When immersive systems leave controlled settings and enter clinical practice, the frictions become visible. Spatial layout, device handling, hygiene routines, and tracking stability repeatedly appear as determinants that shape feasibility during deployment [12-14]. Recent reports from hospital-based XR programs add further texture, noting the need for clear protocols on headset management, staffing, patient selection, and environmental safety [15]. Rehabilitation, primary care, and mental health studies echo this pattern. Many show therapeutic potential, yet routine adoption remains uncommon [12,14,16]. Organizational constraints, workflow misalignment, limited training, and inconsistent technical support continue to slow implementation. Work in chronic pain paints a similar picture [17].

Across settings, the pattern is similar. Many implementation problems only become visible once XR is used in routine care, which leaves teams reacting to issues that could have been anticipated earlier. This gap makes it harder to evaluate feasibility, coordinate staff and equipment, or clarify how XR fits within existing care pathways. A more structured way of assessing readiness may help shift this work to the planning stage, when organizations still have room to adjust. To substantiate this landscape, Multimedia Appendix 1 provides a detailed comparative overview of the methodological, behavioral, and organizational determinants reported across the primary studies reviewed.

Attempts to explain how XR becomes embedded in clinical and educational settings have expanded quickly. Kouijzer et al [12] offer one of the most extensive syntheses and show how similar determinants recur across contexts, including clinician capability, workflow fit, organizational backing, and patient acceptability. They also note substantial variation in methodological quality, which complicates comparison.

In chronic pain, Elser et al [17] used the theoretical domains framework to highlight the weight of beliefs about consequences, emotional responses, and self-efficacy, yet their findings also imply that behavioral determinants alone cannot explain organizational pressures or service-level constraints. Work in education reaches similar conclusions. Lie et al [14] describe how training structures, institutional support, and professional identity shape the use of immersive tools.

Taken together, these scoping reviews show that the most relevant determinants have been identified, although they remain scattered across studies and have not yet been arranged into a structure that helps with planning or priority-setting.

In digital health, evaluation typically relies on established implementation science models. These frameworks illustrate how individuals, organizations, and technologies interact, building on behavioral determinants described by Michie et al [18], the organizational complexity mapped by Greenhalgh and Abimbola [19], and the processes synthesized in the consolidated framework for implementation research (CFIR) [20]. Broader overviews classify these models into determinant frameworks, process models, and theories, each covering a different part of the implementation landscape [21]. More recent work stresses the importance of understanding mechanisms of action, noting that evaluation depends on clarifying how behavioral, organizational, and contextual influences shape implementation outcomes [22].

These perspectives are valuable, although none was designed to address the experiential or embodied features of immersive systems. While specific guidelines such as the Reporting of Applications, Technologies, and Experiments in Extended Reality (RATE-XR) guideline [23] contribute essential descriptors for safety and reporting quality, they focus on technical transparency rather than organizational strategy. To place XR within this broader space, Table 1 summarizes the implementation and behavioral frameworks most relevant to the field, detailing their core purpose, main strengths, limitations in immersive contexts, and specific conceptual contributions to the MCDA-XR framework.

A smaller body of work has begun to move beyond single-theory evaluations by explicitly linking behavioral, organizational, and technological perspectives. Within the immersive field, Chung et al [30] combined the theoretical domains framework; capability, opportunity, motivation-behaviour (COM-B); the behaviour change wheel, and implementation strategy mapping to inform the implementation of VR in psychiatry. This work represents one of the first structured applications of behavioral implementation frameworks to XR. Its scope, however, remains largely confined to mental health settings and centers mainly on clinician behavior, with limited attention to organizational governance, resource allocation, or system-level decision-making.

Beyond XR, recent digital health research offers relevant methodological precedents for the integrative structure proposed here. Pereira Guerreiro et al [31] showed that non-adoption, abandonment, scale-up, spread, sustainability (NASSS) and COM-B can be applied in parallel to assess complex medical devices, suggesting that behavioral and organizational lenses can be meaningfully combined within a single evaluation process. In a complementary direction, Deason et al [32] demonstrated the feasibility of using multicriteria decision analysis to support virtual care selection, shifting the focus from narrative appraisal to structured and transparent scoring. Taken together, these studies point to a clear opportunity. While integrative and multicriteria approaches are already being used in broader digital health, they have not yet been translated into a unified governance framework tailored to the specific characteristics and risks of XR.

