Anxiety, depression, and chronic pain are among the most prevalent and disabling health conditions globally, affecting hundreds of millions of people and contributing significantly to personal, economic, and societal burden [1-5]. Digital therapeutics (DTx), particularly virtual reality (VR)–based interventions (virtual reality therapeutics [VRx]), are emerging as clinically validated, nonpharmacological treatment options for a range of behavioral and neurological conditions [6-14]. Evidence indicates that VRx can reduce symptoms of anxiety, depression, and pain across diverse populations, including patients with cancer, older adults with chronic disease, and individuals with pain syndromes, in some cases, approaching the effectiveness of standard care [9-12,14-22]. By leveraging mechanisms such as cognitive distraction, immersive meditation, and behavioral engagement, VRx offers a scalable, patient-centered approach to symptom management that may improve access, adherence, and patient experience [9,14,23-29].

Despite promising evidence and reported acceptability, widespread adoption of DTx and VRx remains limited [21,30-35]. Barriers include insufficient provider training, unclear prescribing pathways, cost concerns, and inadequate support for patient onboarding and sustained use [12,25,27,28,36-38]. Additional challenges include workflow integration, technology limitations, and lack of structured implementation strategies [10,17,21,27,33,35,39,40]. While user-related factors such as engagement and personalization have shown positive influence on the adoption of digital interventions [32,39,41], provider behavior, organizational context, and implementation strategy are increasingly recognized as critical but underdeveloped drivers [28,34,42-44].

Many existing VRx studies do not incorporate implementation frameworks or assess perceptual and behavioral outcomes, focusing primarily on symptom efficacy [37,44-50]. As a result, there is limited insight into how patients and providers interact with VRx in real-world contexts, particularly during independent at-home use [7,11,51-54]. While prior studies suggest that patients can use VRx at home with initial training and remote support, most evidence is derived from structured or guided programs, leaving gaps in understanding unsupervised use, adherence, and integration into routine care [11,43,51,54-57]. Addressing these gaps requires experimental evaluation of implementation strategies that more closely reflect real-world use, particularly with respect to provider involvement and support structures.

Research in related areas underscores the importance of implementation structure for digital interventions. Factors such as guidance, integration into daily life, and social connectedness are key influences on engagement with digital health interventions [39], while provider training, resource allocation, and reimbursement remain barriers to VR adoption in clinical settings [58]. Organizational buy-in and workflow integration are also critical for facilitating successful implementation [59]. However, these studies provide limited insight into provider involvement or delivery models for at-home VRx use. Collectively, while evidence supporting digital and VR-based therapeutic interventions is growing, practical strategies for enabling scalable, real-world adoption remain insufficiently defined.

Implementation science provides a useful framework to address these challenges. The Behavior and Acceptance Framework (BEAR) supports systematic identification of barriers and facilitators influencing digital health implementation [60]. It integrates behavioral change theory and technology acceptance concepts to evaluate feasibility, acceptability, usability, and fidelity as key constructs for translating efficacy into real-world adoption [60]. Despite rapid growth in efficacy-focused VRx research, there remains limited experimental evidence evaluating how specific implementation strategies such as the BEAR affect real-world adoption.

Despite demonstrated clinical efficacy, many digital health technologies fail to scale due to barriers in implementation rather than therapeutic limitations. A key gap remains in understanding how provider engagement and structured implementation strategies impact DTx adoption. This study addresses this gap by evaluating how different implementation approaches affect acceptability, usability, and fidelity of VRx among healthy individuals in nonclinical settings. Participants were randomly allocated to one of three conditions: (1) unguided self-use, (2) self-directed use with asynchronous support materials, and (3) provider-supported onboarding with simulated prescription and guided setup. Acceptability and usability were assessed using the technology acceptance model (TAM) [61-63] and System Usability Scale (SUS) [64], and qualitative findings from exit interviews were mapped to BEAR [60] domains to identify implementation-related barriers and facilitators.

