The SMART-OR Framework for Implementing Artificial Intelligence in the Operating Room

The SMART-OR Framework for Implementing Artificial Intelligence in the Operating Room | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article The SMART-OR Framework for Implementing Artificial Intelligence in the Operating Room Hillary Lia, Divya Kewalramani, Muhammad Uzair Khalid, Justin Benton, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8338990/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Intraoperative artificial intelligence (AI) decision support systems hold promise for improving surgical outcomes, yet significant barriers impede their translation from development to clinical deployment. A systematic implementation framework informed by stakeholder perspectives is needed to guide responsible adoption in operating room (OR) settings. Methods We conducted a qualitative study using the Consolidated Framework for Implementation Research at a single North American academic institution. Phase I involved semi-structured interviews with OR personnel (surgeons, trainees, nurses, biomedical engineers) recruited through purposive maximum variation and snowball sampling until thematic saturation. Phase II comprised focus groups with patients recruited via convenience sampling. Interview and focus group transcripts underwent iterative thematic analysis using both deductive and inductive coding approaches. Results Twenty-two stakeholder interviews and two patient focus groups (n = 8) identified unique barriers and facilitators that coalesced into five major themes defining implementation requirements for intraoperative AI decision support: intuitive design, adequate training, maximizing adaptability, ongoing support, and fostering buy-in. These themes were contextualized across the surgical timeline: pre-implementation, implementation, and post-implementation phases, to create a comprehensive SMART-OR framework. Key barriers included overreliance concerns, automation bias risks, workflow disruption, team coordination challenges, and medico-legal ambiguity. Facilitators included perceived accuracy improvements, real-time guidance utility, and enhanced educational opportunities. Conclusions This study provides the first comprehensive implementation framework for intraoperative AI decision support, offering practical guidance across the technology lifecycle. The framework addresses critical gaps between AI development and clinical deployment by integrating diverse stakeholder perspectives into actionable recommendations. Future implementation efforts should prioritize transparent validation, coordinated training, and clear governance structures to ensure responsible adoption. Artificial intelligence intraoperative decision support operating room implementation science human factors adoption SMART-OR Implementation framework Figures Figure 1 Figure 2 Introduction Intraoperative adverse events cause preventable harm to surgical patients worldwide [ 1 ]. Recent advances in artificial intelligence (AI) have enabled the development of real-time decision support systems that recognize instruments, delineate anatomy, classify surgical phases, and identify unsafe zones of dissection during ongoing procedures [ 2 – 4 ]. These technologies aim to augment surgeons' perception at critical moments, with particular promise for resource-limited settings where specialist support is unavailable [ 1 , 5 ]. Despite impressive technical performance, translation into routine practice has lagged substantially. For most intraoperative AI tools, rigorous implementation studies remain sparse, and few have progressed beyond pilot testing to become evidence-based standards of care [ 6 ]. This disconnect between algorithm development and clinical deployment represents a critical barrier to realizing the patient safety benefits these technologies promise. Successful deployment requires alignment with operating room (OR) workflows, human-factors-informed interface design, role-specific training, mechanisms for performance monitoring, and generalizability across sites with varying equipment and staffing [ 7 – 10 ]. While established implementation science frameworks provide valuable theoretical guidance [ 11 – 13 ], they operate at a level of abstraction requiring substantial interpretation for the unique demands of the OR: real-time decision-making under pressure, strict sterility requirements, complex multi-professional teams, and zero tolerance for failures that could compromise patient safety. Context-specific translation of these frameworks into actionable guidance for intraoperative AI remains an unmet need. The GoNoGoNet algorithm, a validated computer vision system for identifying safe versus unsafe zones of dissection during laparoscopic cholecystectomy, exemplifies this challenge [ 14 ]. Despite demonstrated potential to reduce bile duct injuries, GoNoGoNet’s clinical impact remains unrealized without effective implementation strategies [ 15 ]. To address the critical gap between intraoperative AI development and clinical deployment, this study aims to qualitatively identify barriers, facilitators, and implementation requirements for intraoperative AI decision support. By engaging diverse stakeholders, we developed the first implementation framework and actionable checklist for deploying AI in the OR, offering concrete guidance for future implementation efforts. Methods Study Design and Theoretical Framework A two-phased qualitative study using thematic analysis methodology was conducted. Thematic analysis was adopted to allow for flexible data analysis, combining inductive and deductive approaches for integrating existing literature with results emerging from the study [16]. The study design was informed by Consolidated Framework for Implementation Research (CFIR), which provided a structured lens for exploring implementation determinants across five domains: intervention characteristics, outer setting, inner setting, characteristics of individuals, and implementation process. 11 The Replicating Effective Programs (REP) framework additionally guided our focus on practical, phase-specific implementation steps [12]. The University of Health Network Research Ethics Board approval was obtained for this study and all participants provided written informed consent (#23-517). Data collection occurred between 11/29/2023 and 08/28/2024. To ground discussions in clinical reality, we used GoNoGoNet, a validated deep learning algorithm [17,18], as a specific use case throughout all interviews and focus groups. Participants viewed video demonstrations and received lay-language explanations of the algorithm's functionality, enabling specific feedback rather than abstract speculation. Participant Recruitment and Data Collection Phase I: We employed purposive maximum variation sampling to recruit OR personnel (surgeons, trainees, nurses, biomedical engineers) at one academic institution, with diverse experience levels, and attitudes toward AI adoption, supplemented by snowball sampling. Thematic saturation was operationalized as no new themes across three consecutive interviews [19]. Semi-structured interviews were conducted via Microsoft Teams (Microsoft Corporation, Redmond, WA) using an interview guide (Supplementary Material, Section 1), structured around CFIR domains. The interviewer followed the participant lead to explore emerging topics and perspectives. Interviews were audio-recorded, de-identified, transcribed verbatim, verified for accuracy. Phase II : Patients were recruited through convenience sampling from a research volunteer database, regardless of prior surgical experience. Two focus groups with four participants each were conducted using semi-structured discussion guides (Supplementary Material, Section 2). Each 60-minute session began with standardized education about intraoperative AI, followed by moderated discussion of patient perspectives on benefits, risks, disclosure preferences, and trust considerations. Focus groups were audio recorded, de-identified, transcribed, and verified. Data Analysis Two researchers independently coded all transcripts using NVivo software (Lumivero, Denver, CO). An initial deductive coding framework based on REP implementation phases was applied to all transcripts, then expanded inductively as new concepts emerged. Discrepancies between coders were resolved through consensus between the full research team. Rigor was maintained through negative case analysis and audit trail maintenance [20]. Framework Development Identified themes were synthesized into a three-phase S ystematic M ethodology for A I R eadiness and T ranslation in the O perating R oom (SMART-OR) implementation framework with actionable checklists. The preliminary framework underwent member-checking, who confirmed relevance and suggested refinements that were incorporated to produce the final stakeholder-informed framework. Results Phase I enrolled 20 OR personnel (91% recruitment rate) including 10 surgeons, 6 nurses, 2 surgical trainees, and 2 biomedical engineers to the study. Participants represented practice settings across Canada and the United States (Table 1). Phase II enrolled six patient participants (recruitment rate: 6%), comprising three males and three females who participated in two focus groups. In Phase I, each interview lasted a mean of 29 minutes (range, 18–45 minutes). Five major themes emerged from the analysis that defined requirements for implementing intraoperative AI decision support: intuitive design, adequate training, maximizing adaptability, ongoing support, and fostering buy-in. These themes span pre-implementation, implementation, and post-implementation phases and address technical, educational, organizational, and cultural dimensions of technology adoption (Figure 1). Perceived Benefits of Intraoperative AI Decision Support Participants across all stakeholder groups identified three anticipated benefits. First, surgeons described the technology as enhancing patient safety by serving as a cognitive checkpoint during routine cases: “ the surgeon could be pretty much on autopilot for a straight-forward case and then something like this causes a warning to pop up and grab your attention somehow. It’s like a slow-down moment .” This function was viewed as particularly valuable for surgeons practicing in resource-limited settings without immediate access to expert consultation. Second, surgeons and trainees identified educational applications. The visual overlay was described as facilitating intraoperative teaching discussions: “ if the red zone (no-go) is marked out and the surgeon asks the trainee to go in the red zone or themselves approach the red zone, that becomes an automatic discussion point. ” Another trainee noted “ the capability to turn the AI model off so the learners can see what safe and unsafe zones actually looks like on normal anatomy is really helpful. ” Third, participants also noted value in post-operative video review for feedback and coaching, where the annotated anatomy could clarify dissection decisions for trainees. Perceived Risks of Intraoperative AI Decision Support Surgeons, non-medical staff, and patients described a wariness toward the technology due to the potential for overreliance, leading to poor decision making. One surgeon questioned, “ would you be falsely reassured by this technology that you're in the Go Zone despite not actually being there and then causing a bile duct injury? ” Surgeons emphasized the importance of positioning AI as an adjunct to surgical judgment rather than a replacement. One surgeon described the importance of appropriate framing: " it's not telling you how to do an operation, and it's not making a judgment for you. It's an added layer of safety. " Participants highlighted the importance of surgeons learning the algorithm's limitations and recognizing situations when clinical judgment should override AI recommendations. Participants also expressed concern about cognitive overload from introducing additional technology into the OR. One surgeon stated, " the goal is not to distract the surgeon, because they still bear the primary responsibility for the patient’s outcome. " Another surgeon noted that if the technology presents excessive clinically irrelevant information, " it's actually not helping and could be perceived as a distraction more often than not. " Additionally, concerns about medicolegal implications of AI use were also raised. Views on liability varied, with some believing AI use would provide protection in litigation and others fearing it could increase