{"paper_id":"2e8e1ff6-a6f5-490f-8f9c-ea0b7bf7fbba","body_text":"Stakeholder Perspectives of Implementation Barriers of Artificial Intelligence in Eye Care: A qualitative framework-based study | 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 Stakeholder Perspectives of Implementation Barriers of Artificial Intelligence in Eye Care: A qualitative framework-based study Judy Nam, Angelica Ly, Sarita Herse, Chris Lim, Mary-Anne Williams, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7911455/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Purpose Despite the revolution of artificial intelligence (AI), its integration remains limited in healthcare. A comprehensive understanding of the barriers to implementation is crucial to enhance the utilisation of AI. This study applies a conceptual framework-based analysis, to explore stakeholder perspectives of implementation barriers of AI in digital diagnosis in eye care. Methods Purposive sampling was used to identify key individuals across stakeholder groups, including technology developers, clinicians, patients, and healthcare leaders. Semi-structured interviews were conducted with 37 stakeholders. Using the Updated Consolidated Framework for Implementation Research (CFIR), responses to the question: ‘What is the biggest barrier to digital diagnosis or AI, specifically age-related macular degeneration (AMD) in Australia?’ were analysed. Barriers identified by stakeholders were mapped to thematic constructs of Updated CFIR and the relative importance of each implementation barrier was measured. Results For clinicians and developers, ‘innovation’ domain was the most frequently cited. Clinicians were most concerned of the costs involved; whereas for developers the lack of evidence of the innovation in real world applications was the main challenge. For leaders and patients, ‘individuals’ domain was the most frequently cited. Leaders were focused on the innovation deliverers: expressing the potential risk of over-reliance on the innovation, and the subsequent consequence of clinician deskilling. Patients were more concerned about innovation recipients: emphasising the perceived lack of human empathy with the implementation of AI. Conclusions Differences were revealed in the identified barriers to the implementation of AI across stakeholder groups. A co-design approach to address the misalignment in key barriers may be essential to successful implementation of AI in digital health innovations. artificial intelligence barriers implementation stakeholders Figures Figure 1 KEY POINTS Applications of artificial intelligence hold the potential to transform eye care, but its integration in current practice remains limited. Stakeholder analysis reveals the perceived barriers to implementation of artificial intelligence in digital diagnostics across groups of developers, clinicians, patients, and healthcare leaders in eye care. While there are some shared barriers, a fundamental misalignment in perspectives highlights the importance of co-design in guiding successful artificial intelligence implementation into the future. INTRODUCTION Artificial intelligence (AI) is transforming everyday lives with its ability to simulate human intelligent behaviour ( 1 ) across a wide range of applications ( 2 ). It is becoming an increasingly popular tool utilised across various fields, including healthcare ( 3 ), where it can improve diagnostic accuracy and quality of care ( 4 ). As AI can enhance clinical decision-making ( 5 , 6 ), there has been a growing interest in augmenting its use for digital diagnosis and improving patient outcomes ( 4 ) across disciplines. In eye care, AI can aid in preventing avoidable blindness and vision impairment. One key example of its potential to revolutionise current clinical practice is in diagnosing and managing age-related macular degeneration (AMD). AMD is a leading cause of vision loss with a global prevalence of 8.7 per cent ( 7 ), and is a growing public health concern as it is frequently underdiagnosed and misdiagnosed in several healthcare settings( 8 , 9 ). This leads to delayed management and vision-threatening consequences, as the progression of AMD is characterised by central vision loss. Furthermore, AMD has two presentations known as ‘dry’ (non-neovascular) or ‘wet’ (neovascular) AMD, which is important to differentiate to provide appropriate management for the patient. AI can be applied to any stage of the patient journey, from improving early detection and diagnosis, to management and triaging for treatment ( 10 , 11 ). However, it is evident from the literature, that despite the significant potential of AI to transform eyecare, its integration remains limited across healthcare ( 12 ). Although there are known challenges to the general implementation of AI including cost, liability and data privacy ( 13 – 15 ), the real-world barriers to its application in digital diagnostic tools in eye care, specifically AMD, have not been explored. This study used qualitative interviews with key stakeholders to explore the barriers to adopting AI in macular disease and the factors contributing to its lack of clinical utilisation in diagnosis. Recognising a co-design approach is fundamental to ensuring both the success and broad uptake of these technologies, clinicians, patients, healthcare leaders and technology developers were surveyed to capture their diverse perspectives. For analysis, the Updated CFIR ( 16 ) – a well-established determinant framework in implementation science to evaluate contextual factors influencing real-world outcomes, and overall implementation effectiveness, was used. The purpose was to identify key factors that could hinder the effective implementation of AI-driven diagnosis in eye care. METHODS Study Design A qualitative study based on data from a series of semi-structured interviews with key stakeholders (including clinicians, developers, leaders, and patients) was conducted by researchers at the UNSW Sydney School of Optometry and Vision Science and School of Business (Supplementary file 1). PARTICIPANT SELECTION Sampling Purposive sampling was used to identify key individuals from the professional contact list of authors to have a diverse representation across stakeholder groups. All participants were 18 years old or older and had an interest and/or experience in macular disease. Method of Approach (Recruitment) and Sample Size From September 2022 to March 2023, 147 potential participants were contacted by email. Participation was voluntary, and no financial incentives were offered for the completion of the study. Informed consent was obtained from all participants in accordance with the Declaration of Helsinki and approved by the University of New South Wales, Sydney Human Research Ethics Advisory Committee (HC210986; February 2022). DATA COLLECTION Demographic data including age, sex, and occupation was obtained using an online questionnaire (Qualtrics, Utah, USA). All interviews were conducted in English via Zoom (Zoom Video Communications Inc, California, USA) by authors AL and SH ( 17 ). Interview Guide A semi-structured interview guide was developed by research investigators following a literature review to explore key questions for each stakeholder group ( 17 ). The questions were designed to encourage open dialogue of AI and diagnostic technologies for AMD in Australia and were structured in sections: 1) experiences, 2) attitudes, 3) enablers, 4) barriers, 5) possible futures (anticipated personal and business use). Interviewers were encouraged to re-word, re-order or clarify the questions to facilitate the discussions. Participants were provided with a copy of the guide before their interview and were asked to only complete the interview once. All participants were informed that the purpose of the interview was to hear their experiences, views and attitudes toward digital