For decades, group therapy has been an acknowledged treatment for a broad range of anxiety disorders, including generalized anxiety disorder, panic disorder, social anxiety disorder, and specific phobias [1-5]. Clinicians, public health administrators, and researchers have highlighted its scalability and cost-effectiveness advantages over individual therapy [4]. With the advent of telehealth services, online group therapy, such as online individual therapy, has also emerged as a treatment that can facilitate access across geographic barriers to underserved regions and populations.

Yet the reduced interactivity of online group therapy may be limiting its impact and broad deployment [6]. In particular, concerns have been raised about the effects of standard video-enabled platforms on group dynamics and group identity. With new digital advances such as extended reality (XR) devices, which include head-mounted virtual reality (VR) and augmented reality (AR) devices, there may be new ways to address these limitations, enhancing the online group therapy experience. This study aims to review the research in group therapy to better understand how the incorporation of XR technology may help online group therapy realize its full potential.

A sizeable body of research suggests that group therapy can be as effective as individual therapy for anxiety disorders [7,8], with the additional benefits of decreasing isolation, providing social support, containing costs, and increasing access [9]. In a meta-analysis of 67 studies involving 3656 adults with generalized anxiety disorder, social anxiety disorder, or panic disorder, group therapy, mostly in the cognitive behavioral therapy (CBT) mode, was more effective than no treatment and not significantly different from individual therapy or pharmacotherapy [7]. While there is some heterogeneity in results across individual studies, another meta-analysis with strict comparison protocols linked the divergence to researcher bias, with researcher allegiance to group or individual therapy moderating study findings [10].

Several factors have been posited as affecting responses to group therapy. Those include the type of therapy [11], individual characteristics such as comorbid conditions [12], and the quality of group dynamics. Although head-to-head comparisons are limited, it is generally accepted that group CBT is superior to other forms of group therapy for anxiety disorders, including group cognitive therapy sans a behavioral component, group interpersonal therapy, group psychodynamic therapy, and group social skills training [7]. Furthermore, the presence of comorbid conditions seems to decrease efficacy, although the number of group therapy sessions seems to moderate this effect [12].

In addition, group dynamics appear to have an impact on effectiveness [13]. A total of 2 studies suggest that perceived similarity with other group members has a positive effect on treatment response [14,15], and an analysis of 43 transdiagnostic CBT groups on anxiety disorders found that group cohesion was a crucial factor in group recovery outcomes and that the effectiveness of group therapy stems not only from the presence of a facilitator but also from the group itself [16]. This would seem consistent with data on groups improving well-being [17], offering solidarity [18], and contributing to a social identity or sense of self as a group member that is separate from one’s personal identity [19]. Shared social identification, specifically, can lead to improvements in well-being, including when group belonging fosters a sense of being “special” [20]. Research in anxiety disorders, specifically, suggests that identifying with a treatment group can dramatically increase the chances of improvement compared with other factors [20]. In addition, within therapeutic relationships, social identification with a therapist strongly predicts a therapeutic alliance, which in turn can predict positive outcomes [21]. Taken together, the data suggest that group social and identity dynamics play a crucial role in group effectiveness.

Most group therapy research, including the social identification–based research discussed earlier, has focused on in-person group therapy due to the relatively low prevalence of online group therapy in the past 20 years. Since the COVID-19 pandemic, however, there has been a dramatic increase in online group therapy offerings [22,23], with many mental health practitioners continuing to provide it after the pandemic [22,24]. The increase in online therapy availability has helped broaden access to care in underserved regions and among underserved populations, including racial and ethnic minority individuals [25], and parents with children [26]. Besides increasing access, the cost-effectiveness and scalability considerations suggest that online individual and group therapy will have staying power and will play a determining role in the future of mental health care [27]. Online group therapy (through group video calls) has also been shown to be highly effective in treating anxiety disorders for over a decade [28-30].

However, as the advantages become more evident, so do the limitations. Besides an up-to-date device, access to online tools still requires reliable, speedy internet coverage, an ongoing barrier worldwide [31,32]. In addition, individuals with lower digital literacy, including some older adults, may be unable to install apps or navigate complex digital menus to access online health tools [33]. Of particular relevance to group therapy is recent research that suggests that users’ ability to engage with others, including identification with a group, may be limited on online video platforms [34]. For example, the narrow focus on the screen in online video calls appears to negatively affect people’s ability to be creative in group settings [35]. Furthermore, a review by Aagaard [36] argued that online video group interactions can give rise to “Zoom fatigue” due to a unique combination of the awkward tendencies around taking turns to speak, heightened self-awareness, limited mobility, lack of true eye contact, and an inability to be spontaneous. Still, another study [37] argued that the intensity of gaze can be psychologically taxing—normally, people do not stare intensely at proximity in person for long periods. This gaze presents a further challenge for users: in video calls, most webcams are situated slightly above the screen, contributing to a feeling that the participants are being looked down on and intensely observed. Both Aagaard [36] and Bailenson [37] argue that online video interactions effectively require developing a new way of communicating, because standard social dynamics must be unlearned due to different social cues and communication patterns.