liability in the event of adverse outcomes. Theme 1: Intuitive Design Participants emphasized that interface design and ease of setup directly influence technology adoption, with well-designed systems reducing training burden and facilitating integration into existing workflows. One surgeon stated, “ if the integration is not smooth, barriers to adoption increase significantly. ” A biomedical engineer explained: “ if you can make it as easy as plug ‘A’ into ‘A’ and ‘B’ into ‘B’ and everything works, then everything will go smoothly. ” Participants suggested that interface elements such as labeling, color-coding, and information density should reinforce appropriate use of AI as a decision-making adjunct rather than a directive authority. Participants also emphasized that achieving intuitive design requires iterative usability testing in both simulation and clinical environments. One surgeon stated, “ usability studies analyzing the workflow, studying the environment and how these cases progress are essential to optimize product design and increase adoption. ” Theme 2: Adequate Training Participants emphasized that effective training must be tailored to professional roles and should communicate the purpose and value of the technology. One participant stated, " When new technologies are introduced without sufficient context, learners understandably question their relevance. It is therefore critical to demonstrate how these tools concretely advance patient care. " Surgeons described requiring conceptual understanding of algorithm function and limitations to appropriately integrate AI into clinical decision-making. One surgeon stated, “ I need to understand how the algorithm was trained, which datasets it was trained on, and how it's been validated. ” This algorithmic literacy enables surgeons to interpret predictions accurately, recognize scenarios where AI may be unreliable, and make informed decisions when AI recommendations conflict with clinical judgment. Nurses anticipated primary responsibility for equipment setup and frontline troubleshooting, necessitating practical, hands-on training. They described the need to practice setting up and troubleshooting the equipment outside of the OR to gain confidence. They also emphasized the value of quick-reference materials: " Physical tipsheets to refer back to the troubleshooting steps would be beneficial. " Similarly, biomedical engineers described needing advanced technical knowledge for system maintenance and complex problem-solving. One engineer stated, “ Troubleshooting tips for common technical failure reasons including cables or hardware components are essential. ” Inefficiencies related to inadequate training and resources may result in underutilization of otherwise useful technology. Theme 3: Maximizing Adaptability Participants emphasized that successful implementation requires adaptable technology capable of accommodating diverse user preferences and practice contexts rather than a standardized, one-size-fits-all approach. This need for adaptability manifested across three domains: use cases, display design, and physical infrastructure. Surgeons described fundamentally different anticipated workflows. Some envisioned intermittent use: " toggle on and off to ensure that you're still in the ‘Go” zone. " Others anticipated continuous monitoring with the overlay persistently displayed. Additional use cases included automatic alert-triggered displays that appear only when approaching high-risk zones, used exclusively as an educational tool for trainees rather than for clinical decision support, and split-display configurations where trainees view the overlay while supervising surgeons see the unaugmented field. One surgeon described, " the trainee doing the surgery can look at the screen with the overlay but someone teaching it could see it on the regular display and see how they are doing. " Display preferences also varied substantially. Some surgeons preferred high-confidence-only overlays: " I'd prefer to see the overlay only at 90% or higher confidence.” Others wanted graduated confidence displays with intermediate warning zones labelled as a “yellow” zone, prompting them to slow down during the dissection. Ideally, the software would be designed to accommodate multiple preferences. Nurses and biomedical engineers identified variability in OR physical space and availability of technology as a barrier. Providing a second display was an attractive option when compatibility across vendors could be an issue. However, “ adding another display introduces logistical challenges related to transport, setup, and physical space constraints. ” Ideally, the technology could be integrated with existing equipment, but it was acknowledged that this was not always possible. Theme 4: Ongoing Support Participants described the need for phased implementation support that is intensive during initial deployment. One nurse stated, " the vendor/ researcher should be present in the OR to help support arising issues as the team uses the technology for the first time. " This on-site presence allows for real-time troubleshooting of technical issues and observation of workflow challenges not anticipated during training. As implementation progresses, support can transition from continuous on-site presence to on-call availability and eventually to standard institutional IT support channels with clear escalation pathways for complex issues. Beyond initial deployment support, participants emphasized the importance of sustained responsiveness to user feedback. One surgeon stated, " be open-minded to feedback as the technology matures. Soliciting feedback and then being responsive to it helps adoption. " This ongoing engagement enables iterative refinement of both the technology and training materials based on real-world implementation challenges and evolving user needs. Theme 5: Fostering Buy-In Participants emphasized that successful implementation depends on cultivating an organizational culture that supports technology adoption and engaging key stakeholders who can champion the change. Surgeons described the need to establish a professional environment where using AI decision support is viewed as aligned with surgical excellence rather than a sign of inadequacy. One surgeon stated, “ sometimes we hesitate to ask for help because it risks appearing inadequate; a supportive culture is instead open to ideas, inquisitive, and committed to improvement. ” Participants identified trust as foundational to adoption, with surgeons requiring transparent evidence about the algorithm before integrating recommendations into clinical practice. Finally, participants emphasized selecting early adopters as champions. A surgeon warned, “ failing to choose the right first adopters (surgeons) is a major barrier to sustainable adoption. ” Champions who are involved early develop a sense of ownership and can advocate for the technology among colleagues, provide peer-to-peer mentorship, and serve as conduits for feedback between end-users and technology developers. Implementation Framework and Checklist This study integrates expert opinion from various stakeholders to explore the perceived risks and benefits of intraoperative AI decision support and to elicit the aforementioned key themes that were coalesced into the SMART-OR implementation framework with actionable checklists (Figure 2). Discussion While AI for intraoperative decision support has advanced rapidly, translation into routine clinical practice remains limited. Existing reporting and evaluation frameworks include the IDEAL-Robotics framework (Idea, Development, Exploration, Assessment, Long-term study for surgical robotics) which provides stage-gated evaluation of innovation [21]; DECIDE-AI (Developmental and Exploratory Clinical Investigations of Decision-support systems driven by AI) which standardizes early-phase evaluation [22]; CPI-AI (Clinical Practice Integration for Artificial Intelligence) which emphasizes safe clinical integration [23]; and Consolidated Standards of Reporting Trials – Artificial Intelligence extension (CONSORT-AI) which extend reporting standards for AI-driven trials [24,25]. Collectively, these frameworks clarify whether a surgical AI system works safely and effectively. However, they provide limited guidance on how to implement AI systems sustainably in complex, high-stakes surgical environments. In contrast, the broader field of digital health implementation has matured over the past two decades, producing robust, theory-driven frameworks such as the CFIR, the Non-Adoption, Abandonment, Scale-up, Spread, and Sustainability (NASSS) framework [26], and the Expert Recommendations for Implementing Change (ERIC) taxonomy [13]. However, they require extensive interpretation for intraoperative use, where decision-making occurs in seconds, sterility constraints limit physical interaction with devices, and coordination among multidisciplinary teams is both dynamic and hierarchical. The SMART-OR framework directly closes the translational gap between high-level implementation science and the realities of intraoperative AI adoption. Its novelty lies in merging AI-specific safeguards (e.g., explainability, automation-bias mitigation, post-deployment drift monitoring) with OR constraints such as limited physical space, latency, and team coordination. A comparison with existing frameworks and their limitations is presented in Table 2. Additionally, the checklist is intended to be used as a practical guide when transitioning from research to intraoperative implementation. The patterns we identified build on three decades of electronic medical record (EMR) implementation. Kaiser Permanente's successful $4 billion HealthConnect deployment demonstrated that phased rollout, physician leadership, extensive training, and workflow integration enabled large-scale health IT adoption [27]. Conversely, the United Kingdom's £10 billion National Programme for Information Technology (IT) failed due to top-down mandates, insufficient end-user engagement, one-size-fits-all design, and underestimated complexity, precisely the barriers our stakeholders identified as critical to avoid [28]. Three lessons apply directly. First, complexity across multiple domains predicts failure. Our identification of diverse user preferences signals inherent complexity requiring adaptable systems rather than standardized solutions. Second, physician engagement from inception is non-negotiable. Our emphasis on identifying surgeon champions early and framing AI as professional enhancement rather than remediation aligns with this principle [13, 29-31]. A consistent facilitator was the explicit positioning of intraoperative AI as an adjunct for surgical decision-making, reinforcing team trust. Third, training investment determines adoption. The consistent theme across stakeholder groups highlighting the importance of comprehensive training in technology uptake, mirrors findings from successful EMR implementations. These parallels suggest that surgical AI faces fundamentally similar organizational challenges to other health IT tools, with the additional complexity of real-time intraoperative use. Our findings also highlight barriers to intraoperative AI adoption. Overreliance on AI emerged as a central concern. Surgeons identified increased risk through automation bias and deskilling unless design and training explicitly teach critical appraisal of AI outputs [32]. Unlike AI in radiology, intraoperative tools must deliver actionable, non-disruptive guidance in real time, imposing stringent requirements for latency, accuracy, and interface ergonomics [33]. Implementation must also account for tightly interdependent OR teams (surgeons, anesthesiologists, nurses, technicians) necessitating coordinated, role-specific training and clear accountability for AI-related issues. Finally, medico-legal ambiguity remains a barrier; aligning policy and guidance to recognize AI as a decision-support adjunct will be crucial to responsible adoption. There are several limitations to this study. First, all participants were drawn from a single North American academic institution, which may limit the transferability of our findings to other practice settings. Future studies should include global stakeholders. Second, several sub-themes from this study require further exploration, including legislation governing intraoperative AI use, impacts on medical litigation, AI-introduced distractions, and optimal methods for usability testing. Lastly, purposive sampling may have introduced selection bias toward individuals