By systematically comparing these conditions, this study is, to our knowledge, one of the first to experimentally evaluate the impact of provider-led support versus self-directed implementation strategies on VRx adoption. These findings provide early evidence to inform scalable models for integrating DTx into routine care, particularly by clarifying the role of clinician involvement and structured support in enabling real-world use.

The objective of this feasibility pilot study was to evaluate how different implementation strategies influence early adoption of a VRx intervention in a nonclinical setting. Specifically, the study compared unguided VRx use, self-directed VRx use with implementation materials, and provider-supported VRx use with implementation materials to assess differences in acceptability, usability, fidelity, and tolerability. We hypothesized that participants receiving clinician-supported implementation would demonstrate more favorable adoption-related outcomes than those receiving self-directed support alone or no additional support.

This was a 3-arm pilot study evaluating the acceptability, usability, and fidelity of a VR-based intervention (VRx) and accompanying implementation package across 3 experimental conditions. The study design was guided by the Virtual Reality Clinical Outcomes Research Experts framework [65], and the implementation package design was informed by the BEAR [60] and formative qualitative research with clinicians and end users. This study is reported in accordance with the CONSORT (Consolidated Standards of Reporting Trials) 2010 extension for pilot and feasibility studies [66]. The completed checklist is provided in Checklist 1. The CONSORT-eHEALTH (Consolidated Standards of Reporting Trials of Electronic and Mobile Health Applications and Online Telehealth) checklist [67] was also completed and is provided in Checklist 2.

Participants were allocated to 1 of 3 experimental conditions. As a pilot study, a formal power calculation for hypothesis testing was not performed. Consistent with guidance for feasibility research, approximately 10‐15 participants per study arm (total N=30‐45) were targeted for enrollment, which was considered sufficient to assess study procedures, characterize variability in key outcomes, and explore patterns across implementation conditions, while remaining appropriate for an early-stage study [68,69]. Accordingly, the study was not powered to detect statistically significant differences between groups, and findings should be interpreted as exploratory.

A mixed methods approach was used to provide a comprehensive assessment of how implementation strategies and provider engagement influence early adoption. Quantitative measures, such as task completion rates and structured questionnaires, assessed acceptability, usability, and fidelity, and enabled comparison across groups. Qualitative insights from exit interviews contextualized findings by capturing user perspectives, identifying usability challenges, and highlighting perceived benefits and areas for improvement.

Screening procedures were used to minimize the risk of adverse events. Individuals who did not meet the inclusion criteria were not enrolled. No adverse events were reported during the study. Cybersickness susceptibility was assessed during screening using the Motion Sickness Susceptibility Questionnaire—Short Form [72]. Participants with severe susceptibility to motion sickness were excluded. Prior to VRx use and immediately following, participants were screened using the Cybersickness in Virtual Reality Questionnaire (CSQ-VR) [73] to minimize cybersickness risk and assess symptoms. Participants were instructed to discontinue use if symptoms occurred, with study personnel present to ensure safety.

Participants were randomly allocated to 1 of 3 experimental conditions with a target 1:1:1 allocation ratio. The allocation sequence was generated in Microsoft Excel using the RAND() function prior to study initiation. Simple randomization was used; no formal blocking or stratification was applied. Assignments were made by study team members at the time of scheduling, with each enrolled participant assigned to the next available condition in the allocation sequence. Because recruitment, scheduling, and data collection occurred in parallel over a defined recruitment window (December 2024-February 2025), target group sizes were monitored during enrollment; when a scheduled participant did not attend their session, a subsequent enrolled participant was assigned to the affected condition to maintain approximate group sizes across arms. Self-reported technology comfort level and prior VR experience were monitored as baseline characteristics but were not used as formal allocation variables.

Allocation concealment from study personnel was not implemented, as schedulers required visibility of assignments to prepare condition-specific materials in advance of each session. Full blinding of participants and study personnel was not feasible, given the unavoidably visible differences between study arms, for example, the presence of a provider in condition 3 and the presence of physical implementation materials in conditions 2 and 3. To minimize expectancy effects, participants were not informed of their assigned condition until arrival for their study appointment. The absence of allocation concealment and blinding is acknowledged as a limitation.

