Social anxiety, also known as social phobia, is characterized by a marked fear of social situations where individuals may be scrutinized by others, often leading to avoidance behaviors and distress [1,2]. It, often arising in adolescence, is among the most common anxiety disorders affecting youth and can substantially disrupt academic performance, peer relationships, and emotional well-being [3,4]. A central feature of social anxiety is extreme sensitivity to negative evaluation, perpetuating cycles of self-doubt and anticipatory anxiety [4]. Among young people, persistent concerns over potential judgment can lead to social withdrawal and diminished social support, thereby exacerbating loneliness and isolation [5]. Indeed, in an investigation involving adolescents and young adults, researchers found that social anxiety often manifests before major depression, underscoring the urgent need for early screening and targeted interventions in this population [6].

Virtual reality (VR)–based interventions provide immersive and controllable environments in which anxiety-provoking social or performance situations can be simulated [7,8]. For youth with social anxiety, VR may be particularly useful because feared situations can be presented gradually, repeated safely, and adjusted according to the participant’s level of difficulty; however, youth-specific evidence remains limited and requires further controlled evaluation [9]. Compared with in vivo exposure alone, VR may reduce some practical barriers by allowing exposure tasks to be delivered in a private and standardized setting, although transfer to real-world functioning should be examined empirically rather than assumed [7,8,10]. Therapist facilitation may further strengthen VR-based interventions. A trained therapist can help participants identify feared predictions and safety behaviors, prepare for exposure tasks, monitor distress, and consolidate learning through debriefing and homework planning [11,12]. However, evidence on therapist-facilitated VR interventions for youth social anxiety remains limited, and the extent to which symptom improvement transfers to daily functioning requires further empirical investigation [9,10].

Robust evidence on VR treatment for adolescent social anxiety is still sparse, particularly from randomized controlled trials (RCTs). This gap is noteworthy because adolescents, often described as “digital natives,” are generally receptive to technology-mediated care—a factor that can boost both engagement and adherence [9]. Although a small body of research indicates that self-guided VR programs can reduce anxiety by providing convenient, on-demand, and relatively low-cost exposure practice [7,8], the added value of therapist-facilitated VR remains largely untested. In a guided format, clinicians can give real-time feedback, fine-tune exposure tasks, and address comorbid issues such as cognitive distortions or emotional dysregulation as they emerge [13]. This active professional involvement is likely to improve treatment adherence, deepen therapeutic engagement, and optimize clinical outcomes by keeping the exposure simultaneously challenging and safely supported [13].

Additionally, the lack of economic evaluations for VR-based psychological interventions is a critical issue, particularly during times of global economic downturns, when funding for mental health support and research often declines [14]. Without rigorous cost-effectiveness analyses, policymakers may face challenges in allocating resources efficiently, potentially limiting access to innovative digital interventions that could provide scalable, engaging, and sustainable mental health support for young individuals. Addressing these gaps is important for understanding whether VR interventions can be both potentially useful and feasible for broader use in youth mental health care [15].

Our aim is to test the preliminary efficacy of a therapist-facilitated VR intervention for reducing social anxiety symptoms among young individuals, compared with a waitlist control condition. The primary research question is: “Does the therapist-facilitated VR intervention, when compared to waitlist group, result in a reduction of social anxiety symptoms among youths at 1-month (post-treatment) and 3-month follow-up?” Additionally, the study aims to investigate the cost-effectiveness of this intervention by determining “What is the cost per unit reduction in the SPIN score (if applicable)?” The study does not aim to establish implementation effectiveness in routine service settings or definitive generalization to real-world behavior. Our hypotheses are as follows:

The design is a 2-arm, statistician-blind, parallel-group RCT to test whether the VR intervention, compared with a waitlist control condition, leads to a reduction in social anxiety among youth. Assessments will be carried out at 0 (baseline, T0), 1 (posttreatment, T1), and 3 (follow-up, T2) months. A summary of the trial design can be seen in Figure 1.