with greater AI exposure, despite efforts to include both AI enthusiasts and skeptics. Conclusion This study describes the first stakeholder-informed SMART-OR implementation framework specifically designed to bridge the gap between development and implementation of intraoperative AI decision support. By systematically eliciting perspectives from surgeons, surgical trainees, nurses, biomedical engineers, and patients we identified five core implementation requirements and synthesized these into an actionable checklist. While validation through implementation is essential and ongoing, this formative work establishes a roadmap for researchers, clinicians, and healthcare organizations seeking to responsibly translate intraoperative AI from development to clinical practice. Declarations Ethics approval and consent to participate: This study was approved by the University of Health Network Research Ethics Board (Protocol #23-5170). All participants provided written informed consent prior to participation. The study was conducted in accordance with the Declaration of Helsinki and adhered to Good Clinical Practice guidelines. Consent for publication: Not applicable. This manuscript does not contain any individual person's data in any form (including individual details, images, or videos). All data are presented in aggregate form only. Availability of data and material: The qualitative datasets generated and analyzed during this study are not publicly available due to participant privacy and confidentiality requirements outlined in the ethics approval and informed consent documents. Aggregate data supporting the findings are available from the corresponding author upon reasonable request and subject to appropriate data sharing agreements. Competing interest: All authors declare no financial or non-financial competing interests. Funding: This work received no funding Authors' contributions: HL: Conceptualization, study design, recruitment, data collection, data analysis, framework generation, preparation of tables and figures, manuscript preparation DK: preparation of tables and figures, manuscript writing, manuscript editing MUK: Data collection, data analysis, thematic segmentation, manuscript writing, manuscript editing JB: manuscript editing CM: Conceptualization, methodology, data verification, manuscript editing RLC: manuscript editing TJL: manuscript editing MN: manuscript editing WS: manuscript editing AM: Conceptualization, supervision, framework generation, preparation of tables and figures, manuscript preparation Acknowledgements: None Clinical Trial Number: Not applicable References Kewalramani D, Loftus TJ, Coleman JR, Kaafarani H, Narayan M. Using AI to bridge global surgical gaps: high tech, high impact. The Lancet [Internet]. 2024 May [cited 2025 Dec 11];403(10438):1746–7. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0140673623027538 Laplante S, Madani A. Artificial intelligence in surgery. In: Artificial Intelligence in Clinical Practice [Internet]. Elsevier; 2024 [cited 2025 Dec 11]. p. 211–6. Available from: https://linkinghub.elsevier.com/retrieve/pii/B978044315688500019X Khalid MU, Laplante S, Madani A. Machines with vision for intraoperative guidance during gastrointestinal cancer surgery. Front Med (Lausanne). 2022;9:1025382. Hashimoto DA, Rosman G, Rus D, Meireles OR. Artificial Intelligence in Surgery: Promises and Perils. Annals of Surgery [Internet]. 2018 July [cited 2025 Dec 11];268(1):70–6. Available from: https://journals.lww.com/00000658-201807000-00013 Kewalramani D, Loftus TJ, Mayol J, Narayan M. Artificial intelligence in surgery: a global balancing act. British Journal of Surgery [Internet]. 2024 Mar 2 [cited 2025 Dec 11];111(3):znae062. Available from: https://academic.oup.com/bjs/article/doi/10.1093/bjs/znae062/7634459 Morris ZS, Wooding S, Grant J. The answer is 17 years, what is the question: understanding time lags in translational research. J R Soc Med [Internet]. 2011 Dec [cited 2025 Dec 11];104(12):510–20. Available from: https://journals.sagepub.com/doi/10.1258/jrsm.2011.110180 Finkelstein J, Gabriel A, Schmer S, Truong TT, Dunn A. Identifying Facilitators and Barriers to Implementation of AI-Assisted Clinical Decision Support in an Electronic Health Record System. J Med Syst [Internet]. 2024 Sept 18 [cited 2025 Dec 11];48(1):89. Available from: https://link.springer.com/10.1007/s10916-024-02104-9 Hassan M, Kushniruk A, Borycki E. Barriers to and Facilitators of Artificial Intelligence Adoption in Health Care: Scoping Review. JMIR Hum Factors [Internet]. 2024 Aug 29 [cited 2025 Dec 11];11:e48633. Available from: https://humanfactors.jmir.org/2024/1/e486339. Petersson L, Larsson I, Nygren JM, Nilsen P, Neher M, Reed JE, et al. Challenges to implementing artificial intelligence in healthcare: a qualitative interview study with healthcare leaders in Sweden. BMC Health Serv Res [Internet]. 2022 Dec [cited 2025 Dec 11];22(1):850. Available from: https://bmchealthservres.biomedcentral.com/articles/10.1186/s12913-022-08215-8 Nair M, Svedberg P, Larsson I, Nygren JM. A comprehensive overview of barriers and strategies for AI implementation in healthcare: Mixed-method design. Six S, editor. PLoS ONE [Internet]. 2024 Aug 9 [cited 2025 Dec 11];19(8):e0305949. Available from: https://dx.plos.org/10.1371/journal.pone.0305949 Damschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implementation Sci [Internet]. 2009 Dec [cited 2025 Dec 11];4(1):50. Available from: http://implementationscience.biomedcentral.com/articles/10.1186/1748-5908-4-50 Kilbourne AM, Neumann MS, Pincus HA, Bauer MS, Stall R. Implementing evidence-based interventions in health care: application of the replicating effective programs framework. Implementation Sci [Internet]. 2007 Dec [cited 2025 Dec 11];2(1):42. Available from: http://implementationscience.biomedcentral.com/articles/10.1186/1748-5908-2-42 Powell BJ, Waltz TJ, Chinman MJ, Damschroder LJ, Smith JL, Matthieu MM, et al. A refined compilation of implementation strategies: results from the Expert Recommendations for Implementing Change (ERIC) project. Implementation Sci [Internet]. 2015 Dec [cited 2025 Dec 11];10(1):21. Available from: http://implementationscience.biomedcentral.com/articles/10.1186/s13012-015-0209-1 Khalid MU, Laplante S, Masino C, Alseidi A, Jayaraman S, Zhang H, et al. Use of artificial intelligence for decision-support to avoid high-risk behaviors during laparoscopic cholecystectomy. Surg Endosc [Internet]. 2023 Dec [cited 2025 Dec 11];37(12):9467–75. Available from: https://link.springer.com/10.1007/s00464-023-10403-4 Peters DH, Tran NT, Adam T. Implementation research in health: a practical guide. Geneva: Alliance for Health Policy and Systems Research, World Health Organization; 2014. 67 p. Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology [Internet]. 2006 Jan [cited 2025 Dec 11];3(2):77–101. Available from: https://www.tandfonline.com/doi/full/10.1191/1478088706qp063oa Laplante S, Namazi B, Kiani P, Hashimoto DA, Alseidi A, Pasten M, et al. Validation of an artificial intelligence platform for the guidance of safe laparoscopic cholecystectomy. Surg Endosc [Internet]. 2023 Mar [cited 2025 Dec 11];37(3):2260–8. Available from: https://link.springer.com/10.1007/s00464-022-09439-9 Madani A, Namazi B, Altieri MS, Hashimoto DA, Rivera AM, Pucher PH, et al. Artificial Intelligence for Intraoperative Guidance: Using Semantic Segmentation to Identify Surgical Anatomy During Laparoscopic Cholecystectomy. Annals of Surgery [Internet]. 2022 Aug [cited 2025 Dec 11];276(2):363–9. Available from: https://journals.lww.com/10.1097/SLA.0000000000004594 Morse JM. The Significance of Saturation. Qual Health Res [Internet]. 1995 May [cited 2025 Dec 11];5(2):147–9. Available from: https://journals.sagepub.com/doi/10.1177/104973239500500201 Creswell JW. Research design: qualitative, quantitative, and mixed methods approaches. 3rd ed. Thousand Oaks, Calif: Sage Publications; 2009. 260 p. Marcus HJ, Ramirez PT, Khan DZ, Layard Horsfall H, Hanrahan JG, Williams SC, et al. The IDEAL framework for surgical robotics: development, comparative evaluation and long-term monitoring. Nat Med [Internet]. 2024 Jan [cited 2025 Dec 11];30(1):61–75. Available from: https://www.nature.com/articles/s41591-023-02732-7. Vasey B, Nagendran M, Campbell B, Clifton DA, Collins GS, Denaxas S, et al. Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI. Nat Med [Internet]. 2022 May [cited 2025 Dec 11];28(5):924–33. Available from: https://www.nature.com/articles/s41591-022-01772-9 Farrow L, Meek D, Leontidis G, Campbell M, Harrison E, Anderson L. The Clinical Practice Integration of Artificial Intelligence (CPI-AI) framework: a proposed application of IDEAL principles to artificial intelligence applications in trauma and orthopaedics. Bone Joint Res [Internet]. 2024 Sept 18 [cited 2025 Dec 11];13(9):507–12. Available from: https://boneandjoint.org.uk/doi/10.1302/2046-3758.139.BJR-2024-0135.R1 Cruz Rivera S, Liu X, Chan AW, Denniston AK, Calvert MJ, Ashrafian H, et al. Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension. The Lancet Digital Health [Internet]. 2020 Oct [cited 2025 Dec 11];2(10):e549–60. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2589750020302193 Liu X, Cruz Rivera S, Moher D, Calvert MJ, Denniston AK, Ashrafian H, et al. Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension. The Lancet Digital Health [Internet]. 2020 Oct [cited 2025 Dec 11];2(10):e537–48. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2589750020302181 Greenhalgh Trisha, Abimbola Seye. The NASSS Framework – A Synthesis of Multiple Theories of Technology Implementation. In: Studies in Health Technology and Informatics [Internet]. IOS Press; 2019 [cited 2025 Dec 11]. Available from: https://www.medra.org/servlet/aliasResolver?alias=iospressISBN&isbn=978-1-61499-990-4&spage=193&doi=10.3233/SHTI190123 Wheatley B. Transforming care delivery through health information technology. Perm J. 2013;17(1):81–6. Justinia T. The UK’s National Programme for IT: Why was it dismantled? Health Serv Manage Res [Internet]. 2017 Feb [cited 2025 Dec 11];30(1):2–9. Available from: https://journals.sagepub.com/doi/10.1177/0951484816662492 Gotlib Conn L, McKenzie M, Pearsall EA, McLeod RS. Successful implementation of an enhanced recovery after surgery programme for elective colorectal surgery: a process evaluation of champions’ experiences. Implementation Sci [Internet]. 2015 Dec [cited 2025 Dec 11];10(1):99. Available from: https://implementationscience.biomedcentral.com/articles/10.1186/s13012-015-0289-y Khalid MU, Mac A, Biderman M, Errett L, Sriharan A. Partnering to build surgical capacity in low-resource settings: a qualitative study of Canadian global surgeons. BMJ Open [Internet]. 2023 Mar [cited 2025 Dec 11];13(3):e070148. Available from: https://bmjopen.bmj.com/lookup/doi/10.1136/bmjopen-2022-070148 Latham SC. The Importance of a Physician Champion in Implementing Practice Change. American Journal of Infection Control [Internet]. 2006 June [cited 2025 Dec 11];34(5):E48. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0196655306007395 Kewalramani D, Jawa RS, Martin CA, Gosain A, Wan D, Varghese TK, et al. Position statement from the society of University surgeons, surgical education committee: Artificial intelligence in surgical training for medical students, residents, and fellows. Surgery [Internet]. 2026 Feb [cited 2025 Dec 11];190:109849. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0039606025007019 Rockall AG, Shelmerdine SC, Chen M. AI and ML in radiology: Making progress. Clinical Radiology [Internet]. 2023 Feb [cited 2025 Dec 11];78(2):81–2. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0009926022007085 Tables Table 1. Demographic and Practice Characteristics of Operating Room Personnel (Phase 1) Characteristics Value Attending Surgeons, n Female, n (%) 5 (50) Male, n (%) 5 (50) Surgical Trainees, n Female, n (%) 1 (50) Male, n (%) 1 (50) Postgraduate level, range PGY3-PGY4 Operating Room Nurses, n Female, n (%) 4 (66) Male, n (%) 2 (33) Biomedical Engineers, n Female, n (%) 2 (100) Male, n (%) 0 Practice Setting Academic medical center, n (%) 14 (70) Community hospital, n (%) 6 (30) Table 2. Translating General Implementation Science Principles to Surgical AI-Specific Guidance Comparison of implementation guidance provided by established frameworks (Consolidated Framework for Implementation Research [CFIR], Expert Recommendations for Implementing Change [ERIC], Replicating Effective Programs [REP]) versus context-specific guidance developed in this study for intraoperative artificial intelligence. Implementation Domain Generic Framework Guidance SMART-OR Framework Training Staff Conduct educational meetings (ERIC) Provide training (CFIR) Surgeons: Algorithm architecture, training data, validation results, limitations, decision frameworks when AI conflicts with judgment Nurses: Hands-on setup practice outside OR, troubleshooting decision trees, quick-reference cards for intraoperative use Engineers: System architecture, network requirements, hardware maintenance, advanced troubleshooting (XX hours) Local Adaptation Assess for readiness and identify barriers (CFIR) Allow for local adaptation (REP) Specific adaptations identified: Display modes: manual toggle, automatic alert-triggered, continuous, teaching-only, post-op review Confidence thresholds: user-customizable (90%) Physical setup: integrated vs. separate display options Vendor compatibility assessment and workarounds Ongoing Support Provide ongoing consultation (ERIC) Develop a formal implementation blueprint (REP) Phased support protocol: Weeks 1-2: On-site presence for all cases Weeks 3-4: Rapid response available within minutes Weeks 5-8: Scheduled check-ins plus as needed Months 3+: Standard IT channels with clear escalation pathway Technology Risks Address challenges to using the intervention (CFIR) AI-specific risk mitigation: Training explicitly addresses overreliance and deskilling Interface labels prompt critical thinking vs. commands Cognitive load reduction through usability testing Clarity on when to override AI recommendations Strategies for when AI and clinical judgment diverge Fostering Adoption Identify and prepare champions (ERIC) Assess for organizational culture (CFIR) OR culture-specific strategies: Frame as professional standard, not remediation Leadership visible use and endorsement Champion characteristics: clinically respected, patient with iteration Evidence package: broad training data, expert labeling, performance metrics, failure modes transparently reported Physical Integration Not addressed in general frameworks OR-specific requirements: Sterile field compatibility Physical space constraints varying by OR Cable/connection specifications Latency requirements for real-time use Point-of-use troubleshooting materials Patient Perspectives Engage patients (CFIR) Nuanced patient considerations: Develop disclosure language surgeons can customize Additional Declarations No competing interests reported. Supplementary Files 4.SupplementaryMaterial.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8338990","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":563449218,"identity":"3107447a-7a1a-4539-98a9-34a139d14173","order_by":0,"name":"Hillary Lia","email":"","orcid":"","institution":"University Health Network","correspondingAuthor":false,"prefix":"","firstName":"Hillary","middleName":"","lastName":"Lia","suffix":""},{"id":563449219,"identity":"11e805ea-11fc-4060-b08c-b4f65ad75a43","order_by":1,"name":"Divya Kewalramani","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABC0lEQVRIiWNgGAWjYDCCM4zNDz5U2ABZB1DEmfFoYW4znHEmDa5Fgggt7A3SnG2H4XzCWvjOHGwwZmA7n29w8HTyB8Y9h+vk3Xsff2CosE5swKFF8mxjw+MCntuWGw6c3SbB8OywhOGZ42YSDGfScWoxOM/YYDxD4raBGVAL0D9ALTPS2BgY2w7j1SLNY3AOpGXzB7CW+c+YPzD+w6MF6DBpnoQDIC0bJEBa5CXYGCQYG3BrkTxzEBjIB5IN7EF+STiQLrmBJ41NIuFYujEuLXxn0h8/+PjPzkByBtBhHw5Y88u3H2P+8KHGWhaXFgSQOMDAkAByKpAGMwgDfqip8oRNHwWjYBSMghEGAIOqZtNHJavWAAAAAElFTkSuQmCC","orcid":"","institution":"Rutgers Robert Wood Johnson Medical School","correspondingAuthor":true,"prefix":"","firstName":"Divya","middleName":"","lastName":"Kewalramani","suffix":""},{"id":563449220,"identity":"a755fec3-3e89-4bc1-b37d-3815a14cfb5f","order_by":2,"name":"Muhammad Uzair Khalid","email":"","orcid":"","institution":"University Health Network","correspondingAuthor":false,"prefix":"","firstName":"Muhammad","middleName":"Uzair","lastName":"Khalid","suffix":""},{"id":563449221,"identity":"2f30162d-14dd-4272-b8c2-789e99b81649","order_by":3,"name":"Justin Benton","email":"","orcid":"","institution":"Rutgers Robert Wood Johnson Medical School","correspondingAuthor":false,"prefix":"","firstName":"Justin","middleName":"","lastName":"Benton","suffix":""},{"id":563449222,"identity":"91524f84-8b39-4cbb-b826-564d80afa797","order_by":4,"name":"Caterina Masino","email":"","orcid":"","institution":"University Health Network","correspondingAuthor":false,"prefix":"","firstName":"Caterina","middleName":"","lastName":"Masino","suffix":""},{"id":563449223,"identity":"2bbf02c1-0702-45fd-9051-d36c7d1cff45","order_by":5,"name":"Rachel L. Choron","email":"","orcid":"","institution":"Rutgers Robert Wood Johnson Medical School","correspondingAuthor":false,"prefix":"","firstName":"Rachel","middleName":"L.","lastName":"Choron","suffix":""},{"id":563449224,"identity":"96eede15-cad3-4726-a2ec-0f7cde7de7df","order_by":6,"name":"Tyler J. Loftus","email":"","orcid":"","institution":"University of Florida Health","correspondingAuthor":false,"prefix":"","firstName":"Tyler","middleName":"J.","lastName":"Loftus","suffix":""},{"id":563449225,"identity":"d26292f3-4a5d-40f4-9d03-d3fee6320b7b","order_by":7,"name":"Mayur Narayan","email":"","orcid":"","institution":"Rutgers Robert Wood Johnson Medical School","correspondingAuthor":false,"prefix":"","firstName":"Mayur","middleName":"","lastName":"Narayan","suffix":""},{"id":563449226,"identity":"9f815928-1ef1-4ef2-bcce-7f7ac71534e8","order_by":8,"name":"Wagner H. Souza","email":"","orcid":"","institution":"University Health Network","correspondingAuthor":false,"prefix":"","firstName":"Wagner","middleName":"H.","lastName":"Souza","suffix":""},{"id":563449227,"identity":"25be92e6-f5a8-44cd-a802-80bfba5a4e2c","order_by":9,"name":"Amin Madani","email":"","orcid":"","institution":"University Health Network","correspondingAuthor":false,"prefix":"","firstName":"Amin","middleName":"","lastName":"Madani","suffix":""}],"badges":[],"createdAt":"2025-12-11 17:08:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8338990/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8338990/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":99308138,"identity":"66752e5e-fbbd-4d50-805a-1272bc9eb76f","added_by":"auto","created_at":"2025-12-31 16:07:49","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4774687,"visible":true,"origin":"","legend":"","description":"","filename":"1.MansucriptImplementationFramework.docx","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/0b016e495b40a47f9246cb5a.docx"},{"id":99308540,"identity":"78412814-f076-4430-ac15-14612f290f9d","added_by":"auto","created_at":"2025-12-31 16:08:44","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11413,"visible":true,"origin":"","legend":"","description":"","filename":"df8fc8fd62514db4a4a098c2701ec0aa.json","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/863a36ccae45b6c5d687e465.json"},{"id":99308166,"identity":"6817ed2d-2374-4927-b99d-46065835aa43","added_by":"auto","created_at":"2025-12-31 16:07:53","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23631,"visible":true,"origin":"","legend":"","description":"","filename":"4.SupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/5e67ca2f6ee85322370add86.docx"},{"id":98845656,"identity":"c2139c31-79be-41aa-a69f-6e2b96d0eddc","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":105110,"visible":true,"origin":"","legend":"","description":"","filename":"df8fc8fd62514db4a4a098c2701ec0aa1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/9f10482bc22701e6be6fa53d.xml"},{"id":98845658,"identity":"bfe886d2-0e5b-41f4-a1c6-68047d8a1d09","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":624276,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/964f0cd3419a559bb1c116e3.jpeg"},{"id":98845664,"identity":"9f4e9307-5e56-4f87-a222-ccdede22a07f","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":247248,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/fbdbd323cb6d9affebb38d4f.png"},{"id":98845665,"identity":"d0245f39-4b5f-461d-9c30-20ca78c793f6","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":163613,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/e6212bc66b64f0ffd27ab17b.png"},{"id":98845663,"identity":"c6abe888-d77d-4e5c-aee6-69a0369e735c","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":89383,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/18a4e72f9c48c9801e4eff40.png"},{"id":98845659,"identity":"3cebc861-a5a8-40f3-b035-838d4d028aff","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102959,"visible":true,"origin":"","legend":"","description":"","filename":"df8fc8fd62514db4a4a098c2701ec0aa1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/cfc9f54f7ac8299fd5a528c0.xml"},{"id":98845666,"identity":"f71bc9bf-2048-4c99-9118-609c09ece3f8","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"html","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":120667,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/278619fda2919a7303a0aede.html"},{"id":98845653,"identity":"f498cca7-8498-4e38-ba27-d39c007868fb","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":624276,"visible":true,"origin":"","legend":"\u003cp\u003eThematic Framework for Implementing Intraoperative Artificial Intelligence Decision Support Systems\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/3b9f9148491f9cbd288bd686.jpeg"},{"id":98845655,"identity":"0d726c69-8225-4848-bb75-9c0519c79162","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":332676,"visible":true,"origin":"","legend":"\u003cp\u003eThree-Phase Implementation Framework and Actionable Checklist for Intraoperative Artificial Intelligence\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/f4c9cacb82a9f38cc5af7f58.png"},{"id":106725186,"identity":"51f1ab88-a81f-4c09-89ca-7d05f8161e1b","added_by":"auto","created_at":"2026-04-12 18:31:44","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1844897,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/64684d5a-fd70-4615-a80f-15b57977866f.pdf"},{"id":98845654,"identity":"d993eda0-ff5a-4074-b51f-afd5d1857a5a","added_by":"auto","created_at":"2025-12-23 04:28:52","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":23631,"visible":true,"origin":"","legend":"","description":"","filename":"4.SupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-8338990/v1/fd19975284438611f4ccbb93.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"The SMART-OR Framework for Implementing Artificial Intelligence in the Operating Room","fulltext":[{"header":"Introduction","content":"\u003cp\u003eIntraoperative adverse events cause preventable harm to surgical patients worldwide [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Recent advances in artificial intelligence (AI) have enabled the development of real-time decision support systems that recognize instruments, delineate anatomy, classify surgical phases, and identify unsafe zones of dissection during ongoing procedures [\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. These technologies aim to augment surgeons' perception at critical moments, with particular promise for resource-limited settings where specialist support is unavailable [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Despite impressive technical performance, translation into routine practice has lagged substantially. For most intraoperative AI tools, rigorous implementation studies remain sparse, and few have progressed beyond pilot testing to become evidence-based standards of care [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. This disconnect between algorithm development and clinical deployment represents a critical barrier to realizing the patient safety benefits these technologies promise.\u003c/p\u003e \u003cp\u003eSuccessful deployment requires alignment with operating room (OR) workflows, human-factors-informed interface design, role-specific training, mechanisms for performance monitoring, and generalizability across sites with varying equipment and staffing [\u003cspan additionalcitationids=\"CR8 CR9\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. While established implementation science frameworks provide valuable theoretical guidance [\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], they operate at a level of abstraction requiring substantial interpretation for the unique demands of the OR: real-time decision-making under pressure, strict sterility requirements, complex multi-professional teams, and zero tolerance for failures that could compromise patient safety. Context-specific translation of these frameworks into actionable guidance for intraoperative AI remains an unmet need. The GoNoGoNet algorithm, a validated computer vision system for identifying safe versus unsafe zones of dissection during laparoscopic cholecystectomy, exemplifies this challenge [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Despite demonstrated potential to reduce bile duct injuries, GoNoGoNet\u0026rsquo;s clinical impact remains unrealized without effective implementation strategies [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eTo address the critical gap between intraoperative AI development and clinical deployment, this study aims to qualitatively identify barriers, facilitators, and implementation requirements for intraoperative AI decision support. By engaging diverse stakeholders, we developed the first implementation framework and actionable checklist for deploying AI in the OR, offering concrete guidance for future implementation efforts.