The intervention consisted of art therapy delivered using the application OpenBrush on an Oculus Quest 2 headset. Participants were instructed to complete a series of drawing tasks over a 20-minute session. Further information about OpenBrush and selection of the application is provided in Multimedia Appendix 2.

Condition-specific materials (eg, comfort accessories, tutorials, chatbot, or provider consultation) were provided to participants as outlined in Table 1. Detailed implementation package elements, equipment lists, and task protocols are provided in Multimedia Appendix 3. Differences between experimental conditions across stages of implementation are summarized in Figure 1. Representative screenshots of the VRx application interface are shown in Figure 2.

The implementation package was developed based on prior qualitative research conducted by the study team exploring barriers to VRx adoption and was informed by the BEAR [60] and Virtual Reality Clinical Outcomes Research Experts framework [65] to address key implementation barriers across 6 stages of the implementation process. Package components included comfort accessories to improve device ergonomics, paper-based quick reference guides, a clinical evidence infographic, and an in-app step-by-step tutorial. The VR system menu was preconfigured for conditions 2 and 3 to display only the therapeutic application, reducing navigation complexity. A support chatbot was developed to provide real-time troubleshooting assistance. A complete description of all implementation package components, including photographs, equipment specifications, training documents, and the chatbot development process, is provided in Multimedia Appendix 3.

In the provider-supported condition, participants engaged in a standardized, scripted interaction delivered by a trained research team member with clinical training (eg, medical students). The interaction simulated a brief clinical encounter introducing VRx, including an explanation of the intervention, expected benefits, guidance on use, and a brief safety screening and device fitting. Interactions were time-limited (<5 minutes) and delivered using a predefined script to ensure consistency, as provided in Multimedia Appendix 3.

All individuals delivering the interaction were trained on the script and protocol prior to study initiation, and the script was followed closely to ensure standardized delivery. A checklist-based approach was used to ensure that all key components of the interaction were delivered consistently across participants. The provider consultation script and materials were developed based on prior qualitative research conducted by the study team exploring clinician perspectives on DTx and VRx and were designed to simulate a realistic clinical interaction.

Participants completed questionnaires at session initiation and immediately following intervention use. All measurement instruments administered in this study, including screening tools and outcome measures, are provided in Multimedia Appendix 4. Observational data, including task performance, errors, help requests, and adherence, were recorded by trained observers using a structured checklist (Multimedia Appendix 5). For conditions 2 and 3, the use of implementation system components and the time spent were also recorded.

Exit interviews were conducted immediately following the session to capture user perceptions, barriers, and facilitators related to intervention use.

Primary outcomes included acceptability, usability, and fidelity; tolerability was assessed as a secondary outcome. Validated measurement instruments and techniques were used to collect both subjective and objective data, which are summarized in Table 2.

This study was reviewed and approved by the UNC Institutional Review Board (IRB) at The University of North Carolina at Chapel Hill (IRB: 24‐2894). The study was determined to meet criteria for approval under federal regulations governing human subjects research, and all procedures were conducted in accordance with applicable ethical guidelines. All participants provided informed consent prior to enrollment. The consent process included a detailed explanation of the study purpose, procedures, risks, and the voluntary nature of participation. The informed consent form is provided in Multimedia Appendix 1. To protect participant privacy and confidentiality, all data collected were deidentified and stored securely using REDCap [70,71]. Personally identifiable information was removed from all research records, and data were stored without participant names. No identifying information is disclosed in any publications resulting from this research. Figures depicting the VR intervention (Figure 2) show application screenshots only, and no photographs of participants were included in the multimedia appendices. Participants did not receive financial compensation for their participation in this study. This study was not prospectively registered, as it was designed as a pilot feasibility study focused on evaluating implementation strategies and process outcomes (eg, acceptability, usability, and fidelity) rather than clinical efficacy or health outcomes. This approach is consistent with guidance indicating that formative and feasibility studies examining implementation processes and user perceptions do not require trial registration. All study procedures and outcome measures were predefined in the IRB-approved protocol prior to data collection, and no outcomes were modified after enrollment began.