Despite their contributions, existing frameworks fall short when applied to immersive technologies. Behavioral models such as COM-B and the theoretical domains framework [18,28] clarify capability, opportunity, and motivation, but do not engage with the sensory, ergonomic, or embodied features that shape XR use [4-9,23]. Organizational frameworks such as CFIR and NASSS describe complexity and system dynamics [19,20], yet offer little guidance on perceptual mechanisms or user experience. Usability approaches assess interaction quality [13], but do not address institutional support or sustainability [12,14-17]. RATE-XR provides safety and reporting descriptors [23] rather than readiness or strategic planning criteria.

Each model captures a relevant part of the picture. What is missing is a way to integrate behavioral, organizational, experiential, and contextual determinants into a structure that supports comparison, prioritization, and planning. In practice, decisions about XR adoption require weighing clinical value, safety, workflow compatibility, organizational capacity, regulatory constraints, and cost. These dimensions rarely move in the same direction, and no single variable captures the trade-offs that determine whether an initiative is viable within a health care organization.

Across the literature, there is agreement that the determinants influencing XR adoption are broadly known, yet they remain dispersed across frameworks with different aims [12,14,17,30]. No current model captures behavioral, organizational, and experiential determinants within a single governance structure that can be applied explicitly to strategic portfolio management in digital health. Existing tools describe barriers, but they do not translate them into assessable criteria or link them to resource allocation, which limits their usefulness for planning.

A multicriteria decision approach offers a way to address this gap. By allowing organizations to weigh criteria according to local priorities, such approaches support comparison between initiatives that differ in purpose, maturity, or resource demand. An organization may prioritize safety during early piloting, focus on workflow integration during scale-up, or emphasize evidence requirements when projects compete for investment. Rather than relying on intuition, multicriteria methods make these trade-offs explicit and transparent.

To address this governance gap, this study describes the development and formative refinement of MCDA-XR, a multicriteria governance framework designed to support feasibility assessment, readiness evaluation, and strategic prioritization of XR initiatives within healthcare organizations.

Phase 1 aimed to identify the behavioral, organizational, and technical variables that define the strategic feasibility of XR initiatives. The work unfolded in 3 sequential steps that moved from theoretical framework selection to empirical evidence validation and final criterion expansion.

In the first step, theoretical models suitable for strategic decision-making were identified through an analytical literature review (Table 1). Four frameworks were selected for their complementary coverage of system-level, behavioral, perceptual, and XR-specific considerations. Their joint use followed established precedents for integrating behavioral, organizational, and decision-analytic perspectives in digital health (Table 2). Core domains from the selected frameworks were intersected to derive an initial set of provisional evaluative dimensions.

To test the deductive structure derived in Step 1, a targeted review of XR implementation studies was conducted. Multimedia Appendix 1 compiles the literature informing framework development, including both the foundational and methodological precedents used to support the conceptual structure of MCDA-XR [12,13,15-17,30-32] and the targeted empirical corpus of XR implementation studies used for determinant extraction in Step 2 [33-45]. Determinants were defined as explicit barriers, facilitators, or requirements related to XR implementation and were extracted exclusively from the “Results” and “Discussion sections to preserve analytical fidelity.

This process used a large language model (LLM)–assisted workflow using ChatGPT (OpenAI GPT-family models) within a constrained extraction protocol designed to identify source-grounded verbatim segments. All extracted items generated with the final prompt were subsequently checked manually against the source documents to confirm attribution accuracy and verbatim fidelity. The resulting extraction matrix, together with the full extraction protocol and prompt, is provided in Multimedia Appendix 2.

Following extraction and verification, a thematic synthesis was performed within each evaluative dimension using a second constrained LLM-assisted workflow. Extracted determinants were examined to identify recurring patterns and higher-order themes, with conceptually overlapping determinants grouped through rule-guided semantic clustering. All outputs were manually reviewed to confirm attribution accuracy, conceptual consistency, and alignment with the original determinant set and study context. Themes were classified according to their frequency across studies and their potential strategic impact, with particular attention to safety, ethical, and regulatory issues.

This synthesis informed the refinement and expansion of the analytical structure. Multimedia Appendix 3 documents the thematic synthesis procedure, including the analytical prompt, clustering logic, and resulting evidence matrix by criterion. Multimedia Appendix 4 then presents the final theoretical-empirical alignment used to define each criterion operationally and to support the participatory appraisal and refinement activities conducted in Phase 2.