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cem\u003eStudy Design and Theoretical Framework\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eA two-phased qualitative study using thematic analysis methodology was conducted. Thematic analysis was adopted to allow for flexible data analysis, combining inductive and deductive approaches for integrating existing literature with results emerging from the study [16]. The study design was informed by Consolidated Framework for Implementation Research (CFIR), which provided a structured lens for exploring implementation determinants across five domains: intervention characteristics, outer setting, inner setting, characteristics of individuals, and implementation process.\u003cstrong\u003e\u003csup\u003e11\u003c/sup\u003e\u003c/strong\u003e The Replicating Effective Programs (REP) framework additionally guided our focus on practical, phase-specific implementation steps [12].\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe University of Health Network Research Ethics Board approval was obtained for this study and all participants provided written informed consent (#23-517). Data collection occurred between 11/29/2023 and 08/28/2024. To ground discussions in clinical reality, we used GoNoGoNet, a validated deep learning algorithm [17,18], as a specific use case throughout all interviews and focus groups. Participants viewed video demonstrations and received lay-language explanations of the algorithm\u0026apos;s functionality, enabling specific feedback rather than abstract speculation.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eParticipant Recruitment and Data Collection\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cu\u003ePhase I:\u003c/u\u003e We employed purposive maximum variation sampling to recruit OR personnel (surgeons, trainees, nurses, biomedical engineers) at one academic institution, with diverse experience levels, and attitudes toward AI adoption, supplemented by snowball sampling. Thematic saturation was operationalized as no new themes across three consecutive interviews [19].\u003csup\u003e\u0026nbsp;\u003c/sup\u003eSemi-structured interviews were conducted via Microsoft Teams (Microsoft Corporation, Redmond, WA) using an interview guide (Supplementary Material, Section 1), structured around CFIR domains. The interviewer followed the participant lead to explore emerging topics and perspectives. Interviews were audio-recorded, de-identified, transcribed verbatim, verified for accuracy.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cu\u003ePhase II\u003c/u\u003e: Patients were recruited through convenience sampling from a research volunteer database, regardless of prior surgical experience. Two focus groups with four participants each were conducted using semi-structured discussion guides (Supplementary Material, Section 2). Each 60-minute session began with standardized education about intraoperative AI, followed by moderated discussion of patient perspectives on benefits, risks, disclosure preferences, and trust considerations. Focus groups were audio recorded, de-identified, transcribed, and verified.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eData Analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTwo researchers independently coded all transcripts using NVivo software (Lumivero, Denver, CO). An initial deductive coding framework based on REP implementation phases was applied to all transcripts, then expanded inductively as new concepts emerged. Discrepancies between coders were resolved through consensus between the full research team. Rigor was maintained through negative case analysis and audit trail maintenance [20].\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFramework Development\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIdentified themes were synthesized into a three-phase \u003cstrong\u003eS\u003c/strong\u003eystematic \u003cstrong\u003eM\u003c/strong\u003eethodology for \u003cstrong\u003eA\u003c/strong\u003eI \u003cstrong\u003eR\u003c/strong\u003eeadiness and \u003cstrong\u003eT\u003c/strong\u003eranslation in the \u003cstrong\u003eO\u003c/strong\u003eperating \u003cstrong\u003eR\u003c/strong\u003eoom (SMART-OR) implementation framework with actionable checklists. The preliminary framework underwent member-checking, who confirmed relevance and suggested refinements that were incorporated to produce the final stakeholder-informed framework.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003ePhase I enrolled 20 OR personnel (91% recruitment rate) including 10 surgeons, 6 nurses, 2 surgical trainees, and 2 biomedical engineers to the study. Participants represented practice settings across Canada and the United States (Table 1). Phase II enrolled six patient participants (recruitment rate: 6%), comprising three males and three females who participated in two focus groups. In Phase I, each interview lasted a mean of 29 minutes (range, 18\u0026ndash;45 minutes). Five major themes emerged from the analysis that defined requirements for implementing intraoperative AI decision support: intuitive design, adequate training, maximizing adaptability, ongoing support, and fostering buy-in. These themes span pre-implementation, implementation, and post-implementation phases and address technical, educational, organizational, and cultural dimensions of technology adoption (Figure 1).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003ePerceived Benefits of Intraoperative AI Decision Support\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParticipants across all stakeholder groups identified three anticipated benefits. First, surgeons described the technology as enhancing patient safety by serving as a cognitive checkpoint during routine cases: \u0026ldquo;\u003cem\u003ethe surgeon could be pretty much on autopilot for a straight-forward case and then something like this causes a warning to pop up and grab your attention somehow. It\u0026rsquo;s like a slow-down moment\u003c/em\u003e.\u0026rdquo; This function was viewed as particularly valuable for surgeons practicing in resource-limited settings without immediate access to expert consultation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSecond, surgeons and trainees identified educational applications. The visual overlay was described as facilitating intraoperative teaching discussions: \u0026ldquo;\u003cem\u003eif the red zone (no-go) is marked out and the surgeon asks the trainee to go in the red zone or themselves approach the red zone, that becomes an automatic discussion point.\u003c/em\u003e\u0026rdquo; Another trainee noted \u0026ldquo;\u003cem\u003ethe capability to turn the AI model off so the learners can see what safe and unsafe zones actually looks like on normal anatomy is really helpful.\u003c/em\u003e\u0026rdquo; Third, participants also noted value in post-operative video review for feedback and coaching, where the annotated anatomy could clarify dissection decisions for trainees.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003ePerceived Risks of Intraoperative AI Decision Support\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSurgeons, non-medical staff, and patients described a wariness toward the technology due to the potential for overreliance, leading to poor decision making. One surgeon questioned, \u0026ldquo;\u003cem\u003ewould you be falsely reassured by this technology that you\u0026apos;re in the Go Zone despite not actually being there and then causing a bile duct injury?\u003c/em\u003e\u0026rdquo; Surgeons emphasized the importance of positioning AI as an adjunct to surgical judgment rather than a replacement. One surgeon described the importance of appropriate framing: \u0026quot;\u003cem\u003eit\u0026apos;s not telling you how to do an operation, and it\u0026apos;s not making a judgment for you. It\u0026apos;s an added layer of safety.\u003c/em\u003e\u0026quot; Participants highlighted the importance of surgeons learning the algorithm\u0026apos;s limitations and recognizing situations when clinical judgment should override AI recommendations.\u003c/p\u003e\n\u003cp\u003eParticipants also expressed concern about cognitive overload from introducing additional technology into the OR. One surgeon stated, \u0026quot;\u003cem\u003ethe goal is not to distract the surgeon, because they still bear the primary responsibility for the patient\u0026rsquo;s outcome.\u003c/em\u003e\u0026quot; Another surgeon noted that if the technology presents excessive clinically irrelevant information, \u0026quot;\u003cem\u003eit\u0026apos;s actually not helping and could be perceived as a distraction more often than not.\u003c/em\u003e\u0026quot; Additionally, concerns about medicolegal implications of AI use were also raised. Views on liability varied, with some believing AI use would provide protection in litigation and others fearing it could increase liability in the event of adverse outcomes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTheme 1: Intuitive Design\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParticipants emphasized that interface design and ease of setup directly influence technology adoption, with well-designed systems reducing training burden and facilitating integration into existing workflows. One surgeon stated, \u0026ldquo;\u003cem\u003eif the integration is not smooth, barriers to adoption increase significantly.\u003c/em\u003e\u0026rdquo; A biomedical engineer explained: \u0026ldquo;\u003cem\u003eif you can make it as easy as plug \u0026lsquo;A\u0026rsquo; into \u0026lsquo;A\u0026rsquo; and \u0026lsquo;B\u0026rsquo; into \u0026lsquo;B\u0026rsquo; and everything works, then everything will go smoothly.\u003c/em\u003e\u0026rdquo; Participants suggested that interface elements such as labeling, color-coding, and information density should reinforce appropriate use of AI as a decision-making adjunct rather than a directive authority. Participants also emphasized that achieving intuitive design requires iterative usability testing in both simulation and clinical environments. One surgeon stated, \u0026ldquo;\u003cem\u003eusability studies analyzing the workflow, studying the environment and how these cases progress are essential to optimize product design and increase adoption.\u003c/em\u003e\u0026rdquo;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTheme 2: Adequate Training\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParticipants emphasized that effective training must be tailored to professional roles and should communicate the purpose and value of the technology. One participant stated, \u0026quot;\u003cem\u003eWhen new technologies are introduced without sufficient context, learners understandably question their relevance. It is therefore critical to demonstrate how these tools concretely advance patient care.\u003c/em\u003e\u0026quot;\u003c/p\u003e\n\u003cp\u003eSurgeons described requiring conceptual understanding of algorithm function and limitations to appropriately integrate AI into clinical decision-making. One surgeon stated, \u0026ldquo;\u003cem\u003eI need to understand how the algorithm was trained, which datasets it was trained on, and how it\u0026apos;s been validated.\u003c/em\u003e\u0026rdquo; This algorithmic literacy enables surgeons to interpret predictions accurately, recognize scenarios where AI may be unreliable, and make informed decisions when AI recommendations conflict with clinical judgment.\u003c/p\u003e\n\u003cp\u003eNurses anticipated primary responsibility for equipment setup and frontline troubleshooting, necessitating practical, hands-on training. They described the need to practice setting up and troubleshooting the equipment outside of the OR to gain confidence. They also emphasized the value of quick-reference materials: \u0026quot;\u003cem\u003ePhysical tipsheets to refer back to the troubleshooting steps would be beneficial.