Of the 39 participants enrolled, 31 (79.5%) completed the study and were included in the final analysis. In total, 8 participants did not complete the study due to logistical factors (n=6) or study-related factors (n=2), specifically the lack of monetary incentive.

Participants had a mean age of 38.6 (SD 12.3; range 18‐64) years, and the sample was predominantly female (20/31, 64.5%). Figure 3 presents a CONSORT-style participant flow diagram, and Table 3 summarizes the demographic characteristics by experimental condition.

Participant characteristics, including self-reported technology comfort and prior VR experience, were monitored during enrollment to assess balance across conditions. Participants were categorized as having low (16/31, 51.6%) or medium (15/31, 48.4%) technology comfort levels. Participants were assigned to experimental conditions according to a prespecified allocation sequence, with condition assignments made at the time of scheduling. Group comparisons indicated no significant differences in baseline demographic characteristics, including age, education level, or employment status.

Exploratory subgroup analyses examining associations between participant demographic characteristics and primary outcomes are reported in Multimedia Appendix 6. With the exception of acceptability scores, which differed significantly across age groups (χ²₄=10.1; P=.04; ηH²=0.235), no significant associations were identified. Given the small subgroup sizes and exploratory nature of this analysis, the age-related acceptability finding should not be interpreted as evidence of confounding and warrants further investigation in future adequately powered studies.

Fidelity outcomes by experimental condition are summarized in Table 11, including overall protocol adherence, time-based adherence, and task-based adherence. Mean fidelity scores were highest in condition 3 (mean 70.9, SD 11.4) relative to conditions 1 (mean 58.0, SD 16.5) and 2 (mean 57.5, SD 14.0). Time-based adherence differed across conditions (Kruskal-Wallis χ²₂=6.7; P=.04), with post-hoc comparisons indicating higher adherence in condition 3 relative to condition 1 (P.adj=.046). No other pairwise differences were observed.

Task-based adherence remained high across all conditions (mean 94.7%, SD 4.3%). No statistically significant between-group differences were observed for overall fidelity scores (Kruskal-Wallis χ²₂=5.4; P=.07) or task-based adherence (Kruskal-Wallis χ²₂=1.31; P=.52). Although the difference in overall fidelity scores did not reach statistical significance, the P value approached the conventional threshold (P=.07), suggesting a potential trend toward higher protocol adherence in condition 3 that warrants examination in future adequately powered studies.

Participants in condition 3 were less likely to explore the environment before completing assigned tasks when compared to conditions 1 and 2, suggesting greater adherence to the prescribed task sequence. No significant differences were observed for overall fidelity scores between conditions. Given the exploratory nature of these analyses and small sample size, findings should be interpreted as descriptive.

Participants completed the CSQ-VR prior to and following intervention exposure. Conditions 1 and 2 completed the assessment independently, whereas participants in condition 3 were administered the assessment as part of the VR Fitness Assessment conducted by the mock provider. Tolerability outcomes, measured using the CSQ-VR before and after the intervention, are presented by experimental condition in Table 12.

No meaningful increases in cybersickness symptoms were observed following VRx exposure. Mean preintervention symptom scores were low across conditions (mean 6.29, SD 0.69), corresponding to absent to very mild symptoms. Postintervention scores remained similar (mean 6.45, SD 0.96), with no evidence of symptom worsening reported across conditions.

Overall, 74% (23/31) of participants reported no symptoms following the intervention. Among those who reported symptoms, the most common were visual discomfort, dizziness, eye strain, and imbalance or instability. These findings suggest that the VR-based intervention was well tolerated in this sample under the implemented safety and screening procedures.

Qualitative insights are presented according to 14 themes derived from reflexive thematic analysis and mapped to BEAR [60] domains to contextualize implementation-relevant barriers and facilitators. Figure 6 presents key themes organized into 3 stages: onboarding, intervention perceptions, and engagement and sustained use, with representative participant quotes from exit interviews. The full BEAR thematic analysis mapping is provided in Multimedia Appendix 6.