diagnosis for macular disease, especially neovascular AMD. Audio Recording and Transcription All interviews were audio-recorded, transcribed verbatim and analysed using a theory-driven approach based on the Updated CFIR ( 16 ) in NVivo (Lumivero, Colorado, USA). The updated CFIR enabled systematic analysis of interview data and assessment of contextual factors influencing real-world implementation. Definitions Definitions have been adapted from the Updated CFIR ( 16 ), with the elimination of sub-domains that were not identified (Table 1 ). Table 1 Adapted CFIR and definitions I. INNOVATION DOMAIN The “thing” being implemented Project Innovation: Artificial intelligence and related technologies for the digital diagnosis of AMD, especially neovascular AMD; which remain in practice when implementation is complete and is distinct from the implementation process and strategies used to implement the technology. Construct Name Construct Definition The degree to which : A. Innovation Source The group that developed and/or visibly sponsored use of the innovation is reputable, credible, and/or trustable. B. Innovation Evidence-Based The innovation has robust evidence supporting its effectiveness. C. Innovation Relative Advantage The innovation is better than other available innovations or current practice i.e. better than a human clinician’s ability to diagnose AMD D. Innovation Adaptability The innovation can be modified, tailored, or refined to fit local context or needs. E. Innovation Complexity The innovation is complicated, which may be reflected by its scope and/or the nature and number of connections and steps. F. Innovation Cost The innovation purchase and operating costs are affordable. II. OUTER SETTING DOMAIN The setting in which the Inner Setting exists. There may be multiple Outer Settings and/or multiple levels within the Outer Setting. Project Outer Setting(s) : Community and local health district systems (public and private), state-wide health systems, national health schemes (within Australia), and other companies including affiliations, organisations and referral networks. Boundary between the outer and inner setting is the physical premises of healthcare clinics, bounded by brick and mortar. Construct Name Construct Definition The degree to which : A. Critical Incidents Large-scale and/or unanticipated events disrupt implementation and/or delivery of the innovation. B. Local Attitudes Sociocultural values (e.g., shared responsibility in helping recipients) and beliefs (e.g., convictions about the worthiness of recipients) encourage the Outer Setting to support implementation and/or delivery of the innovation. C. Partnerships & Connections Networks with external entities, including referral networks, academic affiliations, and professional organization networks. D. Financing Funding from external entities (e.g., grants, reimbursement) is available to implement and/or deliver the innovation. E. External Pressure External pressures drive implementation and/or delivery of the innovation. III. INNER SETTING DOMAIN The setting in which the innovation is implemented. There may be multiple Inner Settings and/or multiple levels within the Inner Setting. Project Inner Setting(s) : Single or multi-site clinics united by a single brand. Construct Name Construct Definition The degree to which : A. Structural Characteristics – Information Technology Infrastructure Technological systems for tele-communication, electronic documentation, and data storage, management, reporting, and analysis support functional performance of the Inner Setting. B. Culture There are shared values, beliefs, and norms across the Inner Setting. C. Access to Knowledge & Information Guidance and/or training is accessible to implement and deliver the innovation. IV. INDIVIDUALS DOMAIN The roles and characteristics of individuals. Project Role(s) : Innovation deliverers and recipients. Project Characteristic(s) : Need, capability and motivation of individuals’ roles. Construct Name – Roles Construct Definition The degree to which : A. Innovation Deliverers Individuals who are directly or indirectly delivering the innovation. B. Innovation Recipients Individuals who are directly or indirectly receiving the innovation. Construct Name - Characteristics Construct Definition The degree to which : 1. Need The individual(s) has deficits related to survival, well-being, or personal fulfillment, which will be addressed by implementation and/or delivery of the innovation. 2. Capability The individual(s) has interpersonal competence, knowledge, and skills to fulfill Role. 3. Motivation The individual(s) is committed to fulfilling Role. V. IMPLEMENTAION PROCCESS DOMAIN The activities and strategies used to implement the innovation. Project Implementation Process : Dynamic Sustainability Framework; activities and strategies. Construct Name – Roles Construct Definition The degree to which : A. Assessing Needs – Innovation Recipients Collects relevant? information about the priorities, preferences, and needs of recipients to guide implementation and delivery of the innovation. DATA ANALYSIS Data Coding AL extracted participants’ responses to the question ‘What is the biggest barrier to digital diagnosis or AI for macular disease in Australia? ’. Other relevant remarks including the keyword ‘barrier’ were also extracted for the analysis. Two authors AL and JN annotated the extracted text segments of the transcripts with the specific barrier to AI adoption using an inductive approach, with no predefined barrier list. Although each participant was encouraged to describe the single biggest barrier to implementation, if more than one was mentioned, it was counted separately. If the same barrier was mentioned more than once by a participant during their interview, it was only counted once. Once all the barriers had been summarised, iterative meetings were performed to go through all the coded texts and crosscheck the results. Agreement was reached on 65 of all 72 identified barriers, resulting in inter-coder reliability of 90%. Disagreements were discussed by AL and JN until an agreement was made on the optimal concept to code a specific fragment of the text by checking the full transcript context and re-reading the section of interest. Where ambiguity remained, the CFIR framework definitions were consulted to guide coding decisions. The usual sources of disagreement were the scope of one concept (3/7) and the specific barriers and facilitators that one concept should encompass (4/7). In the latter cases, a single concept was split into two separate concepts until agreement was achieved. The derived index of specific barriers was then mapped to the constructs of the Updated CFIR. Any disagreements about which Updated CFIR construct was the most appropriate for the concept were resolved by discussing the possible options until an agreement was reached. All authors verified that the domains and sub-domains appropriately described the extracted data. Assessment of Key Barriers To quantify the relative importance of the barrier to implementation, the frequency of each cited barrier coded under each domain and sub-domain across stakeholders was compared (Table 2 ). RESULTS A total of 37 individuals across 4 stakeholder groups participated in the study. Of the participants, there were 12 clinicians (32%), 10 healthcare leaders including senior health managers and professional organisation representatives (27%), 8 patients with AMD (22%), and 7 developers including researchers or technologists (19%). Stakeholder groups were labelled as clinicians (C), developers (D), leaders (L), and patients (P). Despite the interview guide stating that the purpose is to explore AI and related technologies for the digital diagnosis of macular disease, specifically neovascular AMD; participants were more focused on the innovation aspect (i.e. AI) during the interview. The ensuing discussions were not limited to macular disease but covered the implementation of digital diagnosis in eye health care in general. Table 2. Barrier counts CFIR Domain CFIR Sub-Domain Specific Barriers Clinicians n=12 Developers n=7 Leaders n=10 Patients n=8 Totals n=37 Innovation Innovation Source Limited credibility of the innovation source 0 1 0 0 1 Innovation Evidence-Base Lack of evidence and uncertainty whether sufficient resources are available to prove clinical usefulness . 