a large body of studies investigating yoga in these individuals [33,34].

It is unclear why some populations were enrolled in studies investigating remotely delivered yoga, whereas others were not. In an attempt to gain a better understanding of this, information related to the authors’ motivation to implement remote delivery was extracted. However, as reported in the Results section, half of the studies (6/12, 50%) [15,17-19,22,23] did not provide any reason for choosing the remote delivery method. Previous guidelines for the development of yoga interventions have highlighted that intervention delivery must be considered in intervention design and are discussed in the second domain of dose and delivery of yoga [35]. There are multiple reasons why remote delivery may be beneficial, some of which have been mentioned in the reviewed studies. For example, previous studies have shown that multiple factors can impede engagement in exercise and physical activity, including transportation barriers, lack of time, decreased motivation, feelings of intimidation because of physical or environmental barriers, costliness, stigma, and a general lack of resources [36-38]. These barriers may also impede access to in-person yoga, and as such, a better understanding of facilitators and barriers associated with in-person and remote delivery methods is needed. In fact, previous studies support the notion that interventions delivered through technology show higher adherence in older adults [39].

When examining delivery method characteristics, most studies (10/12, 83%) implemented asynchronous delivery. This is similar to the results of another scoping review that investigated web-based mindfulness interventions for people with physical health conditions, which showed that 69% (11/16 studies) implemented asynchronous delivery [40]. Furthermore, another systematic review investigating remotely delivered therapy for mental health showed that 73% (8/11 studies) implemented asynchronous delivery [41]. This may be because asynchronous delivery requires fewer resources; however, this has not been formally established. Specifically, for the purpose of this review, no studies to date have compared synchronous and asynchronous remote delivery methods in yoga interventions. A total of 2 previous reviews also showed that there were no studies comparing synchronous and asynchronous remote delivery methods for their respective interventions: web-based mindfulness [40] and remotely delivered therapy [41]. One of the studies comparing synchronous and asynchronous delivery of tele-exercise for individuals with spinal cord injury showed significantly higher adherence and average weekly training load for the synchronous training group compared with the asynchronous group, suggesting that synchronous training may offer added benefits [42]. Thus, a current gap in the literature relates to the investigation of synchronous remote delivery of yoga, despite promising outcomes for other interventions [40-42].

Although more studies included in our review implemented asynchronous delivery than synchronous delivery, the one study investigating synchronous delivery reported high adherence and enjoyment [16]. However, it also showed some feasibility challenges. It should be acknowledged that we cannot arrive at conclusions about the feasibility of all synchronously delivered, remotely delivered yoga interventions as there was only one study included in this review that implemented synchronous delivery. In addition, the other scoping reviews that investigated remotely delivered interventions did not speak to the feasibility of synchronously delivered interventions [40,41]. A previously published study indicated that certain videoconferencing platforms and practices may facilitate web-based engagement in remote research [43]. Thus, synchronous yoga interventions using widely available and commonly used videoconferencing platforms should be implemented and investigated to determine the approaches that limit technological challenges.

When examining the other intervention characteristics, such as the implemented style and limbs of yoga, our findings are similar to those found in the in-person yoga literature [44]. Although some of the studies reviewed here did not report the yoga style or limbs used, others reported a style of yoga that fell under the umbrella of Hatha yoga (the physical practice of yoga) without further specification. Limited available guidance on the optimal style may be is the reason behind a lack of specific reporting surrounding yoga style in general. In fact, a previously published systematic review [45] indicated that RCTs using different yoga styles did not differ significantly in their odds of achieving their desired outcomes. This demonstrated that there may not be an optimal style of yoga, and other factors such as preference and availability can help determine style [45]. However, future work should strive to report the style and limbs of yoga being implemented in interventions to help identify whether there are population-specific intervention characteristics that can optimize outcomes. For instance, a specific yoga style may be more beneficial for a given population. For example, a study comparing meditative yoga to power yoga for stress reduction in physically active, yoga-naive women showed significant improvements in salivary cortisol and state anxiety following a single meditative yoga session compared with power yoga [46], demonstrating that certain types of yoga may be more beneficial in specific situations. In addition, one yoga style may be delivered remotely more easily than another. For example, it may be more challenging for individuals with chronic health conditions and limited yoga experience to engage in a Vinyasa flow intervention involving rapid transitions from posture to posture while trying to follow a prerecorded video. However, this has yet to be investigated and could not be examined in this review or previously published reviews [6,45]. Thus, future studies should investigate the interactions among yoga style, study population, and other intervention characteristics, especially delivery method.

This scoping review had some limitations. First, only studies with pre- and posttesting were included. This was stipulated in the eligibility criteria and was intended to facilitate the exploration of preliminary efficacy. However, this may have resulted in the exclusion of studies, such as feasibility studies or qualitative analyses, which could have provided additional insight into the current state of the literature. Specifically, we are aware of 4 studies [47-50] that were excluded because of a lack of explicit pre- and posttesting. Second, the heterogeneity of the included studies and assessed outcomes prevented the identification of conclusive preliminary efficacy findings. Third, the searches were completed in April 2021. With the continuation of the COVID-19 pandemic, publications involving telerehabilitation in general [51-53] and remotely delivered yoga interventions, in particular, may increase over the next few years. This could possibly require an update to this review. Regardless of these limitations, this scoping review identifies key gaps in the related literature and provides a strong foundation to optimize future research.

This scoping review had several strengths. A broad range of databases was searched, which allowed a comprehensive search. This review provides a robust quality assessment of the included studies, provides a realistic picture of the literature and facilitates the interpretation of the findings. In addition, this review covers a broad range of content areas, including (1) the populations studied, (2) the intervention characteristics (delivery methods and intervention components implemented), (3) the safety and feasibility of the interventions, and (4) the preliminary efficacy of the interventions. This broad range of content allows readers to obtain a full picture of the state of the related literature and understand the current gaps. This is intended to help provide a path forward to optimize future research.

Multiple steps can be taken to address the gaps identified in this review and optimize future research. Future studies involving larger sample sizes should assess populations similar to those enrolled in the reviewed studies to determine whether the results can be replicated. In addition, populations not examined in the included studies, such as those with neurological conditions or other populations that have been shown to benefit from in-person yoga, should be enrolled in future studies that implement remotely delivered yoga. Future studies should justify the choice of delivery methods and relate this justification to population-specific needs. Moreover, future studies should consider and investigate the interactions among delivery methods, yoga intervention components, and other study characteristics. They should explore the implementation of synchronous delivery and compare different delivery methods. Specifically, synchronous yoga interventions using widely available and commonly used videoconferencing platforms should be investigated to determine whether this approach limits technological challenges and facilitates feasibility. Finally, future studies should report information regarding adverse events, adherence, and other safety and feasibility measures to provide robust information regarding the implementation of these interventions.

This review synthesized the literature regarding the remote delivery of yoga and provided information about gaps in the literature related to study populations, intervention characteristics, intervention safety and feasibility, and intervention efficacy. Overall, this review revealed a broad gap in the literature, showing that little attention has been paid to yoga intervention delivery methods. Future studies and yoga intervention development guidelines should further consider the delivery methods when developing interventions. For instance, population-specific needs and barriers should be accounted for when determining delivery methods. In addition, more studies implementing synchronous delivery methods and studies comparing delivery methods should be conducted, and robust reporting of intervention characteristics is required.