Approximately 280 million people worldwide are experiencing depression [1], and nearly half of those people also experience comorbid anxiety [2]. Symptoms of either disorder can cause substantial distress and impair functioning across all domains of life. Worse yet, the prevalence and severity of these disorders are growing [3]. Despite the increasing evidence base of effective interventions, including cognitive behavioral therapy (CBT), the prevalence of depression and other mental health disorders continues to rise [4]. Accordingly, there is a demand for effective, accessible, and scalable interventions to manage these mental health conditions.

CBT is one of the most effective interventions for depression and anxiety [5]. The model underlying this intervention describes how one’s cognitions influence and are influenced by their emotional, behavioral, and even physiological consequences [6]. Based on this model, CBT teaches clients a variety of cognitive behavioral skills to identify, evaluate, and change maladaptive thoughts and behavioral patterns to indirectly change emotions. CBT has been deemed an empirically supported treatment through direct comparisons to waitlist controls and other bona fide interventions such as antidepressant medications [7]. Not only is CBT generally effective during its acute phase but it also has an enduring effect on symptoms that lasts beyond the end of treatment [8]. Despite its many strengths, CBT-oriented care is often difficult to access and scale due to the lack of trained providers, high costs, and stigma associated with seeking mental health care [9].

Two major innovations in the mental health care field could help to address the aforementioned problems of accessibility—the deployment of lay therapists and the use of technology. First, nonprofessional lay therapists (hereafter referred to as “coaches”) can be trained to provide peer support and teach cognitive behavioral skills. A primary driver of the effects seen in CBT is clients’ use of the specific cognitive behavioral skills taught [10,11], suggesting that cognitive behavioral skills may be a worthwhile training target for mental health support. Prior research has shown that individuals without any prior formal professional training can be taught to deliver effective cognitive behavioral interventions [12], therefore, reducing the burden on and expense of the limited workforce of highly trained, specialized therapists.

Second, digital mental health interventions can deliver cognitive behavioral skills training nearly anywhere at any time and have outcomes comparable to face-to-face CBT [13-18]. Although technology can remove the need for therapists (eg, through completely self-guided interventions or provided through the use of artificial intelligence–based computerized cognitive behavioral programs), digital CBT programs have experienced high dropout rates, with reviews reporting average dropout rates of 32%-74% depending on dropout definition and level of support [19-21]. Fortunately, human support provided via peers or coaches can help improve accountability, engagement, and efficacy [22]. Furthermore, human-mediated digital mental health interventions can help promote nonspecific effects that may contribute to outcomes [23]. Social connection has been posited as a key contributor to engagement in digital mental health interventions [24], indicating that digital interventions must strike a balance between accessibility and human interaction in order to be scalable while remaining effective. While existing CBT programs have aimed to leverage lay therapists and digital modalities to deliver interventions remotely, at scale, and at a lower cost, their uptake and effectiveness remain inconsistent. This highlights the clear need for innovative approaches that can enhance both engagement and efficacy.

Recently, the technology of the metaverse, defined as internet-connected social digital environments that allow end users to interact as avatars, has been adopted to facilitate digital mental health interventions delivered by humans [25]. The metaverse can transport an individual into a personalized 3D digital world that promotes immersion, the extent to which an individual perceives, experiences, and interacts with a digital environment as if it were a real environment [26]. Individuals can create an anonymous avatar and become immersed in a world where they communicate by speaking to other people, each of whom is also represented by an avatar. This technology affords a unique opportunity to deliver an intervention that is simultaneously immersive and anonymous, potentially allowing individuals to feel more comfortable being vulnerable [25]. The level of anonymity that accompanies the use of avatars may also attract end users who typically do not seek out mental health care, whether due to shyness or stigma [25,27]. Finally, the digital nature of the metaverse can reduce barriers to access related to cost, lack of trained providers, or other physical hurdles. Metaverse environments can be accessed through commonly owned flat-screen devices (eg, computer or phone) or through virtual reality technology (ie, head-mounted displays that generate graphics in real time that simulate a 360° view of an environment that one can interact with using hand-held controllers). Virtual reality enables end users to feel strongly immersed in a 3D space, which can increase the sense of presence in a particular environment despite being physically situated in a different place [28]. The metaverse is a suitable setting for coach-based mental health interventions designed to facilitate cognitive behavioral skills training, integrating digital facilitation of effective mental health interventions with realistic anonymous interpersonal interactions. Despite growing interest in virtual reality and metaverse apps for mental health, research on integrating virtual reality and the metaverse with coach-supported CBT skills remains limited.