Findings highlight several key factors influencing acceptability, usability, and fidelity. First, emotional responses during onboarding shaped early engagement trajectories. Participants across conditions described initial apprehension, particularly around device setup and controller use, that transitioned into enjoyment and relaxation once interaction with the application began. Participants who engaged in a provider-led consultation (condition 3) expressed higher levels of confidence and reduced uncertainty. Participants who received no support (condition 1) were more likely to describe frustration and discouragement during initial setup. Participants emphasized that the first interaction was a critical adoption threshold, noting that “first touch needs to be positive or people will walk away from using it, especially older people” (condition 1). Provider involvement was consistently identified as the most influential factor in shaping perceived safety, confidence, and motivation during onboarding, with participants describing cybersickness screening, device fitting, and expectation-setting as contributing to a greater sense of comfort and legitimacy.

Perceptions of VRx as a therapeutic modality were strongly influenced by how the intervention was framed and presented. Several participants initially viewed VR as a gaming platform, expressing skepticism about its legitimacy as a health intervention. Provider endorsement as well as evidence-based communication materials shifted these perceptions, with participants noting that professional framing made them take the intervention more seriously. A distinct barrier emerged around identity conflict, in which participants struggled to categorize VRx as medicine rather than entertainment, describing engagement with a game-like interface as therapeutic felt like self-deception without clinical framing.

Concerns related to data privacy and the use of consumer technology platforms for health interventions also emerged, with participants expressing preferences for offline functionality, local data storage, and health-dedicated devices. Limited prior awareness of DTx as a therapeutic modality was common across conditions.

Environmental constraints, such as limited home space, concerns about safety, and apprehension about the immersive properties of VR when others are in the home, shaped preferences for more structured settings and centralized locations where patients can go to use VRx outside of the home. For sustained use, participants emphasized the importance of flexible session length, integration into daily routines, and structured accountability and goal setting. Many expressed that without some form of structure, such as goal setting with providers or progress tracking, they would have difficulty maintaining regular engagement. Participants across conditions raised concerns about the affordability of VR hardware and suggested centralized access models such as clinic-based or community “studios” as alternatives for individual device ownership.

Collectively, these findings highlight the central role of provider involvement, structured onboarding, and contextual fit in supporting adoption and sustained use of VR-based DTx.

This pilot study evaluated 3 implementation strategies for VR-based DTx, comparing unguided use, self-directed use with support materials, and provider-supported use. A single brief exposure to VRx was associated with statistically significant improvements in acceptability across all conditions. However, postintervention acceptance scores remained within the neutral range, suggesting that while perceptions shifted favorably, a single exposure may be insufficient to produce strong endorsement or readiness for sustained adoption.

Differences across intervention conditions were observed in acceptability, usability, and fidelity-related outcomes. Participants receiving provider-led support demonstrated the most favorable and consistent pattern of favorable outcomes, including longer engagement times, higher protocol adherence, and more positive usability ratings. In contrast, self-directed materials produced more variable engagement and, in some cases, appeared to increase cognitive burden during onboarding. Tolerability was acceptable across all conditions, with no meaningful increases in cybersickness symptoms observed. Qualitative findings reinforced the importance of the initial interaction, clinician endorsement, and structured onboarding in shaping early adoption trajectories.

The findings suggest that early user responses to VR-based DTx are shaped not only by exposure to technology itself but also by the implementation strategy through which that exposure is delivered. While all participants demonstrated improved perceptions of VRx following a single exposure, the differences across conditions indicate that acceptability, usability, and fidelity are strongly influenced by the surrounding implementation approach rather than inherent properties of the intervention alone.

Participants in the provider-supported condition demonstrated the most consistent pattern of favorable outcomes, suggesting that structured onboarding and perceived clinical endorsement reduce uncertainty and increase confidence in use. In contrast, reliance on self-directed materials alone appeared insufficient to support early engagement and may introduce cognitive burden during initial use.