The protocol and the informed consent form received approval from the Hong Kong Shue Yan University Human Research Ethics Committee (HREC 24-05(M3)). Eligible participants willing to join the study will sign an online consent form. For participants under 18 years of age, additional parental or guardian consent will be obtained. To reduce the likelihood or detection of harm, various safety and security procedures will be implemented, such as excluding potentially psychologically ill participants during the screening stage.

To ensure the confidentiality and security of participant data, all collected information is securely encrypted during transmission and storage, preventing unauthorized access. Access to participant data is strictly limited to authorized research personnel. Additionally, personally identifiable information is anonymized or deidentified wherever possible to safeguard participant privacy. Each eligible participant will receive compensation of approximately HK $500 (HK $1=US $0.128 as of July 15, 2026) for their participation in the study.

Participants will be randomized after the baseline assessment using computer-generated block randomization with a 1:1 allocation ratio within pre-established blocks of 90 participants. To ensure balance, a randomized blocking schema with block sizes of 2, 4, or 6 will be implemented. An independent research assistant, uninvolved in statistical analysis, will conduct the randomization. Allocation concealment will be maintained until the participant has completed eligibility screening, provided informed consent, and undergone baseline assessment. The allocation sequence will be accessible only to the independent staff member responsible for randomization. Research staff involved in recruitment and baseline assessment will not have access to the allocation sequence before assignment. Because of the nature of the VR intervention and the waitlist control condition, participants and therapists cannot be blinded after allocation. However, the statistician conducting the primary analysis will remain blinded to group labels until the primary analysis has been completed. Group labels will be coded as group A and group B during data cleaning and primary analysis to reduce analytic bias.

The trial will enroll 90 participants, allocating 45 to each arm. Allowing for 15% attrition, this sample provides 90% power to detect a treatment effect at T1, assuming an independent samples 2-tailed t test, an α level of .05, and Cohen d of 0.75. The target size also accords with comparable 2-arm VR intervention studies [16].

The first participant was enrolled in March 2026. The study will recruit participants from various settings in Hong Kong, including middle schools, tertiary institutions, and community centers. Adequate participant enrollment will be facilitated through partnerships with collaborating institutions, which will post digital flyers and promotional videos on their social-media platforms to reach the target population. The participants’ inclusion criteria were as follows: youth aged 15 to 25 years with a score of 25 or higher on the Social Phobia Inventory (SPIN). The exclusion criteria were as follows: (1) inability to attempt a baseline assessment (eg, due to being unpermitted to leave a psychiatric ward); (2) photosensitive epilepsy or significant visual, auditory, or balance impairment; (3) currently receiving another intensive psychological intervention; (4) a history of neurological conditions (eg, traumatic brain injury); and (5) insufficient language proficiency to understand the intervention requirements. In this trial, any endorsement of suicidal ideation will trigger crisis intervention procedures. Participants will be promptly flagged, and a trained clinician will notify the school or guardian to ensure appropriate safety measures are taken. School tutors or social workers familiar with students’ mental health will help identify participants unsuitable for the study, especially those whose conditions may be masked in self-report questionnaires. Participants will be recruited and screened online or offline, but will complete all interventions offline. Participants will be briefed through tutorial video clips demonstrating the research flow regarding intervention procedures, timing, and frequency.

Potential adverse effects related to VR exposure will be monitored throughout the intervention. These may include cybersickness symptoms such as dizziness, nausea, eyestrain, headache, or disorientation; temporary dissociative-like discomfort or a reduced sense of presence after leaving the immersive environment; and excessive emotional activation during anxiety-provoking social tasks. Before the first VR session, participants will be informed of these possible reactions and reminded that they may pause or stop the session at any time without penalty.

During each VR session, participants will be monitored through real-time behavioral observation, voice responses, task progression, and physiological indicators. Therapists will remain available to enter the VR room or provide support if the participant requests assistance or if the monitoring data suggest substantial distress. Sessions will be paused or terminated if the participant requests to stop, reports significant discomfort, shows signs of cybersickness, becomes unable to continue the task, or if the therapist judges that continuation may not be clinically appropriate.