\u003c/em\u003e\u0026quot; Similarly, biomedical engineers described needing advanced technical knowledge for system maintenance and complex problem-solving. One engineer stated, \u0026ldquo;\u003cem\u003eTroubleshooting tips for common technical failure reasons including cables or hardware components are essential.\u003c/em\u003e\u0026rdquo; Inefficiencies related to inadequate training and resources may result in underutilization of otherwise useful technology.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTheme 3: Maximizing Adaptability\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParticipants emphasized that successful implementation requires adaptable technology capable of accommodating diverse user preferences and practice contexts rather than a standardized, one-size-fits-all approach. This need for adaptability manifested across three domains: use cases, display design, and physical infrastructure.\u003c/p\u003e\n\u003cp\u003eSurgeons described fundamentally different anticipated workflows. Some envisioned intermittent use: \u0026quot;\u003cem\u003etoggle on and off to ensure that you\u0026apos;re still in the \u0026lsquo;Go\u0026rdquo; zone.\u003c/em\u003e\u0026quot; Others anticipated continuous monitoring with the overlay persistently displayed. Additional use cases included automatic alert-triggered displays that appear only when approaching high-risk zones, used exclusively as an educational tool for trainees rather than for clinical decision support, and split-display configurations where trainees view the overlay while supervising surgeons see the unaugmented field. One surgeon described, \u0026quot;\u003cem\u003ethe trainee doing the surgery can look at the screen with the overlay but someone teaching it could see it on the regular display and see how they are doing.\u003c/em\u003e\u0026quot; Display preferences also varied substantially. Some surgeons preferred high-confidence-only overlays: \u0026quot;\u003cem\u003eI\u0026apos;d prefer to see the overlay only at 90% or higher confidence.\u0026rdquo;\u0026nbsp;\u003c/em\u003eOthers wanted graduated confidence displays with intermediate warning zones labelled as a \u0026ldquo;yellow\u0026rdquo; zone, prompting them to slow down during the dissection. Ideally, the software would be designed to accommodate multiple preferences.\u003c/p\u003e\n\u003cp\u003eNurses and biomedical engineers identified variability in OR physical space and availability of technology as a barrier. Providing a second display was an attractive option when compatibility across vendors could be an issue. However, \u0026ldquo;\u003cem\u003eadding another display introduces logistical challenges related to transport, setup, and physical space constraints.\u003c/em\u003e\u0026rdquo; Ideally, the technology could be integrated with existing equipment, but it was acknowledged that this was not always possible.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTheme 4: Ongoing Support\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParticipants described the need for phased implementation support that is intensive during initial deployment. One nurse stated, \u0026quot;\u003cem\u003ethe vendor/ researcher should be present in the OR to help support arising issues as the team uses the technology for the first time.\u003c/em\u003e\u0026quot; This on-site presence allows for real-time troubleshooting of technical issues and observation of workflow challenges not anticipated during training. As implementation progresses, support can transition from continuous on-site presence to on-call availability and eventually to standard institutional IT support channels with clear escalation pathways for complex issues. Beyond initial deployment support, participants emphasized the importance of sustained responsiveness to user feedback. One surgeon stated, \u0026quot;\u003cem\u003ebe open-minded to feedback as the technology matures. Soliciting feedback and then being responsive to it helps adoption.\u003c/em\u003e\u0026quot; This ongoing engagement enables iterative refinement of both the technology and training materials based on real-world implementation challenges and evolving user needs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTheme 5: Fostering Buy-In\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParticipants emphasized that successful implementation depends on cultivating an organizational culture that supports technology adoption and engaging key stakeholders who can champion the change. Surgeons described the need to establish a professional environment where using AI decision support is viewed as aligned with surgical excellence rather than a sign of inadequacy. One surgeon stated, \u0026ldquo;\u003cem\u003esometimes we hesitate to ask for help because it risks appearing inadequate; a supportive culture is instead open to ideas, inquisitive, and committed to improvement.\u003c/em\u003e\u0026rdquo; Participants identified trust as foundational to adoption, with surgeons requiring transparent evidence about the algorithm before integrating recommendations into clinical practice.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFinally, participants emphasized selecting early adopters as champions. A surgeon warned, \u0026ldquo;\u003cem\u003efailing to choose the right first adopters (surgeons) is a major barrier to sustainable adoption.\u003c/em\u003e\u0026rdquo; Champions who are involved early develop a sense of ownership and can advocate for the technology among colleagues, provide peer-to-peer mentorship, and serve as conduits for feedback between end-users and technology developers.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eImplementation Framework and Checklist\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study integrates expert opinion from various stakeholders to explore the perceived risks and benefits of intraoperative AI decision support and to elicit the aforementioned key themes that were coalesced into the SMART-OR implementation framework with actionable checklists (Figure 2).\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eWhile AI for intraoperative decision support has advanced rapidly, translation into routine clinical practice remains limited. Existing reporting and evaluation frameworks include the IDEAL-Robotics framework (Idea, Development, Exploration, Assessment, Long-term study for surgical robotics) which provides stage-gated evaluation of innovation [21]; DECIDE-AI (Developmental and Exploratory Clinical Investigations of Decision-support systems driven by AI) which standardizes early-phase evaluation [22]; CPI-AI (Clinical Practice Integration for Artificial Intelligence) which emphasizes safe clinical integration [23]; and Consolidated Standards of Reporting Trials \u0026ndash; Artificial Intelligence extension (CONSORT-AI) which extend reporting standards for AI-driven trials [24,25].\u0026nbsp;Collectively, these frameworks clarify whether a surgical AI system works safely and effectively. However, they provide limited guidance on how to implement AI systems sustainably in complex, high-stakes surgical environments.\u003c/p\u003e\n\u003cp\u003eIn contrast, the broader field of digital health implementation has matured over the past two decades, producing robust, theory-driven frameworks such as the CFIR, the Non-Adoption, Abandonment, Scale-up, Spread, and Sustainability (NASSS) framework [26], and the Expert Recommendations for Implementing Change (ERIC) taxonomy [13]. However, they require extensive interpretation for intraoperative use, where decision-making occurs in seconds, sterility constraints limit physical interaction with devices, and coordination among multidisciplinary teams is both dynamic and hierarchical. The SMART-OR framework directly closes the translational gap between high-level implementation science and the realities of intraoperative AI adoption. Its novelty lies in merging AI-specific safeguards (e.g., explainability, automation-bias mitigation, post-deployment drift monitoring) with OR constraints such as limited physical space, latency, and team coordination. A comparison with existing frameworks and their limitations is presented in Table 2. Additionally, the checklist is intended to be used as a practical guide when transitioning from research to intraoperative implementation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe patterns we identified build on three decades of electronic medical record (EMR) implementation. Kaiser Permanente\u0026apos;s successful $4 billion HealthConnect deployment demonstrated that phased rollout, physician leadership, extensive training, and workflow integration enabled large-scale health IT adoption [27]. Conversely, the United Kingdom\u0026apos;s \u0026pound;10 billion National Programme for Information Technology (IT) failed due to top-down mandates, insufficient end-user engagement, one-size-fits-all design, and underestimated complexity, precisely the barriers our stakeholders identified as critical to avoid [28].\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eThree lessons apply directly. First, complexity across multiple domains predicts failure. Our identification of diverse user preferences signals inherent complexity requiring \u003cem\u003e\u003cu\u003eadaptable systems\u003c/u\u003e\u003c/em\u003e rather than standardized solutions. Second, physician engagement from inception is non-negotiable. Our emphasis on identifying \u003cem\u003e\u003cu\u003esurgeon champions\u003c/u\u003e\u003c/em\u003e early and framing AI as professional enhancement rather than remediation aligns with this principle [13, 29-31].\u0026nbsp;A consistent facilitator was the explicit positioning of intraoperative AI as an adjunct for surgical decision-making, reinforcing team trust. Third, \u003cem\u003e\u003cu\u003etraining investment\u003c/u\u003e\u003c/em\u003e determines adoption. The consistent theme across stakeholder groups highlighting the importance of comprehensive training in technology uptake, mirrors findings from successful EMR implementations. These parallels suggest that surgical AI faces fundamentally similar organizational challenges to other health IT tools, with the additional complexity of real-time intraoperative use.\u003c/p\u003e\n\u003cp\u003eOur findings also highlight barriers to intraoperative AI adoption. Overreliance on AI emerged as a central concern. Surgeons identified increased risk through automation bias and deskilling unless design and training explicitly teach critical appraisal of AI outputs [32]. Unlike AI in radiology, intraoperative tools must deliver actionable, non-disruptive guidance in real time, imposing stringent requirements for latency, accuracy, and interface ergonomics [33].\u003cstrong\u003e\u003csup\u003e\u0026nbsp;\u003c/sup\u003e\u003c/strong\u003eImplementation must also account for tightly interdependent OR teams (surgeons, anesthesiologists, nurses, technicians) necessitating coordinated, role-specific training and clear accountability for AI-related issues. Finally, medico-legal ambiguity remains a barrier; aligning policy and guidance to recognize AI as a decision-support adjunct will be crucial to responsible adoption.\u003c/p\u003e\n\u003cp\u003eThere are several limitations to this study. First, all participants were drawn from a single North American academic institution, which may limit the transferability of our findings to other practice settings. Future studies should include global stakeholders. Second, several sub-themes from this study require further exploration, including legislation governing intraoperative AI use, impacts on medical litigation, AI-introduced distractions, and optimal methods for usability testing. Lastly, purposive sampling may have introduced selection bias toward individuals with greater AI exposure, despite efforts to include both AI enthusiasts and skeptics.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study describes the first stakeholder-informed SMART-OR implementation framework specifically designed to bridge the gap between development and implementation of intraoperative AI decision support. By systematically eliciting perspectives from surgeons, surgical trainees, nurses, biomedical engineers, and patients we identified five core implementation requirements and synthesized these into an actionable checklist. While validation through implementation is essential and ongoing, this formative work establishes a roadmap for researchers, clinicians, and healthcare organizations seeking to responsibly translate intraoperative AI from development to clinical practice.