Cognitive behavioral immersion (CBI), a synchronous cognitive behavioral skills group training program delivered by coaches in the metaverse developed by Innerworld, Inc, is the first app to our knowledge that bridges virtual reality, the metaverse, and CBT skills in the pursuit of delivering efficacious, accessible, and scalable mental health care [25]. A pilot study provided support for its feasibility among a group of individuals in recovery from a substance use disorder [25]. A second study on this intervention demonstrated improvement in depression and anxiety symptoms associated with CBI engagement [29]. That study found that internet-based social support was associated with depressive symptom improvement, while the use of virtual reality (vs a flat-screen device) with the intervention was associated with greater anxiety symptom improvement. While this preliminary support for the use of CBI is promising, a direct comparison of this modality against a control has not yet been conducted, and its absolute efficacy is unknown. In addition, it is unknown whether accessing CBI via immersive virtual reality confers any advantage over more readily accessible, less immersive flat-screen devices.

This study is designed to address the following research questions: (1) is CBI efficacious (better than its absence)? and (2) does immersive virtual reality confer any advantage over and above accessing CBI via less immersive flat-screen devices? In line with these questions, this study aims to compare immersive CBI accessed via virtual reality (CBI-VR) to a delayed access control (DAC) condition and less immersive CBI accessed via flat-screen devices (CBI-FS) in a randomized controlled trial. We predict that participants randomized to CBI-VR will experience greater clinical and functional improvement than participants randomized to the other conditions. In addition to this primary aim, we plan to explore predictors of response regardless of condition (prognostic), prescriptive indices that predict differential program response (moderators), and causal mechanisms of clinical and functional change across conditions (mediators).

This study is a three-arm unblinded randomized controlled trial (NCT06418997). Participants are randomly assigned to one of three groups in a 1:1:1 ratio: CBI-VR, CBI-FS, or DAC. In the CBI-VR and CBI-FS conditions, participants participate in up to eight weekly intervention sessions. Those in the CBI-VR condition will be mailed Meta Quest 3 headsets to participate in the intervention, while those in the CBI-FS condition will use their own existing flat-screen device (such as a personal computer, tablet, or smartphone) to participate in the intervention. Participants in the DAC condition will be asked not to access the intervention for 8 weeks but will then be provided with information on and encouraged to access the same intervention. Outcomes will be monitored weekly throughout the 8-week acute phase and monthly for 6 months during follow-ups. All procedures will be conducted on web-based platforms. See Figure 1 for a flowchart of study procedures. Reporting in the protocol was based on the CONSORT-EHEALTH (Consolidated Standards of Reporting Trials of Electronic and Mobile Health Applications and Online Telehealth) and SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) checklists (Multimedia Appendices 1 and 2).

All study procedures received human participant ethics review approval from the University of Southern California’s (USC) Institutional Review Board (UP-23-00491). As part of the approved procedures, all participants will provide informed consent prior to beginning the study and be made aware that they can stop their participation in the study at any time. All data will be collected securely through REDCap (Research Electronic Data Capture; Vanderbilt University) [30,31]. After the study is completed, all data will be deidentified. Participants will be compensated with electronic gift cards sent to their email for completion of study surveys. Participants across all conditions will also receive a new Meta Quest 3 virtual reality headset to keep as part of study participation. Distribution of the headset is dependent on study randomization; those randomized to the CBI-VR condition receive a headset prior to starting their 8-week condition, whereas participants in the other two conditions receive the headset after completion of the 8-week acute condition.

We are recruiting a sample of 306 individuals from across the United States who endorse clinical levels of depression symptoms. Participants are recruited through paid advertisements on popular social media apps (eg, Facebook and Instagram), posts on clinical trial or research study registries (eg, ClinicalTrials.gov), through our research lab’s website, and by word of mouth.

The inclusion criteria for this study are (1) clinical level of depressive symptoms (indicated by a score of ≥10 on the 9-item Patient Health Questionnaire [PHQ-9] [32]), (2) 18 years or older, (3) able and willing to give informed consent, (4) personal computer with a stable internet connection that can run the CBI application, and (5) able to speak and read English. The exclusion criterion for this study is elevated suicide risk at the baseline assessment that precludes inclusion in the study (see Safety Monitoring section for more information).

All prescreening and intake procedures will be managed by a research team at USC. Interested individuals will be directed to a link where they will complete an initial prescreening questionnaire hosted on REDCap containing a brief description of the study, eligibility questions, and the PHQ-9. Those who meet initial inclusion criteria will be emailed a link to schedule a 20-minute intake session conducted via Zoom (Zoom Video Communications).