After each session, therapists will conduct a debriefing to review the participant’s subjective experience, emotional responses, and any adverse reactions. Any adverse events will be documented by the research team and reviewed during supervision. Participants requiring additional psychological support will be referred to appropriate school, community, or clinical services.

A pilot study, conducted from January 2025 to April 2025, interviewed 26 individuals with SPIN scores of 25 or higher in Hong Kong to identify the most anxiety-provoking social situations and inform VR scenario design. Participants were asked: “What social settings do you fear the most, such as public speaking or interviews?” Based on their responses, the top 6 most anxiety-inducing situations were identified, including public speaking and job interviews. These findings will guide the development of VR scenarios that replicate realistic high-anxiety situations, ensuring the intervention effectively targets key challenges faced by individuals with social anxiety. Further details of the pilot study will be provided in a forthcoming article titled “Ecological Drivers of Youth Social Anxiety and Stakeholder Perspectives on Virtual Reality Intervention Design: A Qualitative Study in Hong Kong and Mainland China”.

To obtain a broader ecosystem perspective, interviews were conducted with 4 mental health professionals and 1 parent, eliciting views on adolescent social anxiety and their day-to-day challenges. The semistructured guide elicited participants’ views on adolescent social anxiety across five domains: (1) the current prevalence and presentation of the problem, (2) perceived etiological factors, (3) existing supports and resources, (4) gaps or areas in need of improvement, and (5) opinions on whether, and how, VR technology could assist youth with social anxiety.

According to the pilot results, the VR intervention was designed to address the unique characteristics of youth with social anxiety and incorporate suggestions from local mental health professionals. As a result, it stands out from existing VR interventions with 5 key innovative features:

Participants in the waitlist control group will have access to the VR intervention withheld until after the 3-month follow-up assessment. To prevent confounding effects from relevant concomitant care and interventions, participants will be required to refrain from receiving any other treatments or interventions specifically targeting social phobia during the study period. The rationale for this comparator is that this study is an early-stage RCT designed to estimate the preliminary efficacy, acceptability, feasibility, and safety of a newly developed therapist-facilitated VR intervention in a youth population in Hong Kong. Before conducting a more complex and resource-intensive active-controlled trial, it is necessary to first examine whether the intervention can be delivered as planned, whether participants can tolerate and engage with immersive exposure, and whether the study procedures, assessment schedule, and outcome measures are feasible. The waitlist design will also provide initial estimates of recruitment feasibility, adherence, dropout, outcome variability, and potential treatment signals, which can inform the design and sample size calculation of future trials. Providing delayed access also helps address ethical concerns by allowing control participants to receive the intervention after the primary assessment period.

Basic demographic and clinical data, including age, gender, region, and socioeconomic status, will be collected. The primary outcome will be the SPIN, while the secondary outcome will be the Patient Health Questionnaire-9 (PHQ-9) for depressive symptoms. All outcomes will be self-assessed using the following online measurements at baseline, 1-month (postintervention), and 3-month follow-up (Figure 4).

Findings will be disseminated through peer-reviewed publications, academic conferences, and knowledge exchange activities with collaborating schools, community organizations, and youth mental health stakeholders. Nontechnical summaries will be prepared for service providers and community partners to support future intervention refinement and implementation planning.

This study is expected to provide preliminary evidence on the efficacy, feasibility, and acceptability of a therapist-facilitated VR intervention for reducing social anxiety symptoms among youth in Hong Kong. We hypothesize that participants receiving the VR intervention will show greater reductions in SPIN scores than those in the waitlist or delayed-intervention control group at postintervention and at the 3-month follow-up. We also expect to obtain exploratory findings on depressive symptoms, adherence, physiological responses during VR exposure, perceived real-world application, and cost per unit reduction in SPIN scores.