Phase 2 examined how the refined evaluative dimensions functioned when translated into a practical decision-support instrument. The focus was on interpretability, perceived strategic relevance, and operational usability across professional profiles and care settings.

Participants were recruited purposively to capture a range of perspectives relevant to early appraisal of XR implementation. Eligibility for this phase required attendance at one of the structured group sessions and completion of the prototype MCDA-XR exercise. Four structured group sessions were conducted within the BSA living lab environment and collaborating institutions. Each session followed a common format. Before completing the appraisal exercise, participants received a brief introduction to XR use in health care and a live demonstration of immersive applications used in clinical and educational projects. They were also introduced to the 10 criteria through their guiding questions and illustrative implementation examples, such as cybersickness, workflow compatibility, and data privacy, derived from the operational definitions developed in Phase 1 and documented in Multimedia Appendix 4. This introductory segment was intended to provide a common experiential reference point for the subsequent evaluation part, but it did not constitute formal XR training or extended hands-on use.

Participants then completed a prototype version of the MCDA-XR instrument, rating the importance of each criterion using a 5-point scale and providing brief qualitative comments. These comments were used to identify ambiguities, assess interpretability, and indicate whether any relevant implementation criterion appeared to be missing from the proposed structure.

A total of 33 participants contributed to the Phase 2 construction and rating process. The sample comprised practicing health professionals (eg, medicine, nursing, physiotherapy, and psychology), operational and managerial staff (eg, service leadership and administration), alongside postgraduate psychology students. This composition allowed the criteria to be examined from both an applied organizational perspective and a behavioral science perspective during this formative stage. Detailed participant characteristics are reported in the “Results” section.

This phase was intentionally formative. Its objective was to refine clarity and usability rather than to establish psychometric validity. The combined dataset of quantitative ratings and qualitative feedback informed iterative adjustments to wording and structure, preparing the instrument for empirical testing in Phase 3.

Qualitative material consisted of narrative comments generated during the construction activities and additional documentary sources from exploratory stakeholder work. These data reflected how participants interpreted each criterion and related it to their organizational context. Analysis followed Braun and Clarke’s [46] reflexive thematic approach, combining inductive identification of emergent insights with deductive mapping informed by NASSS and COM-B. The aim in this phase was to refine conceptual wording and identify structural ambiguities requiring clarification.

Quantitative data consisted solely of importance ratings for the 10 criteria. Descriptive statistics, including means and SDs, were calculated for the full sample and for subgroups comparing practicing professionals versus postgraduate psychology students, and participants with no prior XR experience versus those with any prior XR experience. These summaries were used to describe rating patterns across participant profiles and to inform iterative refinement of the instrument, but were not used for inferential analysis. Hypothesis testing and validation of scoring behavior are planned for Phase 3, where the MCDA-XR protocol will be applied in real implementation scenarios.

The study was approved by the UOC Research Ethics Committee (CE25-TE78). Participation is voluntary and all contributors provide informed consent. Information is pseudonymized before analysis and no patient-level data are included. Experiential insights from XR projects are handled in line with the approved protocol and comply with General Data Protection Regulation (GDPR) and Organic Law on Protection of Personal Data and Guarantee of Digital Rights (LOPDGDD) requirements.

Reflexive notes were maintained throughout to monitor the potential influence of the researcher’s embedded role within the organization, consistent with guidance for reflexivity in applied health research.

Phase 1 resulted in the transformation of an initial deductive structure into a governance framework composed of 10 operational criteria. This evolution reflects how theoretical assumptions were reshaped when tested against empirical evidence from XR implementation studies.

The evolution of the framework clarified how different theoretical perspectives held up under empirical testing. NASSS primarily structured system- and organization-level considerations, while COM-B and technology acceptance model (TAM) informed criteria related to adoption, usability, and behavioral feasibility. RATE-XR contributed safeguards linked to safety, transparency, and technical integrity. Empirical mapping confirmed the relevance of these contributions, but also showed that governance, evidence legitimacy, and regulatory alignment could not remain implicit within organizational domains and required explicit representation as independent criteria.

The initial synthesis produced 7 provisional criteria addressing relevance, safety and comfort, usability, workflow integration, resources and cost, training requirements, and patient acceptability. Together, these criteria captured core condition, technology, organizational, and adopter-related considerations relevant to XR adoption.