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u0026nbsp;\u003c/strong\u003eThis study was approved by the University of Health Network Research Ethics Board (Protocol #23-5170). All participants provided written informed consent prior to participation. The study was conducted in accordance with the Declaration of Helsinki and adhered to Good Clinical Practice guidelines.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u0026nbsp;\u003c/strong\u003eNot applicable. This manuscript does not contain any individual person\u0026apos;s data in any form (including individual details, images, or videos). All data are presented in aggregate form only.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material:\u0026nbsp;\u003c/strong\u003eThe qualitative datasets generated and analyzed during this study are not publicly available due to participant privacy and confidentiality requirements outlined in the ethics approval and informed consent documents. Aggregate data supporting the findings are available from the corresponding author upon reasonable request and subject to appropriate data sharing agreements.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interest:\u0026nbsp;\u003c/strong\u003eAll authors declare no financial or non-financial competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eThis work received no funding\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHL: Conceptualization, study design, recruitment, data collection, data analysis, framework generation, preparation of tables and figures, manuscript preparation\u003c/p\u003e\n\u003cp\u003eDK: preparation of tables and figures, manuscript writing, manuscript editing\u003c/p\u003e\n\u003cp\u003eMUK: Data collection, data analysis, thematic segmentation, manuscript writing, manuscript editing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eJB: manuscript editing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCM: Conceptualization, methodology, data verification, manuscript editing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRLC: manuscript editing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTJL: manuscript editing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMN: manuscript editing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWS: manuscript editing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAM: Conceptualization, supervision, framework generation, preparation of tables and figures, manuscript preparation\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u0026nbsp;\u003c/strong\u003eNone\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Trial Number:\u0026nbsp;\u003c/strong\u003eNot applicable\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eKewalramani D, Loftus TJ, Coleman JR, Kaafarani H, Narayan M. Using AI to bridge global surgical gaps: high tech, high impact. The Lancet [Internet]. 2024 May [cited 2025 Dec 11];403(10438):1746\u0026ndash;7. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0140673623027538\u003c/li\u003e\n\u003cli\u003eLaplante S, Madani A. Artificial intelligence in surgery. In: Artificial Intelligence in Clinical Practice [Internet]. Elsevier; 2024 [cited 2025 Dec 11]. p. 211\u0026ndash;6. Available from: https://linkinghub.elsevier.com/retrieve/pii/B978044315688500019X\u003c/li\u003e\n\u003cli\u003eKhalid MU, Laplante S, Madani A. Machines with vision for intraoperative guidance during gastrointestinal cancer surgery. Front Med (Lausanne). 2022;9:1025382.\u003c/li\u003e\n\u003cli\u003eHashimoto DA, Rosman G, Rus D, Meireles OR. Artificial Intelligence in Surgery: Promises and Perils. Annals of Surgery [Internet]. 2018 July [cited 2025 Dec 11];268(1):70\u0026ndash;6. Available from: https://journals.lww.com/00000658-201807000-00013\u003c/li\u003e\n\u003cli\u003eKewalramani D, Loftus TJ, Mayol J, Narayan M. Artificial intelligence in surgery: a global balancing act. British Journal of Surgery [Internet]. 2024 Mar 2 [cited 2025 Dec 11];111(3):znae062. Available from: https://academic.oup.com/bjs/article/doi/10.1093/bjs/znae062/7634459\u003c/li\u003e\n\u003cli\u003eMorris ZS, Wooding S, Grant J. The answer is 17 years, what is the question: understanding time lags in translational research. J R Soc Med [Internet]. 2011 Dec [cited 2025 Dec 11];104(12):510\u0026ndash;20. Available from: https://journals.sagepub.com/doi/10.1258/jrsm.2011.110180\u003c/li\u003e\n\u003cli\u003eFinkelstein J, Gabriel A, Schmer S, Truong TT, Dunn A. Identifying Facilitators and Barriers to Implementation of AI-Assisted Clinical Decision Support in an Electronic Health Record System. J Med Syst [Internet]. 2024 Sept 18 [cited 2025 Dec 11];48(1):89. Available from: https://link.springer.com/10.1007/s10916-024-02104-9\u003c/li\u003e\n\u003cli\u003eHassan M, Kushniruk A, Borycki E. Barriers to and Facilitators of Artificial Intelligence Adoption in Health Care: Scoping Review. JMIR Hum Factors [Internet]. 2024 Aug 29 [cited 2025 Dec 11];11:e48633. Available from: https://humanfactors.jmir.org/2024/1/e486339. \u003c/li\u003e\n\u003cli\u003ePetersson L, Larsson I, Nygren JM, Nilsen P, Neher M, Reed JE, et al. Challenges to implementing artificial intelligence in healthcare: a qualitative interview study with healthcare leaders in Sweden. BMC Health Serv Res [Internet]. 2022 Dec [cited 2025 Dec 11];22(1):850. Available from: https://bmchealthservres.biomedcentral.com/articles/10.1186/s12913-022-08215-8\u003c/li\u003e\n\u003cli\u003eNair M, Svedberg P, Larsson I, Nygren JM. A comprehensive overview of barriers and strategies for AI implementation in healthcare: Mixed-method design. Six S, editor. PLoS ONE [Internet]. 2024 Aug 9 [cited 2025 Dec 11];19(8):e0305949. Available from: https://dx.plos.org/10.1371/journal.pone.0305949\u003c/li\u003e\n\u003cli\u003eDamschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implementation Sci [Internet]. 2009 Dec [cited 2025 Dec 11];4(1):50. Available from: http://implementationscience.biomedcentral.com/articles/10.1186/1748-5908-4-50\u003c/li\u003e\n\u003cli\u003eKilbourne AM, Neumann MS, Pincus HA, Bauer MS, Stall R. Implementing evidence-based interventions in health care: application of the replicating effective programs framework. Implementation Sci [Internet]. 2007 Dec [cited 2025 Dec 11];2(1):42. Available from: http://implementationscience.biomedcentral.com/articles/10.1186/1748-5908-2-42\u003c/li\u003e\n\u003cli\u003ePowell BJ, Waltz TJ, Chinman MJ, Damschroder LJ, Smith JL, Matthieu MM, et al. A refined compilation of implementation strategies: results from the Expert Recommendations for Implementing Change (ERIC) project. Implementation Sci [Internet]. 2015 Dec [cited 2025 Dec 11];10(1):21. Available from: http://implementationscience.biomedcentral.com/articles/10.1186/s13012-015-0209-1\u003c/li\u003e\n\u003cli\u003eKhalid MU, Laplante S, Masino C, Alseidi A, Jayaraman S, Zhang H, et al. Use of artificial intelligence for decision-support to avoid high-risk behaviors during laparoscopic cholecystectomy. Surg Endosc [Internet]. 2023 Dec [cited 2025 Dec 11];37(12):9467\u0026ndash;75. Available from: https://link.springer.com/10.1007/s00464-023-10403-4\u003c/li\u003e\n\u003cli\u003ePeters DH, Tran NT, Adam T. Implementation research in health: a practical guide. Geneva: Alliance for Health Policy and Systems Research, World Health Organization; 2014. 67 p.\u003c/li\u003e\n\u003cli\u003eBraun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology [Internet]. 2006 Jan [cited 2025 Dec 11];3(2):77\u0026ndash;101. Available from: https://www.tandfonline.com/doi/full/10.1191/1478088706qp063oa\u003c/li\u003e\n\u003cli\u003eLaplante S, Namazi B, Kiani P, Hashimoto DA, Alseidi A, Pasten M, et al. Validation of an artificial intelligence platform for the guidance of safe laparoscopic cholecystectomy. Surg Endosc [Internet]. 2023 Mar [cited 2025 Dec 11];37(3):2260\u0026ndash;8. Available from: https://link.springer.com/10.1007/s00464-022-09439-9\u003c/li\u003e\n\u003cli\u003eMadani A, Namazi B, Altieri MS, Hashimoto DA, Rivera AM, Pucher PH, et al. Artificial Intelligence for Intraoperative Guidance: Using Semantic Segmentation to Identify Surgical Anatomy During Laparoscopic Cholecystectomy. Annals of Surgery [Internet]. 2022 Aug [cited 2025 Dec 11];276(2):363\u0026ndash;9. Available from: https://journals.lww.com/10.1097/SLA.0000000000004594\u003c/li\u003e\n\u003cli\u003eMorse JM. The Significance of Saturation. Qual Health Res [Internet]. 1995 May [cited 2025 Dec 11];5(2):147\u0026ndash;9. Available from: https://journals.sagepub.com/doi/10.1177/104973239500500201\u003c/li\u003e\n\u003cli\u003eCreswell JW. Research design: qualitative, quantitative, and mixed methods approaches. 3rd ed. Thousand Oaks, Calif: Sage Publications; 2009. 260 p.\u003c/li\u003e\n\u003cli\u003eMarcus HJ, Ramirez PT, Khan DZ, Layard Horsfall H, Hanrahan JG, Williams SC, et al. The IDEAL framework for surgical robotics: development, comparative evaluation and long-term monitoring. Nat Med [Internet]. 2024 Jan [cited 2025 Dec 11];30(1):61\u0026ndash;75. Available from: https://www.nature.com/articles/s41591-023-02732-7. \u003c/li\u003e\n\u003cli\u003eVasey B, Nagendran M, Campbell B, Clifton DA, Collins GS, Denaxas S, et al. Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI. Nat Med [Internet]. 2022 May [cited 2025 Dec 11];28(5):924\u0026ndash;33. Available from: https://www.nature.com/articles/s41591-022-01772-9\u003c/li\u003e\n\u003cli\u003eFarrow L, Meek D, Leontidis G, Campbell M, Harrison E, Anderson L. The Clinical Practice Integration of Artificial Intelligence (CPI-AI) framework: a proposed application of IDEAL principles to artificial intelligence applications in trauma and orthopaedics. Bone Joint Res [Internet]. 2024 Sept 18 [cited 2025 Dec 11];13(9):507\u0026ndash;12. Available from: https://boneandjoint.org.uk/doi/10.1302/2046-3758.139.BJR-2024-0135.R1\u003c/li\u003e\n\u003cli\u003eCruz Rivera S, Liu X, Chan AW, Denniston AK, Calvert MJ, Ashrafian H, et al. Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension. The Lancet Digital Health [Internet]. 2020 Oct [cited 2025 Dec 11];2(10):e549\u0026ndash;60. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2589750020302193\u003c/li\u003e\n\u003cli\u003eLiu X, Cruz Rivera S, Moher D, Calvert MJ, Denniston AK, Ashrafian H, et al. Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension. The Lancet Digital Health [Internet]. 2020 Oct [cited 2025 Dec 11];2(10):e537\u0026ndash;48. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2589750020302181\u003c/li\u003e\n\u003cli\u003eGreenhalgh Trisha, Abimbola Seye. The NASSS Framework \u0026ndash; A Synthesis of Multiple Theories of Technology Implementation. In: Studies in Health Technology and Informatics [Internet]. IOS Press; 2019 [cited 2025 Dec 11]. Available from: https://www.medra.org/servlet/aliasResolver?alias=iospressISBN\u0026amp;isbn=978-1-61499-990-4\u0026amp;spage=193\u0026amp;doi=10.3233/SHTI190123\u003c/li\u003e\n\u003cli\u003eWheatley B. Transforming care delivery through health information technology. Perm J. 2013;17(1):81\u0026ndash;6.\u003c/li\u003e\n\u003cli\u003eJustinia T. The UK\u0026rsquo;s National Programme for IT: Why was it dismantled? Health Serv Manage Res [Internet]. 2017 Feb [cited 2025 Dec 11];30(1):2\u0026ndash;9. Available from: https://journals.sagepub.com/doi/10.1177/0951484816662492\u003c/li\u003e\n\u003cli\u003eGotlib Conn L, McKenzie M, Pearsall EA, McLeod RS. Successful implementation of an enhanced recovery after surgery programme for elective colorectal surgery: a process evaluation of champions\u0026rsquo; experiences. Implementation Sci [Internet]. 2015 Dec [cited 2025 Dec 11];10(1):99. Available from: https://implementationscience.biomedcentral.com/articles/10.1186/s13012-015-0289-y\u003c/li\u003e\n\u003cli\u003eKhalid MU, Mac A, Biderman M, Errett L, Sriharan A. Partnering to build surgical capacity in low-resource settings: a qualitative study of Canadian global surgeons. BMJ Open [Internet]. 2023 Mar [cited 2025 Dec 11];13(3):e070148. Available from: https://bmjopen.bmj.com/lookup/doi/10.1136/bmjopen-2022-070148\u003c/li\u003e\n\u003cli\u003eLatham SC. The Importance of a Physician Champion in Implementing Practice Change. American Journal of Infection Control [Internet]. 2006 June [cited 2025 Dec 11];34(5):E48. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0196655306007395\u003c/li\u003e\n\u003cli\u003eKewalramani D, Jawa RS, Martin CA, Gosain A, Wan D, Varghese TK, et al. Position statement from the society of University surgeons, surgical education committee: Artificial intelligence in surgical training for medical students, residents, and fellows. Surgery [Internet]. 