We conducted a hybrid type 2, nonrandomized, phased implementation study conducted in public primary care centers (PCC). The rollout was structured in three phases; this paper reports on Phases 1 and 2, which examined feasibility and optimization, while Phase 3 is ongoing and will be reported separately.

Eligibility was limited to adults enrolled in chronic pain group programs routinely offered at the PCCs. Participants had experienced chronic pain for at least 3 months, most often related to fibromyalgia, with no restrictions on pain site or intensity. Exclusion criteria, based on diagnoses documented in the medical record, included severe psychiatric or cognitive disorders, epilepsy, major neurological or sensory impairments, contraindications for physical activity, or barriers that prevented regular attendance at group sessions.

Recruitment was coordinated by the physiotherapist and psychologist leading the groups at each center. Patients invited to join the chronic pain groups were informed that the program included the option of using XR as part of our research study. They could choose to participate with or without the XR component and were free to continue in the group even if they later decided to withdraw from the study. In routine practice, referrals to the chronic pain groups were generally based on patients presenting features such as central sensitization, emotional distress, kinesiophobia, or limited response to prior treatments, although these factors were not applied as formal inclusion criteria for the study. At a dedicated evaluation visit, eligibility was confirmed by the physiotherapist or psychologist responsible for the group, and written informed consent was obtained together with baseline data. All patients who were invited chose to participate.

As the intervention was embedded in standard care, no randomization or blinding was applied. No formal sample size calculation was conducted, as these phases focused on feasibility and to inform the design of subsequent groups. The sample size was pragmatic, determined by patient enrollment in chronic pain groups during the study period. A minimum of 30 participants was considered sufficient to capture feasibility issues and guide refinements for broader implementation.

The intervention was delivered by psychologists, physiotherapists, and family physicians routinely leading group-based pain management programs. They received prior training and continuous technical support from the institutional innovation team, who had specific expertise in implementing XR in clinical care and codesigned the intervention content.

XR content was applied for 15‐20 minutes per session, with duration adjusted to the therapeutic focus and participants’ tolerance. All modules were delivered under direct facilitator supervision. Headsets (Meta Quest 2 and 3) were preloaded with the required content, sanitized between uses, and configured individually. Quest 3 devices were prioritized for AR passthrough sessions when available, given their superior visual quality.

To support acclimatization among first-time users, XR elements were introduced in a stepwise sequence. Early sessions began with a passive, audio-guided VR body scan, after which participants progressed to psychoeducational and mindfulness modules, and later to AR passthrough motor exercises and interactive tasks. This gradual progression reflected strategies already tested in other XR initiatives within our institution, where it proved effective in improving tolerability and reducing cybersickness [18].

The codesign process resulted in 2 integrated components that became the stable structure of REDOCVR sessions (see Figures 1 and 2). The XR content targeted 4 core domains relevant to chronic pain care in primary care: mindfulness and attentional focus; cognitive reframing and self-compassion; distraction for symptom regulation; and exercises to encourage safe, engaging movement.

All immersive content was available offline, preloaded in the headsets. Apps did not collect or transmit data; the only element requiring Wi-Fi was the optional Reality Telling tablet app for synchronizing playback of 360° videos.

As this was an early-phase study, no explicit thresholds for feasibility or clinical outcomes were prespecified. The primary aim was to assess usability, tolerability, and feasibility of implementation under routine primary care conditions. The results were interpreted descriptively to guide adjustments in program delivery and inform the design of subsequent implementation phases, rather than to determine continuation or termination against predefined criteria.

Ethical approval was obtained from the Institut Universitari d’Investigació en Atenció Primària Jordi Gol Research Ethics Committee for both study protocols (refs. 24/047-P and 24/211-ACps), which are registered at ClinicalTrials.gov (NCT06361706 and NCT06964360). The REDOCVR XR content was implemented only as nontherapeutic support within a group psychoeducational program. After review by the Spanish Agency of Medicines and Medical Devices, it was formally classified as nonmedical device software, and no regulatory authorization was required.

All participants provided written informed consent before enrollment. Study data were deidentified prior to analysis, and no personal or health identifiers were collected by the XR software. Participants did not receive compensation for their participation. Figure 2 includes images where participants and clinicians are partially identifiable; written informed consent for publication of these images was obtained from all individuals depicted and archived by the research team.

A total of 42 participants were enrolled across 7 intervention groups at 3 PCCs during 2024. Of these, 36 (85.7%) participants completed the full intervention and both pre- and post-assessments, with no missing outcome data among completers. Analyses of clinical outcomes were based on this final sample (n=36), with a mean age of 57.7 years (SD=9.36; range 40‐79) and a predominance of women (97.2%, n=35). Adherence was high, with the 6 dropouts attributed to patients’ logistical or scheduling conflicts, with no study-related reasons reported. Figure 3 illustrates a CONSORT-style participant flow diagram, while Table 1 details group composition across phases, from initial optimization to the later inclusion of supervised medication tapering.