It is possible that video-based interpersonal interactions are more challenging for those with anxiety disorders [38]. Robust research already suggests that individuals with anxiety conditions are more attentive to social cues [39], and limited research suggests that this is especially true online. For example, 1 study involving 24 users found that patients with social anxiety disorder expressed higher physiological signs of state anxiety in response to an online, video-based social task than controls, with high levels of attention to body language in the anxiety cohort [38].

Perhaps more importantly, social identification with a therapeutic online group purely through video may be more difficult due to fewer nonverbal cues than in in-person interactions. Nonverbal cues, such as posture, lean, and arm or leg crossing, have been found to be important to positive group interactions [40], and the reduction in nonverbal cues online appears to reduce social identification with a group [40]. Studies also suggest that the lack of certain visual cues and the static, artificial environment may make online groups less appealing [41,42].

Emergent XR technologies might represent a solution to some of the social identification and interactivity limitations of group video therapy. XR technologies include VR, where the user experiences a virtual environment that is seen as real or near real, and that is largely due to computer-generated visual information, coupled with the blocking out of key elements of reality. VR technologies include head-mounted displays or a computer screen, where users are blocked from some real visual elements in favor of a virtual environment that they can navigate [43]. XR technologies also include AR, where users are presented with their real environment, plus virtual additions. AR applications include users looking through the camera viewfinder of a mobile phone and seeing a digital object in their real physical space [44] or using a head-mounted display with “passthrough” capabilities that allows them to see the real environment with added digital objects [45].

XR technologies have greatly evolved in the past 20 years. Older platforms were expensive, relied on wired connections, demanded substantial technological expertise [46], or were very limited, using VR through a computer screen or AR through a handheld mobile device. Current offerings are comparatively much more complex and immersive. Major companies, such as Meta and Apple, provide high-quality, fully wireless, and user-friendly XR headsets. These headsets are also reasonably affordable, with prices starting around US $150, and come with apps that make avatar-enabled group chat possible [47]. For example, Apple’s Vision Pro, which is more expensive at approximately US $3500, has persona-driven group chats that allow for realistic, 3D representations of people’s heads as well as dynamic and realistic eye tracking (Multimedia Appendix 1) [48]. More affordable options such as Meta’s Quest 2, Quest 3, and Quest Pro (formerly known as the Oculus Quest series and costing approximately US $200, US $500, and US $1000, respectively) all also have extensive group chat VR capabilities that can track upper body movement using customized avatars (Multimedia Appendix 2) [49]. Meta dominates the XR headset space, representing about 80% of XR headsets ever sold as of 2023 [50]. While it is acknowledged that other XR solutions do exist, like the ones offered by Microsoft, Valve, and ByteDance, this review will focus mostly on Meta due to its popularity and Apple due to its novelty.

Head-mounted XR devices can potentially mitigate some drawbacks associated with video social interactions [36]. Users can enter into a group chat; observe others move about; see head orientation; engage in overlapping conversations against a customizable background; move their eyes to make eye contact at will; have a customized self-image (if VR); and have the ability to better see other people’s body language, including hand gestures and head tilting or nodding [51]. These XR capabilities within social interactions might help address some of the aforementioned shortcomings inherent to online therapy. Furthermore, because practitioners and patients have the ability to customize the background for maximum use, distracting “background events,” which are common in online video calls, are much less likely [52].

XR can help in ways beyond enhancing online social identification and interactivity or mitigating against common issues in video-enabled care—it can deliver evidence-based treatment. A growing body of evidence supports the use of VR to deliver exposure and other interventions for anxiety disorders. A systematic review of reviews from 2020 analyzed data from 23 reviews on the use of individual VR for anxiety disorders, concluding that VR had strong potential as a supportive therapy in conjunction with conventional therapies [43]. Furthermore, a 2023 scoping review in anxiety disorders found that almost every study showed the VR intervention to be as effective as traditional therapies [53]. Similarly, a meta-analysis examining VR exposure therapy for anxiety disorders found large effect sizes in pre-post studies [54]. Overall, despite some methodological issues, all 3 reviews suggest that VR is a promising treatment for anxiety disorders.

In the context of online group therapy, it may be possible to supplement the VR-mediated improvement in group interactivity with VR-mediated exposure or other therapy interventions. Preliminary research suggests the potential success of such an approach. One study in social anxiety disorder, for example, examined the Reddit posts of users of group VR chats and found that aspects of VR (physicality, sense of presence, nonverbal activities, and shared activities) may reduce fear of social interactions in the offline world through a desensitization-like mechanism [51]. Although the data are limited, this suggests the feasibility of introducing VR into online group therapy.