2 2 1 0 5 Innovation Relative Advantage Generalisability : The innovation seemed inaccurate when applied locally and does not improve clinical workflow. 4 1 0 0 5 Innovation Adaptability Interoperability : Limited capacity to work with different imaging instruments, algorithms, and software. 2 1 1 0 4 Innovation Complexity Nature of ocular imaging being difficult to acquire clear images 0 1 0 0 1 Innovation Cost Purchase and operating costs 5 1 1 0 7 13 7 3 0 23 Outer Setting Critical Incidents Hackers obtaining personal health information 1 0 0 0 1 Local Attitudes Uncertainty about whether it would be possible to maintain equity across different organisations and patient populations 0 1 3 0 4 Partnerships & Connections Lack of partnerships and connections to support utilisation 1 0 0 0 1 Financing Uncertainty about financing; role of government subsidy 0 2 1 1 4 External Pressure Misaligned priorities of different stakeholders 0 0 1 0 1 2 3 5 1 11 Inner Setting Structural Characteristics – Information Technology Infrastructure Lack of infrastructure to ensure appropriate data security and privacy ; fear of data misuse 2 1 0 0 3 Large data storage requirements 2 0 0 0 2 Culture Resistance to change 2 1 1 1 5 Access To Knowledge & Information Optometrists’ lack of training in being able to use the AI recommendations 1 2 1 0 4 7 4 2 1 14 Individuals Innovation Deliverers Need Optometrists’ fear of job loss 0 0 1 0 1 Capability Risk of over-reliance 1 1 3 0 5 Fear that professional judgement of clinicians will be diminished 0 0 1 0 1 Motivation Lack of trust in the technology itself 1 1 1 0 3 Innovation Recipients Need Patient need for human empathy and oversight 2 0 1 4 7 Motivation Lack of individual patient awareness regarding the need for eye care 1 1 0 1 3 5 3 7 5 21 Implementation Process Assessing Needs – Innovation Deliverers Training and knowledge of innovation deliverers, persistent need to retrain 1 0 2 0 3 1 0 2 0 3 Table 3 summarises the domain count across the stakeholder groups in order of highest to lowest frequency. For clinicians and developers, ‘innovation’ was the most frequently quoted domain. For leaders and patients ‘individuals’ were the most cited; however, leaders focussed on innovation deliverers, whereas patients were more concerned about innovation recipients. Implementation processes were mentioned the least across the stakeholders as an identified barrier. Table 3. Key domains identified by stakeholder groups. Clinicians Developers Leaders Patients Innovation (13) Innovation (7) Individuals (7) Individuals (5) Inner setting (7) Inner setting (4), Individuals (4) Outer setting (5) Outer setting (1), Inner setting (1) Individuals (5) Innovation (4) Outer setting (2) Outer setting (3) Inner setting (2), Implementation processes (2) Implementation process (1) 1. Clinicians The costs of purchasing and operating an innovation was expressed by many clinicians as a key barrier to its implementation (Figure 1). “The only barrier for practitioners to acquire it is cost…The only barrier for every practice to have it…is cost” – C4 “To be honest, the biggest factor…is cost…assuming it works reasonably well, I think, it’s mainly a cost thing.” – C12 Clinicians also cited concerns of generalisability and the innovations’ lack of relative advantage over their own ability to make accurate diagnoses (Figure 1). “If the AI incorrectly does, like pick up an artefact…an error in judgement can be made…it highlights that there’s something severely wrong and they’re unnecessarily referred [false positive] or the opposite where it misses something, and they’re told…there’s nothing wrong [false negative].” – C8 The structural limitations within their clinical practices were addressed (Figure 1), with the lack of information technology infrastructure to support the functionality of the innovation, including data storage and cyber security (digital health privacy). “We are still some ways away from having the infrastructure to do so…hackers managed to obtain personal information” – C1 “Other barriers…data storage, privacy issues” – C12 2. Developers Developers identified similar barriers to clinicians correlating to the innovation domain. “ I think people have issues with AI and decision support…how can I trust that system has done it the same way that I would have as a clinician? ” – D6 “ It comes back to knowledge, right?...Like how do they know this is legit? That whatever technology that is put forward to me is safe, is accurate, is relevant?” – D5 However, their views were more generalised across subdomains (Figure 1); with the only barrier with repeated citations being the lack of evidence for the innovation and hence the uncertainty of its clinical usefulness in the real world. “I think it's definitely…a near term issue in terms of getting enough data to make that really work” – D1 “So that's a real barrier for us, conducting the research itself in a setting…out into the world to do, you know, pragmatic studies, but that's not currently what we're doing yet.” – D2 3. Leaders Leaders on the other hand focussed more on individuals (innovation deliverers) (Figure 1), highlighting the potential risk of over-reliance on the innovation and the consequent deskilling of clinicians. “AI lends itself to sort of protocol-based care…you feed this information in, you get a green light or a red light…but I think the risk is over-reliance on that…people just sort of blindly trusting it “ – L1 “People get…a bit blasé about re-checking because it’s quite good for a while and then there’s a false negative and they’re not doing the adequate other tests…but it doesn’t have the same sensitivity or specificity”. – L3 “It’s an algorithm, it’s a model….but people in the system have got to remember that that’s what it is. It’s a tool. It’s an aid…So people have to be aware that you can’t over-rely on a model…Practitioners need to appreciate both the strength of an AI diagnosis as well as the shortcomings and where the tool might be at its weakest. So I think that worries me because people have a tendency to over-rely on technology these days.” – L5 Another commonly expressed barrier was the uncertainty of being able to maintain equity across different organisations and patient populations (Figure 1). “It’s also about how you roll out the accessibility and availability of these tools to the broader community” – L5 “One of the problems across the health area is consistency. And so you want everybody to get the same level of care…You’re building up a new silo with people who are for digital health… and people who are not so knowledgeable” – L7 4. Patients Patients posed the most individual-centric outlook (innovation recipients) (Figure 1), with the key barrier being the need and the lack thereof human empathy with the implementation AI. “I think there’s sort of an added element in the personal connection…the patient if they feel safe and secure, is much more able to speak about their concerns….if you just leave them up to a computer, the computer can only do what it’s set up to do.” – P5 “I’m reassured by the fact there is human intervention.” – P6 “I like the personal touch…maybe I’m a bit old fashioned still…whereby I’d like to actually be able to ask the person whatever questions…it’s about having the person in the room with you” – P8 Supplementary file 2 summarises all the other barriers. DISCUSSION