Normality assumptions determined the choice of test: paired-sample t tests were applied to normally distributed continuous outcomes (WEMWBS, HADS, CSI; Table 4), while Wilcoxon signed-rank tests were used for ordinal EQ-5D-5L domains (Table 5).

Table 6 presents pre-post values for clinical outcomes (n=36). Emotional well-being (WEMWBS) increased from 16.9 (SD 4.9) to 21.7 (SD 6.1), with a mean change of 4.8 (95% CI 2.9‐6.7, P<.001, Cohen d=0.86). Anxiety symptoms (HADS-Anxiety) decreased from 13.1 (SD 4.3) to 10.7 (SD 4.3), mean change –2.5 (95% CI –3.8 to –1.2, P<.001, Cohen d=0.66), and depressive symptoms (HADS-Depression) from 10.6 (SD 3.4) to 9 (SD 4.2), mean change –1.6 (95% CI –2.5 to –0.7, P<.001, Cohen d=0.62). Central sensitization (CSI) declined slightly, from 60.7 (SD 14.2) to 57.5 (SD 13.4), mean change –3.2 (95% CI –7.2 to 0.7, P=.11, Cohen d=0.28).

For quality of life (EQ-5D-5L), Mobility improved significantly (median change –1.0, 95% CI –1.0 to 0.0, P=.02, r=0.39), while Pain/Discomfort showed a nonsignificant trend toward improvement (median change –0.5, 95% CI –0.5 to 0.0, P=.08, r=0.29). Self-Care, Usual Activities, and Anxiety/Depression domains showed no significant changes.

This study examined the early implementation of REDOCVR, an XR-enhanced group program for chronic pain in primary care. The findings suggest that immersive technologies can be integrated into existing care structures without disturbing routine delivery.

Satisfaction, tolerability, and usability indicators provide additional insight into how the program was received. Patients rated the sessions positively, reporting high satisfaction and minimal discomfort. Cybersickness was rare, and overall headset tolerance further supported the acceptability of XR in this population. The adapted 8-item SUS for patients reduced response burden while maintaining validity, consistent with prior analyses of SUS modifications [21-24,33]. In contrast, professional SUS scores showed low internal consistency, likely due to the small and heterogeneous sample. Such factors are known to reduce the stability of Cronbach α [34,35], whereas prior studies suggest that 10-item adaptations generally retain reliability [21,22]. Our findings, therefore, probably reflect contextual heterogeneity rather than flaws in the instrument itself.

An exploratory review of the open comments provided complementary insights into participants’ and professionals’ experiences with XR. Feedback suggests that XR was experienced less as a novel intervention in itself and more as a support that appeared to enhance engagement with the program’s therapeutic content. In psychology-led sessions, immersive tools supported mindfulness and self-compassion practices that some patients otherwise found difficult to understand. In physiotherapy-led sessions, AR passthrough helped participants perform movements in a safe and supervised way, while VR increased immersion and motivation, both encouraging patients to attempt physical activities they would normally avoid.

Exploratory clinical findings suggested improvements in emotional well-being, anxiety, and depression, with effect sizes in the moderate-to-large range. Statistically significant improvements were observed in Mobility on the EQ-5D-5L, while Pain/Discomfort showed only a nonsignificant positive trend. Among participants enrolled in the deprescription protocol, benzodiazepine, opioid, and NSAID intake decreased, and several participants achieved complete discontinuation. Although causality cannot be inferred, these results are consistent with the possibility that group-based interventions may complement structured deprescription strategies [36].

While our findings on patient usability and clinical outcomes align with the existing literature [10,37], the primary contribution of this study is its focus on pragmatic implementation within a public primary care setting. We have encountered barriers in early adoption similar to other reports related to technical setup and workflow adjustments [38]. Prior reviews have noted that most VR interventions for pain concentrate on supporting specific aspects of health literacy or self-management rather than integration into routine care [39]. By embedding XR within an established group program rather than running it in parallel, REDOCVR addresses adoption challenges noted in recent reviews [14,15,40] and shows how immersive tools can fit into everyday service delivery. Delivering XR in group sessions brings clear advantages, including peer support, efficiency, and easier integration into existing care pathways, though it comes at the expense of individualized tailoring. Meta-analytic evidence in chronic pain indicates that group-based approaches can achieve outcomes comparable to individual formats [41]. This appears particularly relevant in the context of the public health care system, where resource optimization is not only desirable but often decisive for adoption.