Notably, the disconnect observed between improvements in attitudinal acceptance and more modest shifts in behavioral intention suggests that early exposure may not be sufficient to drive adoption. This reinforces established technology acceptance theories, which suggest that early perceptions may shift quickly, while sustained intention and behavior require continued experience and reinforcement [62]. However, in contrast to prior VR studies that emphasize PU as the primary driver of adoption, our findings suggest that ease of use and structured onboarding may play a more prominent role during initial exposure, particularly for first-time users [79].

Findings in this study align closely with insights from other DTx studies where making DTx feel “ordinary” and familiar accelerated patient adoption and acceptance [42]. In this context, provider involvement and structured implementation strategies may serve as mechanisms that help bridge the gap between initial exposure and sustained engagement.

Findings consistently demonstrate that provider involvement plays a critical role in shaping early adoption of VRx. Participants receiving provider-supported onboarding showed greater improvements across acceptability, usability, and adherence outcomes, suggesting that structured human interaction enhances both perceived legitimacy and user confidence. This aligns with broader DTx literature identifying provider endorsement as a critical driver of patient uptake [30,42,80-82]. Our results extend this work by demonstrating that provider involvement influences not only adoption intent but also early usability, engagement, and adherence behaviors during initial exposure.

Provider-led activities, including safety screening, expectation setting, and device setup, appeared to reduce uncertainty and reinforce the therapeutic credibility of VRx. Participants reported greater willingness to engage when the intervention was introduced by a provider, consistent with prior research that provider endorsement enhances perceptions of safety, credibility, and ease of use [28,42,83].

The brief structured nature of the mock consultation suggests that effective implementation may not require a substantial time burden, but rather targeted and well-designed onboarding as part of a familiarization stage. This extends prior research in VR, which found that a familiarization phase before VR exposure reduced surprise effects [84,85]. Future research should evaluate scalable models, including hybrid approaches that combine brief human interaction with digital guidance.

In contrast, reliance on static materials alone was insufficient to support initial engagement. Participants in the self-directed condition relied solely on their own interpretation or ability to digest the information in a timely fashion. The lack of real-time support or personalized guidance may have led to uncertainty or cognitive overload, preventing participants from fully grasping the benefits and ease of using the VR independently. These findings are consistent with other implementation research that found the complexity of the implementation process and the amount of work involved to learn the technologies as barriers for successful adoption [25,86,87]. Future research should evaluate how different types of training delivery (eg, interactive onboarding, human-led instruction, or personalized digital guidance) influence the acceptability of VRx.

Taken together, these findings suggest that structured, human-supported onboarding may reduce early apprehension and improve engagement with VR therapeutics, particularly among first-time users. Provider endorsement remains a powerful lever for uptake, beyond reimbursement or access [88,89]. Notably, the provider interaction observed in this study was brief and structured, suggesting that effective implementation may not require a significant time burden but rather targeted, well-designed interactions. Future research should evaluate scalable models of clinician-supported onboarding, including hybrid approaches that combine brief human interaction with digital guidance.

Item-level usability analysis provided granular insights on important barriers to sustained adoption. Participants reported strong intention to use, PEU, and a low perception of system complexity. However, lower confidence, learnability, and a high perceived need for assistance indicate that early usability challenges persist even with implementation support [90].

Compared to other single-exposure VR studies reporting higher usability scores, the slightly lower usability observed in this study may reflect the broader scope of the user experience assessed, from initial device interaction through full session completion, rather than isolated in-app tasks [91]. This more naturalistic approach likely introduced variability reflective of real-world use conditions.

A notable qualitative finding was the distortion of time perception during VRx use. Several participants reported that the 20-minute session felt substantially shorter than its actual duration, suggesting a high degree of immersion and cognitive absorption. This aligns with prior research demonstrating the ability of VR to create immersive, attention-capturing experiences that support therapeutic mechanisms such as distraction, presence, and engagement [6,8,92,93]. The relationship between perceived time, immersion, prescription, and clinical outcomes in VRx should be further examined in future studies.

This finding also has important implications for fidelity, as users may be more likely to complete prescribed session durations when the experience feels shorter and more engaging. Related studies on VR use show that the level of immersion may support positive psychological responses, including increased intrinsic motivation, improved mood, and enhanced self-efficacy, which are known contributors to sustained engagement and adherence in rehabilitation and behavioral interventions over time [7,94,95]. Adherence improvements correlate with better clinical outcomes in many DTx studies [30,82]. Further research is needed to show the longitudinal impact of immersion on fidelity and adherence.