Although there has been a revolution of AI applications, it is well-recognised that its integration remains limited in healthcare ( 12 ). Our study addressed this by exploring barriers to the implementation of AI for digital diagnosis in eye care and found that clinicians are concerned about the cost and financing of AI, developers face challenges proving its usefulness in the clinical workflow, leaders foreboded the consequence of clinicians deskilling due to potential over-reliance on AI and patients had apprehension for the absence of human interaction. Most importantly, our results highlight the differing focus of each stakeholder group and their perceived barriers, whereby the misalignment of perspectives itself can be an inherent barrier to the implementation of the innovation. Barriers to AI in healthcare are typically explored from the perspective of clinicians ( 19 ), leaders ( 20 ) and patients ( 21 ) or an incomplete combination of stakeholders ( 22 , 23 ). To our knowledge, there is currently no framework-based analysis of all the stakeholders’ perspectives. Stakeholder co-design is crucial from the initial development phase ( 24 , 25 ) of digital health innovations to successfully deliver patient-centred care ( 25 ). Thus, a synthesised evaluation (Table 4 ) combining multiple stakeholder perspectives is fundamental to achieving a more holistic understanding of barriers to AI in digital diagnosis. Table 4 Summary of barriers with a high count* (frequency of 5 or more) across stakeholders. * The count of ‘Costs’ and ‘Financing’ were added as commonly cited together. Clinicians Developers Leaders Patients Costs Financing ● ● ● ● ● ● Lack of Clinical Usefulness ● ● ● Not Improving Workflow ● ● Deskilling of Clinicians ● ● ● Lack of Human Interaction for Patients ● ● ● Resistance to Change ● ● ● ● By default in the design of our study, all the stakeholders had at least a fundamental awareness of AI. However, the level of knowledge in its potential role and use case of digital diagnosis innovations varied across individuals. This further highlighted the importance of co-design, and the need for establishing a common language and understanding of the technological innovation as a foundation for successful implementation. Costs and Financing The multi-factorial challenge of costs and financing involved in implementing AI digital diagnosis systems was unanimously identified as a key barrier across stakeholder groups, including high initial start-up costs, operating expenses and the uncertainty of financing the implementation of the innovation. These issues are widely cited in literature ( 15 , 26 – 28 ), and highlight the importance of favourable financial models and reimbursement policies in the adoption of innovative technologies such as AI, despite its potential benefits. Further exploration of stakeholders’ willingness to pay ( 29 ) can elicit their perceived value and preferences regarding the implementation of AI and contribute towards building a sustainable financial model to overcome the barrier of costs and financing. Lack of Clinical Usefulness and Not Improving Workflow Another inherent barrier arises from the uncertainty of the usefulness of the AI innovation, and whether it ultimately improves clinical workflow. This barrier combines notions of needing robust evidence supporting AI effectiveness, accuracy, and ability to outperform clinicians, and its lack thereof being a barrier for all innovation deliverers (clinicians, developers, and leaders). There has been a shift in literature, with initial concerns arising from the limited number of randomised controlled trials in the clinical context ( 30 ) to its persistent lack of generalisability and practicality ( 31 ) at present, despite there being an expanding interest in AI. The need for more comprehensive research to minimise publication bias and to prioritise patient-relevant outcomes when evaluating AI in clinical practice has been recognised ( 31 ). Identifying clinical “pain points” for more solution-driven AI innovations and enhancing collaborations between stakeholders ( 30 ) are also acknowledged as potential solutions to improving its usefulness and the overall workflow. Deskilling of Clinicians (Innovation Deliverers) All innovation deliverers expressed the potential risk of over-reliance by clinicians as a barrier as well as the consequence of deskilling; compromising one’s competence, knowledge, and skills. Over-reliance is identified as a precursor to deskilling ( 26 , 32 ), with concerns at present regarding increased dependence on the capabilities of automation ( 26 ). Deskilling in the clinical context can be defined as a “situation where clinicians experience reduced opportunities to exercise particular skills” ( 33 ) as an unintended consequence of AI implementation in the medical profession. This is considered an inevitable process in improving efficiency and reducing costs, as exemplified by the now widespread adoption of electronic health records and clinical practice guidelines ( 34 ). The alternative phenomenon of upskilling or reskilling ( 33 , 35 ) involving enhancing one’s existing skill set or learning new skills, respectively was not discussed by any of the stakeholders. However, it is to acknowledge that there is a growing consensus that “AI is not going to replace humans, but humans with AI are going to replace humans without AI”( 36 ). Thus, focusing on the positive potentials of AI, reskilling efforts should centre on preparing eye care professionals to effectively collaborate with AI systems ( 35 ) whilst maintaining the clinician’s critical role in patient outcomes. Lack of Human Interaction for Patients (Innovation Recipients) Moreover, the lack of human interaction for patients was identified as a key barrier by both innovation deliverers and recipients, but most commonly by the patients. This highlights the significance of the patient-clinician relationship whereby building rapport is crucial for the patient to entrust the clinician to act on behalf of their health and have their best interests at heart ( 33 ). Thus, this barrier arises from the fear that healthcare is shifting away from patient interactions and towards data analytics with a loss of the holistic approach ( 32 ). It is commonly cited in literature that patients prefer human interaction over automation ( 32 , 37 ), despite the capability of AI to outperform humans ( 38 ). The strive for patient-centred care, which involves patients as active participants in their care, is considered an essential part of high-quality healthcare systems ( 39 ). Thus co-design of the implementation of AI should be considered such that it can instead foster the connection between clinicians and patients( 40 ). Resistance to Change Finally, similar views emerged from all stakeholders with resistance to change, being cited as a barrier by at least one stakeholder in each group (Table 4 ). Change resistance ( 41 , 42 ) being an intrinsic and complex problem, is amplified by the sociotechnical implications of AI ( 43 ); with apprehensions raised across healthcare organisations by clinicians ( 44 ), patients ( 45 ) and healthcare leaders ( 27 ) alike. However, resistance to change can be expressed differently amongst the stakeholders ( 43 ). This can take the form of: 1. rejection (disinclination to adopt the innovation, often due to a reluctance to change the status quo), 2. postponement (accepting the innovation in principle but deciding not to adopt it) and/or 3. opposition (actively engaging in strategies to prevent the innovation’s success) ( 46 ). Given the intricate nuances of change resistance, successful implementation of innovations requires effective change management strategies ( 47 ) whereby leaders play a crucial role ( 48 ). Successful navigation of change resistance can be exemplified by the national implementation