Previous reviews suggest that VR exposure therapy may reduce symptoms of social anxiety disorder and may have effects comparable to other exposure-based approaches in some contexts [26,27]. More broadly, VR therapy has been most extensively studied for anxiety-related conditions and posttraumatic stress disorder, although evidence varies by diagnosis, population, and intervention format [7]. However, the existing literature still contains relatively few trials focused specifically on adolescents and young adults, and many studies differ in therapist involvement, scenario design, follow-up duration, and comparator conditions. Therefore, further randomized trials are needed to clarify how VR-supported interventions may be adapted for youth populations and specific sociocultural contexts.

This study contributes to this body of literature by testing a therapist-facilitated VR intervention that combines structured exposure tasks, therapist preparation and debriefing, culturally relevant social scenarios, physiological monitoring, and AI-supported feedback. Unlike fully self-guided VR interventions, the current intervention retains therapist involvement before and after VR exposure. This design allows therapists to help participants identify feared predictions and safety behaviors, prepare for exposure tasks, review their responses, and plan between-session homework. At the same time, the VR component offers a standardized and immersive environment in which exposure tasks can be repeated and adjusted in difficulty.

If the intervention shows preliminary efficacy and acceptability, the findings may inform the design of a larger trial with longer follow-up and a more rigorous comparator condition. Future studies should consider active control groups, such as therapist-led psychoeducation, nonimmersive digital interventions, standard CBT-informed support, or another credible psychological control condition. Such designs would help distinguish the specific contribution of VR from nonspecific effects, such as therapist attention, expectancy, and structured contact. The exploratory cost analysis may also provide initial information on the resources required to deliver the intervention, including therapist time, training, supervision, equipment, software, and administrative costs. However, this trial will estimate the cost per unit reduction in SPIN scores rather than establish full cost-effectiveness in routine clinical or community settings.

Several limitations should be acknowledged. First, given the multimonth research period, participant attrition was a concern. To mitigate this, participants were screened for motivation, supported throughout the study with dedicated staff and reminders, and the VR experience was refined based on pilot feedback to reduce technical issues and enhance engagement. Second, the technological nature of the VR intervention may limit generalizability, as youth more familiar with digital tools may benefit more than older or less tech-savvy individuals. Future research should explore ways to improve accessibility across age groups. Third, the disparity in content and intensity between the VR and control groups may limit interpretability, as symptom improvements could reflect therapist contact or session structure rather than VR-specific effects. Future studies should consider a 3-arm design to better isolate the unique contribution of VR. Fourth, the use of a waitlist control has important methodological limitations. Waitlist control conditions may produce larger apparent treatment effects than more active or usual-care comparators in psychotherapy trials, and they do not control for nonspecific therapeutic factors, such as therapist attention, expectancy, structured contact, or digital engagement [28]. Therefore, the findings will be interpreted as preliminary efficacy compared with delayed intervention rather than as evidence of VR-specific treatment effects. Future trials should include active comparator conditions, such as therapist-led psychoeducation, nonimmersive digital intervention, standard CBT-informed support, or another credible psychological control condition, to better isolate the unique contribution of VR and therapist facilitation.

In conclusion, this protocol describes an RCT evaluating the preliminary efficacy, feasibility, and acceptability of a therapist-facilitated VR intervention for social anxiety among youth in Hong Kong. By combining immersive VR exposure, therapist preparation and debriefing, physiological monitoring, and lightweight AI-supported prompts, the study will provide initial evidence on whether this intervention is associated with reductions in social anxiety symptoms compared with a waitlist control condition. The study will also generate exploratory findings on depressive symptoms, adherence, perceived real-world application, physiological responses, and cost per unit reduction in SPIN scores. Given the use of a waitlist comparator and a 3-month follow-up period, the findings should be interpreted as preliminary and will not establish VR-specific treatment effects, long-term effectiveness, or broad implementation outcomes. These results may inform the design of future active-controlled trials and longer-term evaluations of VR-supported psychological interventions for youth social anxiety.