Prior to attending their intake session, each individual will be asked to sign the study consent form via the REDCap eConsent module and to complete a set of intake measures via a secure REDCap link. During the intake session, study personnel will explain the study; answer any questions; confirm participant eligibility; record the participant’s name, address, and mobile phone number for verifying identity and for mailing the virtual reality headset; collect the participant’s availability for attending a CBI intervention group, and secure informed consent. After completing the intake session, participants will be randomized to one of three conditions.

Participants randomized to the CBI-VR or CBI-FS conditions will be sent an email alerting them of their study conditions and providing directions on how to access the application hosting the CBI intervention. Prior to their first intervention session, participants will take part in a web-based tutorial about how to navigate the CBI intervention and may attend an optional onboarding session with study personnel if additional support is needed navigating the CBI intervention (and virtual reality headset if randomized to the CBI-VR condition). DAC participants will also be emailed to inform them of their study condition, when they can expect to begin receiving surveys, and when they can expect to receive compensation.

Participants are randomized to one of the three groups in a 1:1:1 ratio using a random number generation Python script. The seed Python uses to perform random generation is predetermined for each “wave” of participants to aid replicability. Once the number of recruited participants in a wave has reached the cap indicated by coach and researcher availability, recruitment will be paused, and that wave of participants will be randomly assigned to a group and commence their 8-week condition. This process will repeat every few weeks until the planned sample size of 306 is reached. Randomization will be done in blocks so that participant distribution across each group per wave is equated to ensure CBI classes during each wave have sufficient scheduled participants. The randomization procedures are conducted by the research team at USC.

The CBI intervention will be hosted on the Innerworld app. Innerworld is an English language app freely available for download across several platforms (eg, Windows, Meta Quest headsets, iOS, and Android). Innerworld trains adults who do not necessarily have prior professional training to serve as CBI coaches. Each coach undergoes an intensive 20-hour training program spread over 5 days led by Innerworld support staff to learn how to facilitate cognitive behavioral skills groups. Coaches will be trained to lead an 8-week CBI for depression course, which was developed by investigators who are experts in CBT. See Table 1 for a condensed outline of the coach training program. After completion of the fifth training session, each coach must cofacilitate at least one real meeting in the presence of a trainer (ie, a coach who has extensive experience leading Innerworld groups). The coach is evaluated by the trainer and provided with feedback. To maintain coach adherence to the intervention throughout the trial, coaches will be required to attend weekly supervision led by Innerworld support staff and team investigators. We will also ask participants to complete the Group Session Rating Scale to assess group quality.

The CBI for depression intervention involves participants meeting in groups of up to 10 in which they learn lessons on various strategies to cope with depression based on cognitive behavioral theory [6]. Each group meets for 1 hour once a week for 8 weeks. In addition, user feedback during pilot testing led to the implementation of an “a la carte” model that allows participants to attend a rotating schedule of makeup sessions. Peer coaches will follow a detailed manual describing the content to be followed for each session. See Table 2 for a condensed outline of the session content. Behavioral strategies include but are not limited to activity scheduling, problem-solving, and assertion training. Each session begins with setting the agenda and reviewing homework from the previous session. Cognitive strategies include but are not limited to identifying the connections between thoughts, feelings, physiology, and behaviors, and catching and examining negative automatic thoughts. Each session ends with a session summary, homework assignment, and feedback about the session. If participants cannot attend their assigned session, participants are encouraged to attend a makeup for that session via email prompts sent by the research team. If participants need support navigating the app or using their virtual reality headset after beginning the intervention, they will be directed to connect with the app support staff, who are available as avatars in the app 24/7.

We will take precautions to monitor the safety of the study participants. First, all coaches and study investigators at USC will undergo training on how to manage crisis situations. During each CBI session, coaches will also monitor any discussion participants share regarding their safety in relation to themselves and others. We will also establish automatic thresholds on REDCap for PHQ-9 and 7-item Generalized Anxiety Disorder (GAD-7) assessments completed before, during, and after baseline. If a participant meets the criteria for suicidality (a 2 or above on the PHQ-9 suicide item), crisis resources will be automatically displayed and participants will be prompted to complete the Columbia-Suicide Severity Rating Scale (C-SSRS) Self-Report [33]. High-risk participants (scoring 4 or above on the C-SSRS Self Report) will be contacted by the research team via email to schedule a Zoom call for further assessment using the C-SSRS interview. Participants scoring a 4 or above on the C-SSRS interview will be excluded from the study due to sufficiently elevated suicide risk and will be provided with crisis and mental health resources. If imminent risk is present (C-SSRS interview score of 5), a lead study investigator (IDE) will evaluate the situation and contact local emergency services if deemed necessary. Participants with significantly worsening symptoms (a 20% increase on PHQ-9 or GAD-7 and a score of 15 or above) will be contacted by USC study personnel via email for a brief check-in and provided with mental health referrals. A copy of these referrals will also be publicly posted in the Innerworld app at all times.