The improvements in emotional well-being and functional outcomes align with the psychoeducational and behavioral mechanisms targeted, consistent with results from mindfulness- and acceptance-based approaches [8,13]. The reduction in medication use also mirrors outcomes from other multidisciplinary primary care interventions that combine patient education with supervised tapering [5]. What our findings add is preliminary evidence that such effects may be achievable when immersive tools are embedded into routine group care, rather than tested only in parallel or experimental conditions.

The intervention was delivered to a modest number of groups within a supportive institutional setting, which may limit direct transferability to other contexts. Resources such as trained staff, headset availability, and organizational readiness facilitated implementation and should be considered when adapting the model elsewhere. Data collection on implementation fidelity was pragmatic. For instance, the dose of immersive exposure was guided by a time range (15‐20 min) and adjusted for tolerance, but not strictly logged for each individual patient in every session. Adherence was assessed through attendance records, while adverse effects were monitored through informal therapist feedback and a single post-intervention questionnaire, rather than a structured, session-by-session register. The qualitative component was restricted to an exploratory review of written comments. While this offered useful contextual insights for program refinement, it was not intended as a formal qualitative study.

Other limitations are typical of early-phase implementation research, and exploratory clinical outcomes should be interpreted with caution. The absence of randomization and a control group limits causal inference. Generalizability is constrained by the modest and predominantly female sample (97%), which reflects the higher prevalence of chronic pain among women and the predominance of female participation in primary care group interventions in Spain [2,42]. This imbalance may restrict the applicability of findings to male patients. Participants also represented a heterogeneous chronic pain population, and we did not stratify by phenotype (nociceptive, neuropathic, or nociplastic), which may have masked differential responses. The group delivery model, while resource-efficient and well-suited to public primary care, inevitably restricts opportunities for individual tailoring [40]. The use of self-reported measures introduces potential recall and social desirability bias. Positive findings may have been partly influenced by a novelty effect. Finally, the 5-month follow-up is insufficient to establish the durability of changes over time.

Despite these constraints, the findings have already informed practical adjustments. Feedback from patients and professionals in Phases 1 and 2 reinforced the institutional support. Patients reported high satisfaction and usability, while professionals highlighted integration and workflow challenges. Combined with the clinical findings, these insights guided the Phase 3 refinements. The program was refined in collaboration with the full clinical team, leading to optimizations to improve feasibility.

In physiotherapy sessions, AR proved more suitable for larger groups, where supervision and interaction needed to be fluid, while VR was reserved for smaller groups. Both patients and professionals clearly favored hand-tracking, and all VR content was adapted accordingly, eliminating the need for controllers in the VR exercise games. Synchronizing 360° videos from the tablet required stable Wi-Fi, which was not always available. External routers were used to compensate, adding setup complexity. Efforts are underway to improve connectivity in PCCs and reduce facilitator workload.

Informed by early-phase findings and aligned with Catalan health system guidelines, the program will now follow the conventional format of 12 weekly sessions co-led by a psychologist and a physiotherapist, with selected sessions involving family physicians and nutritionists to reinforce healthy habits. XR use has been reorganized to fit within this structure, concentrated into 4 sessions—2 psychology-led (mindfulness and self-compassion) and 2 physiotherapy-led (motor activation). This targeted approach preserves therapeutic value while optimizing headset use and technical support across centers.

Future evaluation should focus on systematic data collection related to implementation fidelity. With XR use now organized into defined modules, session-by-session monitoring will allow more precise analysis of exposure, usability, and adverse effects across different modalities (eg, VR vs AR, passive vs interactive content). The expansion to additional centers and teams will be accompanied by structured pre-implementation focus groups and complementary usability strategies. Combined with established implementation frameworks, these steps are expected to generate the structured evidence needed to support broader adoption. By linking fidelity monitoring with pragmatic implementation, REDOCVR may help close the gap between immersive research and everyday chronic pain care.

The first 2 phases of REDOCVR suggest that a person-centered, codesigned XR program can be incorporated in primary care chronic pain groups under real-world conditions. The intervention showed high adherence and favorable usability, and both patients and professionals valued its contribution to engagement and session delivery. Exploratory improvements in emotional well-being and reductions in high-risk medication use with supervised tapering are promising but should be interpreted with caution given the feasibility design.

This study is, to our knowledge, the first to develop and evaluate an XR program entirely within primary public health care. The findings offer practical guidance for refining XR-based interventions and preparing for wider scale-up. Ongoing work on Phase 3 across additional centers will further clarify long-term outcomes, scalability, and adaptability to different settings.