of eye health electronic patient records in Scotland and Wales, with increasing adoption across England ( 49 – 51 ). Strengths and Limitations The main strength of our analysis lies in using purposive sampling to identify key stakeholders, which enabled us to capture diverse and informed perspectives on the barriers to implementing AI in eye care diagnostics. Selecting individuals who play a critical role in the implementation process ensured the insights gathered are comprehensive and reflective of real-world challenges. While the sample distribution across stakeholder groups was uneven, this was by design and reflects real-world variability in end-users and their associated willingness to participate. The use of purposive sampling however ensured the inclusion of key informants with relevant experience in AI and AMD, which thereby influenced the specific number and role of the participants ( 17 ). This study was conducted within the Australian healthcare system and focused on AMD. While the findings offer insights into stakeholder perspectives on AI implementation in this context, the transferability of these insights to other diseases, demographics or international settings may be limited by differences in healthcare infrastructure, funding models, and disease-specific workflows. However, the open dialogue format during the semi-structured interviews naturally allowed responses to explore a wider perspective on AI's role in eye care diagnostics. Notions around cost and financing and deskilling of clinicians are likely to be applicable across ophthalmic diseases and digital diagnostics. Sustainable reimbursement is well-recognised as a pivotal enabler for the adoption of novel technologies within healthcare systems ( 15 , 52 , 53 ). Thus, it is critical to understand the stakeholders that determine reimbursement and consider which models, including fee-for-service and value-based care, may best address the financial sustainability of AI implementation in eye care ( 52 ). Recent evaluations of cost-effectiveness and real-world implementation feasibility highlight the importance of aligning AI integration with performance-based reimbursement frameworks; therefore, advocating for hybrid models that incentivise innovation whilst ensuring fiscal sustainability ( 52 , 53 ). Another similarity that is echoed in literature is the need for clinicians to be AI competent. Beyond the imperative for eye care professionals to reskill to utilise AI systems ( 35 ), emerging perspective suggest that AI needs to be integrated at the education level, across different aspects of the clinical curriculum to combine technical skills with ethical reasoning ( 15 , 54 , 55 ). Thus, current and future clinicians require adequate training to be equipped to navigate the complexities of AI-driven healthcare ( 55 ). Recommendations for Future Directions To reduce barriers to adoption, financing models could be adapted through national procurement guidelines, targeted grants, and expanded tax incentives, particularly for small and medium-sized enterprises developing AI tools, as recommended in the Australian Alliance for AI in Healthcare’s National Policy Roadmap ( 56 ). Successful examples, such as Singapore’s semi-automated diabetic retinopathy screening program, demonstrate how cost savings and clinical efficiency can be achieved ( 57 ). However, cost-effectiveness varies by geography, deployment strategy, and healthcare funding model. High-income countries may benefit more due to higher human grader costs, but studies from lower-income settings show mixed results ( 57 ). Fee-for-service, bundled payments and subscription-based deployment strategies are likely to influence uptake. Moreover, tailoring financial strategies to the structure of healthcare systems, whether public, private, or hybrid, can enhance relevance and sustainability. In Australia, corporate-led rollout models have been described as a strong catalyst for uptake, with even a small number of providers able to shift practice norms across the profession ( 17 ). To mitigate deskilling and promote upskilling or reskilling, leaders should proactively redesign training frameworks to preserve core clinical competencies while fostering human–AI collaboration. Strategies include maintaining hands-on learning opportunities, integrating AI literacy into curricula, and using AI as a tool for feedback and simulation-based training. Successful examples include double-reading workflows where clinicians assess cases independently before reviewing AI outputs, and structured validation sessions that reinforce clinical reasoning ( 58 , 59 ). These approaches align with broader calls for safeguards against skill erosion and support a shift toward hybrid intelligence, where clinicians and AI systems co-evolve to enhance diagnostic quality and professional resilience. Concrete examples of successful co-design in healthcare highlight its potential to address both resistance to change and concerns about loss of human connection. A statewide initiative in Australia co-designed AI workflows to analyse patient-reported experience measures, enabling clinicians to integrate narrative feedback into quality improvement cycles while maintaining human oversight and accountability ( 60 ). Similarly, the development of a deep learning skin lesion classifier incorporated a multidisciplinary co-design framework to address ethical, legal, and technical risks—such as overdiagnosis and bias—ensuring the system supported clinician decision-making rather than replacing it ( 60 ). These examples demonstrate how co-design can preserve the human elements of care, enhance trust, and tailor AI systems to clinical realities. CONCLUSION This framework-based analysis of barriers to implementation of AI in digital diagnosis provides a more complete overview of stakeholder perspectives, highlighting differences in the views of clinicians, healthcare leaders, patients and developers including researchers or technologists. Five key barriers were identified as costs and financing, lack of perceived clinical usefulness or not improving workflow, deskilling of clinicians, lack of human interaction for patients and resistance to change. Abbreviations AI Artificial Intelligence AMD Age-related Macular Degeneration CFIR Consolidated Framework for Implementation Research Declarations Ethics approval and consent to participate: Written informed consent was obtained from all participants in accordance with the tenets of the Declaration of Helsinki and approved by the UNSW Sydney Human Research Ethics Advisory Committee (HC210986; February 2022). Consent for publication : Not applicable. Funding: This work was carried out by the UNSW-Roche Digital Diagnostics Project and funded by Roche Products, Australia. The funder had no role in study design, data collection, data analysis, data interpretation, writing of the report, and decision to submit the paper for publication. Author Contribution FS and MAW conceived the idea and supervised the study. Interviews were conducted by AL and SH. Data analysis was performed by JN and AL. JN wrote the manuscript aided by AL. All authors read and approved the final manuscript. Acknowledgements: Not applicable. Availability of data and material: The data and material generated and analysed during the study are not publicly available to protect participant privacy. References Oxford English Dictionary. Oxford University Press; 2023. artificial intelligence, n. Fui-Hoon Nah F, Zheng R, Cai J, Siau K, Chen L. Generative AI and ChatGPT: Applications, challenges, and AI-human collaboration. Journal of Information Technology Case and Application Research. 