All adverse incidents will be documented and immediately reported to lead study investigators, who will then follow them to a satisfactory resolution. They will distinguish serious adverse events from adverse events not judged to be serious using the guidance and definitions provided by the National Institutes of Health notices. All serious adverse events will be reported to the USC Institutional Review Board and the study’s independent Data and Safety Monitoring Board.

All assessments will be managed by a research team at USC and securely collected using REDCap. All participants regardless of condition will receive the same set of web-based surveys at the same time points apart from the two session-related assessments (Group Session Rating Scale and Competencies of Cognitive Therapy Scale). These two session assessments will not be given to the DAC participants because they do not complete sessions during their 8-week condition. In the case that a participant does not complete a survey, study personnel will send up to three reminder emails over the course of 1 week post survey invitation to complete the uncompleted survey. See Table 3 for the assessment schedule.

This study will be the first randomized controlled trial to test the efficacy of CBI. While the cognitive behavioral skills being taught and the therapeutic framework of peer support are each supported empirically, this study investigates an intervention that integrates these components in a novel environment (the metaverse) and compares its efficacy to a minimal treatment control condition. This study will also examine the relative efficacy of accessing the coach-led cognitive behavioral skills training program via immersive virtual reality technology versus less immersive flat-screen devices. We anticipate that participants randomized to both CBI conditions will experience improvements in depression symptoms, anxiety symptoms, and quality of life. We also predict that participants who access the program through virtual reality will experience more positive outcomes than those who access the program through flat-screen devices. Furthermore, we will explore a variety of prognostic variables (ie, cognitive behavioral skills, group alliance, and program engagement), prescriptive variables (ie, demographics and immersion), and mechanisms (ie, cognitive change and social support) of program response.

The findings from this study will test the efficacy of this intervention relative to its absence and determine if and how the access platform affects efficacy. This study will also help to determine if there is a benefit to the increased cost of disseminating CBI through virtual reality headsets versus through devices that many people may already own. If deemed efficacious, this intervention can be promoted to scale and subsequently increase access to effective mental health interventions.

We anticipate some potential challenges. The primary challenge we foresee is a high dropout rate given an elevated dropout rate among digital interventions in prior research [43]. To help reduce the risk of drop out, we have designed a comprehensive onboarding process to get study participants comfortable with the study procedures and address any anticipated technical barriers to study completion. We did incorporate an expected dropout rate of 25% into our power analysis to avoid low statistical power in the event of sizeable attrition rates. We will monitor participant engagement with the sessions and assessments, as well as send electronic reminders of intervention sessions and study assessments. If a participant misses any session or assessment, we will send follow-up reminders and inquire about barriers to completing the study. If any barriers are noted, we will work with the participant to address them. Given that this intervention is delivered on a web-based platform by nonprofessional coaches, there exists the chance that technical or personal issues related to Innerworld, or its coaches, may interfere with the delivery of the intervention to participants. To mitigate this risk, Innerworld provides 24/7 access to live coaches (who are trained on how to detect if a person is in crisis and provide the proper resources) and technical support to address any unexpected obstacles. Finally, we were unable to implement blinding due to differences between interventions (such as the device used to access the intervention) that would be obvious to participants, as well as coaches. This could introduce bias in how participants or coaches behave or perceive their symptoms and the intervention.

If we find a significant difference in favor of CBI versus the DAC condition, this will provide evidence that CBI is efficacious for individuals with depression. If we find a significant difference in favor of CBI accessed via virtual reality headsets over the less immersive CBI accessed via flat-screen devices, this may suggest that the immersive quality of CBI strengthens the program. If we find no such difference, this will suggest that CBI may be effectively implemented with either virtual reality or flat-screen devices (ie, computer or handheld mobile devices), significantly reducing the cost of access since most people have a flat-screen device. This study will produce valuable information on whether CBI can effectively serve as a scalable and accessible mental health intervention among a growing, diverse population of help-seekers. In addition, knowledge of the differential benefit (or lack thereof) of immersive technology will help researchers develop other virtual reality and metaverse interventions, as well as funding agencies in determining whether virtual reality equipment is worthwhile or necessary in clinical studies.