When empirical determinants from the literature were mapped onto this structure, clear limitations became apparent. A substantial proportion of extracted determinants could not be meaningfully accommodated within the original criteria without stretching their scope. In particular, issues related to evidence quality, institutional commitment, and regulatory or ethical compliance repeatedly emerged as distinct considerations rather than secondary aspects of organizational readiness.

To resolve these structural mismatches, the framework was expanded to include 3 additional criteria: evidence and credibility, institutional support, and legal and ethical alignment. Their inclusion marked a shift from an implicitly organizational interpretation of system-level factors to an explicit recognition of governance, legitimacy, and accountability as independent strategic determinants. The resulting 10-criterion structure provided a more faithful representation of the empirical landscape observed across XR implementation studies. The final allocation of determinants to criteria is documented in the extraction matrix (Multimedia Appendix 2).

Following the establishment of the 10-criterion structure, determinants within each domain were examined to identify recurring patterns and higher-order themes. Themes were classified as major when supported by 3 or more independent studies and as critical when directly related to safety, ethical, or regulatory requirements, regardless of frequency.

This synthesis informed the refinement of the scope and wording of each criterion, ensuring conceptual clarity and practical interpretability. Multimedia Appendix 3 presents the thematic structure by criterion, while Multimedia Appendix 4 provides the final operational definitions and example indicators that guided the framework construction activities in Phase 2.

Phase 1 concluded with the definition of the MCDA-XR framework, comprising 10 operational criteria intended to guide strategic decision-making on XR adoption in health care.

The resulting structure is formalized as a governance matrix that brings behavioral, organizational, and technical considerations into a single analytical space. Rather than treating these dimensions as parallel or sequential factors, the matrix reflects how they interact when XR initiatives are assessed, prioritized, and implemented in real-world settings.

This synthesis does not seek to replace existing models. Its purpose is to translate their combined insights into a pragmatic diagnostic tool that makes misalignments visible, particularly those between technological readiness, workforce capability, and organizational or policy constraints. Figure 2 illustrates the structure and logic of the MCDA-XR governance matrix.

The 10 MCDA-XR criteria fall into 3 broad layers of implementation. One layer concerns people, their capability to use XR, and the conditions that make its use realistic in daily care. Another relates to organizational structures, workflows, and resource constraints. A third concerns the technology, its safety, usability, and alignment with sector standards such as RATE-XR.

These layers are not treated as isolated components. They help identify where tensions or enablers usually appear as teams move from conceptual ideas to operational deployment.

Table 3 summarizes the 10 criteria and shows how each one aligns with the NASSS, COM-B, TAM, and RATE-XR constructs. The table also provides the guiding questions that form the basis of the decision-support instrument. Within MCDA-XR, these concepts are defined operationally through the criterion wording and guiding questions. For example, safety refers to tolerability and comfort during XR use, usability to the manageability and reliability of the system in practice, and patient experience to trust, enjoyment, and acceptance as part of care.

To support traceability, the empirical themes that informed these operational definitions are presented in Multimedia Appendix 3, and the detailed conceptual-empirical mappings grounding each criterion in the NASSS, COM-B, TAM, and RATE-XR constructs are provided in Multimedia Appendix 4. Together, these materials document the analytic steps that link the thematic synthesis to the final structure summarized in Table 3.

A total of 33 participants contributed importance-weighting data. The sample was predominantly female (28, 84.8%). Ages ranged from 18 to 69 years and were distributed as follows: 18‐29 (12/33, 36.4%) years, 30‐39 (6/33, 18.2%) years, 40‐49 (6/33, 18.2%) years, 50‐59 (5/33, 15.2%) years, and 60‐69 (4/33, 12.1%) years. Reflecting the study’s aim to capture a broad cross-section of the health care ecosystem, the sample included 21 (63.6%) practicing health professionals from primary care, intermediate care, mental health, and neurorehabilitation services, alongside 12 (36.4%) postgraduate psychologists enrolled in a Master’s program in Psychogerontology. Detailed characteristics, including specific institutional origins (hospital, primary care, and university settings) and professional roles for all participants, are provided in Multimedia Appendix 5. Prior exposure to XR was limited: 16/33 (48.5%) participants reported no previous experience, 13/33 (39.4%) recreational-only use, and 4/33 (12.1%) had been involved in a clinical pilot.