2023;25(3):277–304. Alowais SA, Alghamdi SS, Alsuhebany N, Alqahtani T, Alshaya AI, Almohareb SN, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med Educ. 2023;23(1):689. Ouanes K, Farhah N. Effectiveness of Artificial Intelligence (AI) in Clinical Decision Support Systems and Care Delivery. J Med Syst. 2024;48(1):74. Sim I, Gorman P, Greenes RA, Haynes RB, Kaplan B, Lehmann H, et al. Clinical decision support systems for the practice of evidence-based medicine. J Am Med Inform Assoc. 2001;8(6):527–34. Sutton RT, Pincock D, Baumgart DC, Sadowski DC, Fedorak RN, Kroeker KI. An overview of clinical decision support systems: benefits, risks, and strategies for success. NPJ Digit Med. 2020;3:17. Wong WL, Su X, Li X, Cheung CM, Klein R, Cheng CY, et al. Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis. Lancet Glob Health. 2014;2(2):e106-16. Ly A, Nivison-Smith L, Zangerl B, Assaad N, Kalloniatis M. Advanced imaging for the diagnosis of age-related macular degeneration: a case vignettes study. Clin Exp Optom. 2018;101(2):243–54. Neely DC, Bray KJ, Huisingh CE, Clark ME, McGwin G, Jr., Owsley C. Prevalence of Undiagnosed Age-Related Macular Degeneration in Primary Eye Care. JAMA Ophthalmol. 2017;135(6):570–5. Ferrara D, Newton EM, Lee AY. Artificial intelligence-based predictions in neovascular age-related macular degeneration. Curr Opin Ophthalmol. 2021;32(5):389–96. Kumar H, Goh KL, Guymer RH, Wu Z. A clinical perspective on the expanding role of artificial intelligence in age-related macular degeneration. Clin Exp Optom. 2022;105(7):674–9. Wu K, Wu E, Theodorou B, Liang W, Mack C, Glass L, et al. Characterizing the Clinical Adoption of Medical AI Devices through U.S. Insurance Claims. NEJM AI. 2024;1(1):AIoa2300030. Tseng R, Gunasekeran DV, Tan SSH, Rim TH, Lum E, Tan GSW, et al. Considerations for Artificial Intelligence Real-World Implementation in Ophthalmology: Providers' and Patients' Perspectives. Asia Pac J Ophthalmol (Phila). 2021;10(3):299–306. Li JO, Liu H, Ting DSJ, Jeon S, Chan RVP, Kim JE, et al. Digital technology, tele-medicine and artificial intelligence in ophthalmology: A global perspective. Prog Retin Eye Res. 2021;82:100900. Singh RP, Hom GL, Abramoff MD, Campbell JP, Chiang MF, Intelligence AAOTFoA. Current Challenges and Barriers to Real-World Artificial Intelligence Adoption for the Healthcare System, Provider, and the Patient. Transl Vis Sci Technol. 2020;9(2):45. Damschroder LJ, Reardon CM, Widerquist MAO, Lowery J. The updated Consolidated Framework for Implementation Research based on user feedback. Implement Sci. 2022;17(1):75. Ly A, Herse S, Williams M-A, Stapleton F. Artificial Intelligence for Age-Related Macular Degeneration Diagnosis in Australia: A Novel Qualitative Interview Study. Ophthalmic Physiol Opt. 2025. The Center for Implementation. CFIR 2.0. V2025.01 ed2025. p. Adapted from Damschroder, L. J., Reardon, C. M., Widerquist, M. A. O., et al. (2022). The updated consolidated framework for implementation research based on user feedback. Implementation Science, 17, 75. https://doi.org/10.1186/s13012-022-01245-0 . Fujimori R, Liu K, Soeno S, Naraba H, Ogura K, Hara K, et al. Acceptance, Barriers, and Facilitators to Implementing Artificial Intelligence-Based Decision Support Systems in Emergency Departments: Quantitative and Qualitative Evaluation. JMIR Form Res. 2022;6(6):e36501. Neher M, Petersson L, Nygren JM, Svedberg P, Larsson I, Nilsen P. Innovation in healthcare: leadership perceptions about the innovation characteristics of artificial intelligence-a qualitative interview study with healthcare leaders in Sweden. Implement Sci Commun. 2023;4(1):81. Robinson R, Liday C, Lee S, Williams IC, Wright M, An S, et al. Artificial Intelligence in Health Care-Understanding Patient Information Needs and Designing Comprehensible Transparency: Qualitative Study. JMIR AI. 2023;2:e46487. Liao X, Yao C, Jin F, Zhang J, Liu L. Barriers and facilitators to implementing imaging-based diagnostic artificial intelligence-assisted decision-making software in hospitals in China: a qualitative study using the updated Consolidated Framework for Implementation Research. BMJ Open. 2024;14(9):e084398. Ho V, Brown Johnson C, Ghanzouri I, Amal S, Asch S, Ross E. Physician- and Patient-Elicited Barriers and Facilitators to Implementation of a Machine Learning-Based Screening Tool for Peripheral Arterial Disease: Preimplementation Study With Physician and Patient Stakeholders. JMIR Cardio. 2023;7:e44732. van de Sande D, Chung EFF, Oosterhoff J, van Bommel J, Gommers D, van Genderen ME. To warrant clinical adoption AI models require a multi-faceted implementation evaluation. NPJ Digit Med. 2024;7(1):58. Sanz MF, Acha BV, Garcia MF. Co-Design for People-Centred Care Digital Solutions: A Literature Review. Int J Integr Care. 2021;21(2):16. Ahmed MI, Spooner B, Isherwood J, Lane M, Orrock E, Dennison A. A Systematic Review of the Barriers to the Implementation of Artificial Intelligence in Healthcare. Cureus. 2023;15(10):e46454. He J, Baxter SL, Xu J, Xu J, Zhou X, Zhang K. The practical implementation of artificial intelligence technologies in medicine. Nat Med. 2019;25(1):30–6. Parikh RB, Helmchen LA. Paying for artificial intelligence in medicine. NPJ Digit Med. 2022;5(1):63. Abbas S, Usmani A, Imran M. Willingness To Pay And Its Role In Health Economics. Journal of Bahria University Medical and Dental College. 2018;09:62–6. Grant K, McParland A, Mehta S, Ackery AD. Artificial Intelligence in Emergency Medicine: Surmountable Barriers With Revolutionary Potential. Ann Emerg Med. 2020;75(6):721–6. Han R, Acosta JN, Shakeri Z, Ioannidis JPA, Topol EJ, Rajpurkar P. Randomised controlled trials evaluating artificial intelligence in clinical practice: a scoping review. Lancet Digit Health. 2024;6(5):e367-e73. Loncaric F, Camara O, Piella G, Bijnens B. Integration of artificial intelligence into clinical patient management: focus on cardiac imaging. Rev Esp Cardiol (Engl Ed). 2021;74(1):72–80. Leslye Denisse Dias D. Deskilling of medical professionals: an unintended consequence of AI implementation? Giornale di Filosofia. 2021;2(2). Hoff T. Deskilling and adaptation among primary care physicians using two work innovations. Health Care Manage Rev. 2011;36(4):338–48. Li L. Reskilling and Upskilling the Future-ready Workforce for Industry 4.0 and Beyond. Inf Syst Front. 2022:1–16. Ignatius A, Lakhani, K. AI Won’t Replace Humans — But Humans With AI Will Replace Humans Without AI Harvard Business Review2023 [Available from: https://hbr.org/2023/08/ai-wont-replace-humans-but-humans-with-ai-will-replace-humans-without-ai . Longoni C, Bonezzi A, Morewedge CK. Resistance to Medical Artificial Intelligence. Journal of Consumer Research. 2019;46(4):629–50. Riedl R, Hogeterp SA, Reuter M. Do patients prefer a human doctor, artificial intelligence, or a blend, and is this preference dependent on medical discipline? Empirical evidence and implications for medical practice. Front Psychol. 2024;15:1422177. Greene SM, Tuzzio L, Cherkin D. A framework for making patient-centered care front and center. Perm J. 2012;16(3):49–53. Topol E. Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again: Basic Books; 2019. Talwar S, Dhir A, Islam N, Kaur P, Almusharraf A. Resistance of multiple stakeholders to e-health innovations: Integration of fundamental insights and guiding research paths. Journal of Business Research. 2023;166:114135. Desveaux L, Soobiah C, Bhatia RS, Shaw J. Identifying and Overcoming Policy-Level Barriers to the Implementation of Digital Health Innovation: Qualitative Study. J Med Internet Res. 2019;21(12):e14994. Yang Y, Ngai EWT, Wang L. Resistance to artificial intelligence in health care: Literature review, conceptual framework, and research agenda. Information & Management. 2024;61(4):103961. Prakash AV, Das S. Medical practitioner's adoption of intelligent clinical diagnostic decision support systems: A mixed-methods study. Information & Management. 2021;58(7):103524. Park EH, Werder K, Cao L, Ramesh B. Why do Family Members Reject AI in Health Care? Competing Effects of Emotions. Journal of Management Information Systems. 