Cybersickness was minimal across participants and did not appear to meaningfully impact engagement or completion of the VRx session. While a small number of participants reported mild symptoms (eg, slight dizziness or discomfort), these were transient and did not lead to early discontinuation [96-98]. The low incidence of cybersickness in this study may also reflect awareness of the side effect from providing the cybersickness questionnaire prior to use. Cybersickness remains an important consideration for broader implementation, particularly in patient populations who may be more susceptible. Future research should evaluate tolerability over repeated exposures and across diverse clinical populations, as well as identify design and implementation strategies that further mitigate risk.

The 3 implementation strategies evaluated in this study represent different resource models with important implications for scalability. While provider-supported use produced the strongest outcomes, it introduces additional resource requirements that must be considered in real-world deployment.

The provider interaction in this study was brief (~5 minutes) and delivered by trained personnel, suggesting that effective support may be achievable without substantial burden. In many systems, physicians and psychotherapists act as the main prescribers and gatekeepers, deciding whether DTx are integrated into the treatment pathway [40,43,99]. There is an opportunity for future research to explore alternative delivery models, including nurse-led onboarding, digital-first guidance, or hybrid approaches combining asynchronous support with targeted human interaction.

Participants also identified cost and access as key barriers. Concerns about the affordability of VR hardware and uncertainty around insurance coverage were consistently raised. Suggested solutions included shared-device models, such as clinic-based or community access points, and integration into existing care pathways to reduce both cost and access barriers. Research supports shared-device models, shared digital access, and community or mobile access points as practical, relatively low-cost ways to reduce cost and access barriers, especially when tightly integrated into existing care pathways and supported by appropriate policy and management infrastructure [100-104].

Notably, previous research examining key barriers to DTx adoption showed that improving access to prescriptions and removing cost barriers did not accelerate adoption [34,40,43]. Further investigation found that provider endorsement of DTx significantly impacted patient adoption [43,47]. Adoption of DTx is shaped by intertwined patient, provider, organizational, and regulatory factors [28,42,80,105,106]. Lowering cost or easing prescription alone is insufficient; provider knowledge, trust, workflow fit, and institutional support are central levers for accelerating patient uptake.

While a formal cost analysis was beyond the scope of this study, the relative resource differences across conditions highlight the need for future work evaluating the cost-effectiveness of implementation strategies. Identifying the minimum effective level of provider involvement, as well as opportunities for digital or hybrid onboarding, will be critical for developing scalable models of VRx delivery.

Although provider training was not a primary focus of this study, trained study personnel acting as “mock providers” were able to deliver a brief, structured onboarding interaction that improved perceived safety, usability, and engagement. These materials were designed to be efficient and minimize time burden, addressing a known barrier for provider adoption: limited time during patient encounters [89]. This suggests that a short “introduce and demonstrate” interaction may be sufficient to support initial adoption without adding substantial time burden to clinical workflows.

Research related to prescribing methods for DTx is limited; methods for prescribing and ensuring patient safety are substantial gaps in how DTx and VRx are delivered today [88,89,107]. There is a substantial opportunity for this research to be expanded on to design simplified tools for patient screening and streamlining the clinical workflow to mitigate adoption barriers experienced by providers. Finally, lightweight screening, dose guidance, and simplified demo devices may further reduce provider friction and support safe, repeatable prescribing.

This study integrates complementary frameworks to demonstrate that VRx adoption is shaped by the interaction of user perceptions (TAM), usability experience (SUS), and implementation context (BEAR). While brief exposure improved PU and PEU, these shifts did not translate into strong acceptability, indicating that perception alone is insufficient for adoption [61].