Participants rated all 10 MCDA-XR criteria on a 1‐5 scale, where 1 reflected the lowest strategic importance and 5 the highest. In the pooled Phase 2 sample, mean ratings ranged from 4.03 (SD 0.95) for workflow integration to 4.61 (SD 0.56) for safety and comfort (Figure 3). Safety and comfort received the highest mean score (4.61) and showed the lowest variability (SD 0.56), indicating the highest level of agreement among participants. Greater variability was observed for institutional support (SD 1.06) and legal and ethical alignment (SD 0.97), reflecting a wider spread of ratings across the sample.

Descriptive subgroup analyses are presented in Table 4. Overall, rating patterns were broadly similar across participant profiles, with all criteria remaining in the upper range of the scale. Some differences in relative emphasis were observed. Compared with postgraduate students, practicing professionals gave higher ratings to workflow integration, training requirements, and institutional support, whereas students rated relevance and usability slightly higher. When ratings were examined according to prior XR exposure, participants without prior XR experience gave higher ratings to evidence and credibility, legal and ethical alignment, and institutional support than those with any prior XR experience. Given the small subgroup sizes, these comparisons are descriptive only.

Open-text comments were brief and generally positive, noting perceived clinical usefulness, patient engagement, and comfort. Remarks concerning workload and the need for protected time aligned with the existing criteria of workflow integration, institutional support, and resources and cost. No additional domains were proposed during the sessions, and participant comments did not suggest obvious omissions within the 10-criterion structure in this sample. These observations informed retention of the 10-criterion structure and its subsequent refinement ahead of prospective testing in Phase 3.

Following the refinement of the 10 criteria and their prioritization by stakeholders in Phase 2, the MCDA-XR framework was translated into an initial decision-support format for strategic preimplementation use. The final instrument organizes the framework into a dual-score structure that distinguishes between institutional priorities and current implementation conditions.

For each of the 10 criteria, stakeholders first rate strategic importance on a 5-point scale (1-5). This step captures how critical each dimension is within the local organizational context, independently of any specific XR solution. Importance ratings reflect institutional values, service objectives, and strategic goals, making explicit which considerations carry the greatest weight in adoption decisions.

Stakeholders then assess organizational readiness for each criterion on a 3-point scale (1-3), indicating the extent to which current conditions, resources, and capabilities are in place to support implementation. Readiness scores reflect practical feasibility, including workflow compatibility, staff capacity, governance arrangements, and technical preparedness.

Rather than collapsing these dimensions into a single aggregate score, MCDA-XR relies on their joint interpretation at the level of each criterion. At this stage, the matrix does not impose validated or universal decision thresholds. Instead, prioritization is based on the relative position of each criterion across the importance and readiness axes, as illustrated in Figure 4. Criteria with high importance and high readiness indicate areas that are well aligned for implementation. Criteria with high importance and low readiness identify priority gaps, where preparatory action is likely required before implementation or scale-up, such as targeted training, workflow adaptation, clarification of governance responsibilities, or allocation of additional resources. Criteria with low importance and high readiness indicate areas in which organizational conditions are already favorable, but which are not currently strategic priorities. Criteria with low importance and low readiness suggest lower-priority constraints that may be acceptable or manageable in the short term.

Figure 4 illustrates the application format of MCDA-XR. The left panel shows the evaluation scorecard where stakeholders rate the 10 criteria on strategic importance and organizational readiness. The right panel shows the cross-score interpretation matrix, where these 2 dimensions are plotted to distinguish priority gaps, defined as high importance and low readiness, from implementation-ready conditions, without collapsing the appraisal into a single aggregated score.

This importance-readiness crossing moves the framework beyond a descriptive checklist toward a structured preimplementation appraisal format. In practical terms, governance is operationalized by using this matrix to inform whether an XR initiative can proceed under current conditions, requires targeted preparation, should be adapted before deployment, or should be deferred until priority gaps are addressed. It supports informed discussion among multidisciplinary stakeholders by making trade-offs explicit and by distinguishing between aligned areas, priority gaps, and more manageable constraints under current organizational conditions. This scoring logic provides the initial operational basis of MCDA-XR, while its further refinement and prospective validation, including its translation into prioritized preparatory actions, are being addressed in Phase 3.