2022;39(3):765–92. Kleijnen M, Lee N, Wetzels M. An exploration of consumer resistance to innovation and its antecedents. Journal of Economic Psychology. 2009;30(3):344–57. Grol R, Wensing M. Effective Implementation of Change in Healthcare. Improving Patient Care2020. p. 45–71. Lemak CH, Pena D, Jones DA, Kim DH, Guptill J. Leadership to Accelerate Healthcare's Digital Transformation: Evidence From 33 Health Systems. J Healthc Manag. 2024;69(4):267–79. Wales launches first national digital eye care patient record system [press release]. OpenEyes [press release]. 2022. OpenEyes Platform Proves Concept in Glasgow [press release]. 2024. Abramoff MD, Dai T, Zou J. Scaling Adoption of Medical AI — Reimbursement from Value-Based Care and Fee-for-Service Perspectives. NEJM AI. 2024;1(5):AIpc2400083. El Arab RA, Al Moosa OA, Sagbakken M. Economic, ethical, and regulatory dimensions of artificial intelligence in healthcare: an integrative review. Front Public Health. 2025;13:1617138. Paranjape K, Schinkel M, Nannan Panday R, Car J, Nanayakkara P. Introducing Artificial Intelligence Training in Medical Education. JMIR Med Educ. 2019;5(2):e16048. Syeda LH, Batool Z, Hayder Z, Ali S. Medical undergraduate students' awareness and perspectives on artificial intelligence: A developing nation's context. BMC Med Educ. 2025;25(1):1060. Australian Alliance for Artificial Intelligence in Healthcare. A National Policy Roadmap for Artificial Intelligence in Healthcare. Rajesh AE, Davidson OQ, Lee CS, Lee AY. Artificial Intelligence and Diabetic Retinopathy: AI Framework, Prospective Studies, Head-to-head Validation, and Cost-effectiveness. Diabetes Care. 2023;46(10):1728–39. Nilsson C. The artificial intelligence (AI) competence paradox: how AI reshapes clinical expertise. Transforming Government: People, Process and Policy. 2025. Natali C, Marconi L, Dias Duran LD, Cabitza F. AI-induced Deskilling in Medicine: A Mixed-Method Review and Research Agenda for Healthcare and Beyond. Artificial Intelligence Review. 2025;58(11):356. Canfell OJ, Chan W, Pole JD, Engstrom T, Saul T, Daly J, et al. Artificial intelligence after the bedside: co-design of AI-based clinical informatics workflows to routinely analyse patient-reported experience measures in hospitals. BMJ Health Care Inform. 2024;31(1). Additional Declarations No competing interests reported. Supplementary Files Supplementaryfile1.docx Supplementaryfile2.docx Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 23 Dec, 2025 Reviews received at journal 20 Dec, 2025 Reviews received at journal 07 Dec, 2025 Reviewers agreed at journal 30 Nov, 2025 Reviewers agreed at journal 30 Nov, 2025 Reviewers invited by journal 23 Oct, 2025 Editor assigned by journal 23 Oct, 2025 Submission checks completed at journal 21 Oct, 2025 First submitted to journal 21 Oct, 2025 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-7911455\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":false,\"archivedVersions\":[],\"articleType\":\"Research Article\",\"associatedPublications\":[],\"authors\":[{\"id\":538832053,\"identity\":\"4edc2e35-ee08-48a1-a66b-7e3cdfcdc27e\",\"order_by\":0,\"name\":\"Judy Nam\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"UNSW Sydney\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Judy\",\"middleName\":\"\",\"lastName\":\"Nam\",\"suffix\":\"\"},{\"id\":538832054,\"identity\":\"a4174400-1831-4692-b84e-1b11e66bc57f\",\"order_by\":1,\"name\":\"Angelica Ly\",\"email\":\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwElEQVRIiWNgGAWjYFAC5gaGD1CmBJFaGBsYZ5CshZmHJC38DYxt0rZtd+wNDjAfvM3DYJfYQEiLxAGglty2Z4kbDrAlW/MwJBPWwgDRcjjB7ACPmTQPAzNhLfIgLZZth+3NDvB/A2qpJ6zFAKSFse0w47YDPGxALYcJazE8zNhs2XPucOL+w2zGlnMMjhsT1CJ3vPngjR9lh+0l25sf3nhTUS1LUAsDMwOLBJQBcidB9RC1HwirGQWjYBSMghENAJirOG1DuKxxAAAAAElFTkSuQmCC\",\"orcid\":\"\",\"institution\":\"UNSW Sydney\",\"correspondingAuthor\":true,\"prefix\":\"\",\"firstName\":\"Angelica\",\"middleName\":\"\",\"lastName\":\"Ly\",\"suffix\":\"\"},{\"id\":538832055,\"identity\":\"0bcc2c4f-081c-48dd-a00f-b444953ef0e1\",\"order_by\":2,\"name\":\"Sarita Herse\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"UNSW Sydney\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Sarita\",\"middleName\":\"\",\"lastName\":\"Herse\",\"suffix\":\"\"},{\"id\":538832056,\"identity\":\"49525048-5519-4f16-b7d5-a00a8a82b22b\",\"order_by\":3,\"name\":\"Chris Lim\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"National University Health System\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Chris\",\"middleName\":\"\",\"lastName\":\"Lim\",\"suffix\":\"\"},{\"id\":538832057,\"identity\":\"1548ecb8-d56d-4a22-a683-a1225e5c3589\",\"order_by\":4,\"name\":\"Mary-Anne Williams\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"UNSW Sydney\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Mary-Anne\",\"middleName\":\"\",\"lastName\":\"Williams\",\"suffix\":\"\"},{\"id\":538832058,\"identity\":\"ec72da8e-7300-42de-87d0-47e5214d0d31\",\"order_by\":5,\"name\":\"Fiona Stapleton\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"UNSW Sydney\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Fiona\",\"middleName\":\"\",\"lastName\":\"Stapleton\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2025-10-21 07:08:28\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-7911455/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-7911455/v1\",\"draftVersion\":[],\"editorialEvents\":[],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":95224708,\"identity\":\"0d15162a-9755-4831-8088-f296cefea9f1\",\"added_by\":\"auto\",\"created_at\":\"2025-11-05 16:24:11\",\"extension\":\"docx\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":347176,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"BarriersManuscriptclean.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/a22b4aa270f50df63671d865.docx\"},{\"id\":95224415,\"identity\":\"baed9afd-c7ad-46e5-9b29-7c098656ff54\",\"added_by\":\"auto\",\"created_at\":\"2025-11-05 16:23:42\",\"extension\":\"json\",\"order_by\":1,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":7626,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"b342a191b86a4662a02070a5d3e3ed72.json\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/77b8aaa6df2a1e4f04de021f.json\"},{\"id\":95102601,\"identity\":\"22db1dda-a0ed-4303-9e38-e6c1ad07d12e\",\"added_by\":\"auto\",\"created_at\":\"2025-11-04 10:23:03\",\"extension\":\"docx\",\"order_by\":2,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":60383,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Supplementaryfile1.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/624381f3fbb12ec0c0c69b79.docx\"},{\"id\":95224785,\"identity\":\"463b5564-6f54-4627-adb5-dcf4031dbf2e\",\"added_by\":\"auto\",\"created_at\":\"2025-11-05 16:24:17\",\"extension\":\"docx\",\"order_by\":3,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":20353,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Supplementaryfile2.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/18cca277294bf2c279d68e82.docx\"},{\"id\":95225539,\"identity\":\"dba4c24e-2995-4392-8f5c-acaf5eaf8624\",\"added_by\":\"auto\",\"created_at\":\"2025-11-05 16:25:11\",\"extension\":\"xml\",\"order_by\":4,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":161318,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"b342a191b86a4662a02070a5d3e3ed721enriched.xml\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/ffccdc2e7f6fd563bf84a266.xml\"},{\"id\":95224702,\"identity\":\"740d26a6-3521-4e90-8582-237eeaf1dce2\",\"added_by\":\"auto\",\"created_at\":\"2025-11-05 16:24:11\",\"extension\":\"png\",\"order_by\":6,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":38926,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Onlinefloatimage1.