Usability findings further highlight a gap between PEU and user confidence, suggesting that initial support may be required even for well-designed systems [64]. The BEAR contextualizes these findings by demonstrating that implementation strategies, particularly through provider involvement, reduce uncertainty and reinforce legitimacy [60]. This reinforces the role of implementation strategies as active drivers of engagement, not simply delivery mechanisms. These findings suggest that successful VRx deployment requires alignment across perception, usability, and implementation strategy to support both initial engagement and sustained use.

Taken together, these findings suggest that implementation strategy is a critical determinant of early adoption of VR-based DTx. By isolating onboarding and provider support components, this study extends prior VRx research that has largely focused on clinical efficacy to address how these interventions are introduced and supported. The results highlight the importance of integrating structured onboarding and provider touchpoints to enhance usability, engagement, and adherence. These findings have practical implications for the design and deployment of scalable VRx solutions in real-world settings.

This study has several limitations that affect the interpretation and generalizability of the findings. First, as a pilot study with approximately 10‐11 participants per condition, it was not powered to detect between-group differences. As a result, both null and statistically significant findings should be interpreted cautiously, and effect size estimates should be considered exploratory rather than definitive.

Second, multiple comparisons were conducted across TAM and SUS constructs, and exploratory subgroup analyses were performed without correction for multiplicity. While consistent with the exploratory nature of pilot studies, this increases the risk of type I error [69,77]. Accordingly, findings should be interpreted based on overall patterns rather than individual comparisons.

Third, the study was conducted in a controlled, nonclinical sample of healthy adults, which may limit generalizability to intended patient populations. Engagement, motivation, and usability may differ in clinical populations, where symptom burden and provider recommendation play a more central role. There was an imbalance in sex distribution across conditions, with a higher proportion of female participants in condition 3, which also demonstrated the most favorable outcomes. Prior research suggests differences in technology acceptance by sex, and this imbalance may represent a potential confounding variable. Simple randomization was used rather than stratified or block randomization. While baseline characteristics including technology comfort level were monitored across conditions, the absence of formal stratification may have contributed to residual imbalance, including a higher proportion of female participants in condition 3 and uneven distribution of technology comfort level across conditions. Allocation concealment from study personnel was also not implemented, and full blinding of participants and study personnel was not feasible, given the visible differences across study arms. Future adequately powered studies should incorporate formal stratification, allocation concealment, and, where feasible, blinded outcome assessment.

Fourth, provider involvement was simulated and may not reflect real-world prescribing, symptom-driven motivation, or clinical care pathways. In clinical populations, the presence of a diagnosed condition and provider recommendation may further influence engagement, adherence, and perceived value of VR therapeutics.

Fifth, differences between the laboratory and community-based study settings may have influenced usability and engagement. Future studies should evaluate these strategies in fully naturalistic, real-world settings to better understand their impact on adoption and sustained use.

Finally, the study evaluated a single, short-term exposure to VRx, precluding assessment of sustained engagement, long-term adherence, and clinical outcomes. Future research should examine longitudinal use in clinical populations and real-world settings to better understand adoption, adherence, and implementation at scale.

This study provides early evidence that the implementation strategy plays a critical role in shaping the adoption of VR-based DTx. This work is innovative in experimentally isolating implementation components, such as onboarding support and provider involvement, rather than evaluating the therapeutic content alone, which distinguishes it from prior VRx studies that have primarily focused on clinical efficacy. By directly comparing unguided, self-directed, and clinician-supported approaches, this study offers new insight into how implementation design influences early user adoption.

Findings demonstrate that while brief exposure to VRx improved user perceptions across all conditions, structured onboarding and provider involvement produced more consistent improvements in usability, engagement, and adherence. Notably, even brief provider interaction appeared to increase perceived legitimacy, reduce uncertainty, and support early engagement, suggesting a potential pathway for future scalable implementation through low-burden, hybrid onboarding models, warranting validation in larger studies.

These results contribute to the growing field of DTx by emphasizing that successful deployment depends not only on the technology itself but also on how it is introduced and supported. In real-world settings, integrating lightweight provider touchpoints and structured onboarding may improve uptake and sustained use of VRx interventions. Future research should evaluate these strategies in longitudinal and clinical populations to determine their impact on sustained adherence and clinical outcomes.