2026 Feb [cited 2025 Dec 11];190:109849. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0039606025007019\u003c/li\u003e\n\u003cli\u003eRockall AG, Shelmerdine SC, Chen M. AI and ML in radiology: Making progress. Clinical Radiology [Internet]. 2023 Feb [cited 2025 Dec 11];78(2):81\u0026ndash;2. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0009926022007085\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1. Demographic and Practice Characteristics of Operating Room Personnel (Phase 1)\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCharacteristics\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eValue\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAttending Surgeons, n\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Female, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e5 (50)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Male, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e5 (50)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSurgical Trainees, n\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Female, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e1 (50)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Male, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e1 (50)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Postgraduate level, range\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003ePGY3-PGY4\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOperating Room Nurses, n\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Female, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e4 (66)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Male, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e2 (33)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBiomedical Engineers, n\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Female, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e2 (100)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Male, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePractice Setting\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Academic medical center, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e14 (70)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 409px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Community hospital, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 215px;\"\u003e\n \u003cp\u003e6 (30)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Translating General Implementation Science Principles to Surgical AI-Specific Guidance\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eComparison of implementation guidance provided by established frameworks (Consolidated Framework for Implementation Research [CFIR], Expert Recommendations for Implementing Change [ERIC], Replicating Effective Programs [REP]) versus context-specific guidance developed in this study for intraoperative artificial intelligence.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 125px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eImplementation Domain\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 182px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGeneric Framework Guidance\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 317px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSMART-OR Framework\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 125px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTraining Staff\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 182px;\"\u003e\n \u003cp\u003eConduct educational meetings (ERIC)\u003c/p\u003e\n \u003cp\u003eProvide training (CFIR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 317px;\"\u003e\n \u003cul\u003e\n \u003cli\u003e\u003cstrong\u003eSurgeons:\u003c/strong\u003e Algorithm architecture, training data, validation results, limitations, decision frameworks when AI conflicts with judgment\u0026nbsp;\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eNurses:\u003c/strong\u003e Hands-on setup practice outside OR, troubleshooting decision trees, quick-reference cards for intraoperative use\u0026nbsp;\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eEngineers:\u003c/strong\u003e System architecture, network requirements, hardware maintenance, advanced troubleshooting (XX hours)\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 125px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLocal Adaptation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 182px;\"\u003e\n \u003cp\u003eAssess for readiness and identify barriers (CFIR)\u003c/p\u003e\n \u003cp\u003eAllow for local adaptation (REP)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 317px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecific adaptations identified:\u003c/strong\u003e\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003eDisplay modes: manual toggle, automatic alert-triggered, continuous, teaching-only, post-op review\u003c/li\u003e\n \u003cli\u003eConfidence thresholds: user-customizable (90%)\u003c/li\u003e\n \u003cli\u003ePhysical setup: integrated vs. separate display options\u003c/li\u003e\n \u003cli\u003eVendor compatibility assessment and workarounds\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 125px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOngoing Support\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 182px;\"\u003e\n \u003cp\u003eProvide ongoing consultation (ERIC)\u003c/p\u003e\n \u003cp\u003eDevelop a formal implementation blueprint (REP)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 317px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePhased support protocol:\u003c/strong\u003e\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003eWeeks 1-2: On-site presence for all cases\u003c/li\u003e\n \u003cli\u003eWeeks 3-4: Rapid response available within minutes\u003c/li\u003e\n \u003cli\u003eWeeks 5-8: Scheduled check-ins plus as needed\u003c/li\u003e\n \u003cli\u003eMonths 3+: Standard IT channels with clear escalation pathway\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 125px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTechnology Risks\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 182px;\"\u003e\n \u003cp\u003eAddress challenges to using the intervention (CFIR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 317px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAI-specific risk mitigation:\u003c/strong\u003e\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003eTraining explicitly addresses overreliance and deskilling\u003c/li\u003e\n \u003cli\u003eInterface labels prompt critical thinking vs. commands\u003c/li\u003e\n \u003cli\u003eCognitive load reduction through usability testing\u003c/li\u003e\n \u003cli\u003eClarity on when to override AI recommendations\u003c/li\u003e\n \u003cli\u003eStrategies for when AI and clinical judgment diverge\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 125px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFostering Adoption\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 182px;\"\u003e\n \u003cp\u003eIdentify and prepare champions (ERIC)\u003c/p\u003e\n \u003cp\u003eAssess for organizational culture (CFIR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 317px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOR culture-specific strategies:\u003c/strong\u003e\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003eFrame as professional standard, not remediation\u003c/li\u003e\n \u003cli\u003eLeadership visible use and endorsement\u003c/li\u003e\n \u003cli\u003eChampion characteristics: clinically respected, patient with iteration\u003c/li\u003e\n \u003cli\u003eEvidence package: broad training data, expert labeling, performance metrics, failure modes transparently reported\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 125px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePhysical Integration\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 182px;\"\u003e\n \u003cp\u003eNot addressed in general frameworks\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 317px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOR-specific requirements:\u003c/strong\u003e\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003eSterile field compatibility\u003c/li\u003e\n \u003cli\u003ePhysical space constraints varying by OR\u003c/li\u003e\n \u003cli\u003eCable/connection specifications\u003c/li\u003e\n \u003cli\u003eLatency requirements for real-time use\u003c/li\u003e\n \u003cli\u003ePoint-of-use troubleshooting materials\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 125px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePatient Perspectives\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 182px;\"\u003e\n \u003cp\u003eEngage patients (CFIR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 317px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNuanced patient considerations:\u003c/strong\u003e\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003eDevelop disclosure language surgeons can customize\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Artificial intelligence, intraoperative decision support, operating room, implementation science, human factors, adoption, SMART-OR Implementation framework","lastPublishedDoi":"10.21203/rs.3.rs-8338990/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8338990/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eIntraoperative artificial intelligence (AI) decision support systems hold promise for improving surgical outcomes, yet significant barriers impede their translation from development to clinical deployment. A systematic implementation framework informed by stakeholder perspectives is needed to guide responsible adoption in operating room (OR) settings.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe conducted a qualitative study using the Consolidated Framework for Implementation Research at a single North American academic institution. Phase I involved semi-structured interviews with OR personnel (surgeons, trainees, nurses, biomedical engineers) recruited through purposive maximum variation and snowball sampling until thematic saturation. Phase II comprised focus groups with patients recruited via convenience sampling. Interview and focus group transcripts underwent iterative thematic analysis using both deductive and inductive coding approaches.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eTwenty-two stakeholder interviews and two patient focus groups (n\u0026thinsp;=\u0026thinsp;8) identified unique barriers and facilitators that coalesced into five major themes defining implementation requirements for intraoperative AI decision support: intuitive design, adequate training, maximizing adaptability, ongoing support, and fostering buy-in. These themes were contextualized across the surgical timeline: pre-implementation, implementation, and post-implementation phases, to create a comprehensive SMART-OR framework. Key barriers included overreliance concerns, automation bias risks, workflow disruption, team coordination challenges, and medico-legal ambiguity. Facilitators included perceived accuracy improvements, real-time guidance utility, and enhanced educational opportunities.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThis study provides the first comprehensive implementation framework for intraoperative AI decision support, offering practical guidance across the technology lifecycle. The framework addresses critical gaps between AI development and clinical deployment by integrating diverse stakeholder perspectives into actionable recommendations. Future implementation efforts should prioritize transparent validation, coordinated training, and clear governance structures to ensure responsible adoption.\u003c/p\u003e","manuscriptTitle":"The SMART-OR Framework for Implementing Artificial Intelligence in the Operating Room","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-23 04:28:47","doi":"10.21203/rs.3.rs-8338990/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"bbe84572-2f04-4a9b-b2be-6d3264bc4f22","owner":[],"postedDate":"December 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-09T23:23:55+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-23 04:28:47","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8338990","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8338990","identity":"rs-8338990","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}