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/e225b3618d8a9390cf73ad7d.png\"},{\"id\":95102606,\"identity\":\"d25e4646-ceff-46ed-a4e6-3895fd5af57b\",\"added_by\":\"auto\",\"created_at\":\"2025-11-04 10:23:03\",\"extension\":\"xml\",\"order_by\":7,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":160011,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"b342a191b86a4662a02070a5d3e3ed721structuring.xml\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/c438b40ab54d42cb9e5cfa01.xml\"},{\"id\":95102604,\"identity\":\"1318ff4f-87e8-4ebe-9313-c6897c8ca084\",\"added_by\":\"auto\",\"created_at\":\"2025-11-04 10:23:03\",\"extension\":\"html\",\"order_by\":8,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"acdc-reference\",\"size\":176885,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"earlyproof.html\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/1e5a04ecc639800338590dd8.html\"},{\"id\":95102596,\"identity\":\"95db6ffb-dad4-42e2-923b-52d28b716d7c\",\"added_by\":\"auto\",\"created_at\":\"2025-11-04 10:23:03\",\"extension\":\"png\",\"order_by\":1,\"title\":\"Figure 1\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":176678,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eBarriers mapped across domains and stakeholder types: clinicians (C), developers (D), leaders (L), and patients (P). Adapted from The Center for Implementation’s image (18) of the Updated CFIR (16).\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage124.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/4aa9f4eaa812176c972c68cf.png\"},{\"id\":95230389,\"identity\":\"57a90dc1-2b81-41c5-9070-7a788762d6a9\",\"added_by\":\"auto\",\"created_at\":\"2025-11-05 16:37:24\",\"extension\":\"pdf\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":1850659,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/a8b8dcde-e2ae-468c-9d5e-3fbebec04d90.pdf\"},{\"id\":95102598,\"identity\":\"195ddef1-a9de-4391-a466-f3498836eda6\",\"added_by\":\"auto\",\"created_at\":\"2025-11-04 10:23:03\",\"extension\":\"docx\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":60383,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Supplementaryfile1.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/931b5bb1ed1a459d847fc38e.docx\"},{\"id\":95102597,\"identity\":\"f0ea6aa6-3b60-48cc-aeff-82e2dbfffe3a\",\"added_by\":\"auto\",\"created_at\":\"2025-11-04 10:23:03\",\"extension\":\"docx\",\"order_by\":1,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":20353,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"Supplementaryfile2.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7911455/v1/8b597befc01dbefffe5f88b4.docx\"}],\"financialInterests\":\"No competing interests reported.\",\"formattedTitle\":\"Stakeholder Perspectives of Implementation Barriers of Artificial Intelligence in Eye Care: A qualitative framework-based study\",\"fulltext\":[{\"header\":\"KEY POINTS\",\"content\":\"\\u003cul\\u003e\\n \\u003cli\\u003eApplications of artificial intelligence hold the potential to transform eye care, but its integration in current practice remains limited.\\u003c/li\\u003e\\n \\u003cli\\u003eStakeholder analysis reveals the perceived barriers to implementation of artificial intelligence in digital diagnostics across groups of developers, clinicians, patients, and healthcare leaders in eye care.\\u003c/li\\u003e\\n \\u003cli\\u003eWhile there are some shared barriers, a fundamental misalignment in perspectives highlights the importance of co-design in guiding successful artificial intelligence implementation into the future.\\u003c/li\\u003e\\n\\u003c/ul\\u003e\"},{\"header\":\"INTRODUCTION\",\"content\":\"\\u003cp\\u003eArtificial intelligence (AI) is transforming everyday lives with its ability to simulate human intelligent behaviour (\\u003cspan citationid=\\\"CR1\\\" class=\\\"CitationRef\\\"\\u003e1\\u003c/span\\u003e) across a wide range of applications (\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e). It is becoming an increasingly popular tool utilised across various fields, including healthcare (\\u003cspan citationid=\\\"CR3\\\" class=\\\"CitationRef\\\"\\u003e3\\u003c/span\\u003e), where it can improve diagnostic accuracy and quality of care (\\u003cspan citationid=\\\"CR4\\\" class=\\\"CitationRef\\\"\\u003e4\\u003c/span\\u003e). As AI can enhance clinical decision-making (\\u003cspan citationid=\\\"CR5\\\" class=\\\"CitationRef\\\"\\u003e5\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR6\\\" class=\\\"CitationRef\\\"\\u003e6\\u003c/span\\u003e), there has been a growing interest in augmenting its use for digital diagnosis and improving patient outcomes (\\u003cspan citationid=\\\"CR4\\\" class=\\\"CitationRef\\\"\\u003e4\\u003c/span\\u003e) across disciplines.\\u003c/p\\u003e\\u003cp\\u003eIn eye care, AI can aid in preventing avoidable blindness and vision impairment. One key example of its potential to revolutionise current clinical practice is in diagnosing and managing age-related macular degeneration (AMD). AMD is a leading cause of vision loss with a global prevalence of 8.7 per cent (\\u003cspan citationid=\\\"CR7\\\" class=\\\"CitationRef\\\"\\u003e7\\u003c/span\\u003e), and is a growing public health concern as it is frequently underdiagnosed and misdiagnosed in several healthcare settings(\\u003cspan citationid=\\\"CR8\\\" class=\\\"CitationRef\\\"\\u003e8\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR9\\\" class=\\\"CitationRef\\\"\\u003e9\\u003c/span\\u003e). This leads to delayed management and vision-threatening consequences, as the progression of AMD is characterised by central vision loss. Furthermore, AMD has two presentations known as \\u0026lsquo;dry\\u0026rsquo; (non-neovascular) or \\u0026lsquo;wet\\u0026rsquo; (neovascular) AMD, which is important to differentiate to provide appropriate management for the patient. AI can be applied to any stage of the patient journey, from improving early detection and diagnosis, to management and triaging for treatment (\\u003cspan citationid=\\\"CR10\\\" class=\\\"CitationRef\\\"\\u003e10\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR11\\\" class=\\\"CitationRef\\\"\\u003e11\\u003c/span\\u003e).\\u003c/p\\u003e\\u003cp\\u003eHowever, it is evident from the literature, that despite the significant potential of AI to transform eyecare, its integration remains limited across healthcare (\\u003cspan citationid=\\\"CR12\\\" class=\\\"CitationRef\\\"\\u003e12\\u003c/span\\u003e). Although there are known challenges to the general implementation of AI including cost, liability and data privacy (\\u003cspan additionalcitationids=\\\"CR14\\\" citationid=\\\"CR13\\\" class=\\\"CitationRef\\\"\\u003e13\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e), the real-world barriers to its application in digital diagnostic tools in eye care, specifically AMD, have not been explored.\\u003c/p\\u003e\\u003cp\\u003eThis study used qualitative interviews with key stakeholders to explore the barriers to adopting AI in macular disease and the factors contributing to its lack of clinical utilisation in diagnosis. Recognising a co-design approach is fundamental to ensuring both the success and broad uptake of these technologies, clinicians, patients, healthcare leaders and technology developers were surveyed to capture their diverse perspectives. For analysis, the Updated CFIR (\\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e) \\u0026ndash; a well-established determinant framework in implementation science to evaluate contextual factors influencing real-world outcomes, and overall implementation effectiveness, was used. The purpose was to identify key factors that could hinder the effective implementation of AI-driven diagnosis in eye care.\\u003c/p\\u003e\"},{\"header\":\"METHODS\",\"content\":\"\\u003cdiv id=\\\"Sec3\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eStudy Design\\u003c/h2\\u003e\\u003cp\\u003eA qualitative study based on data from a series of semi-structured interviews with key stakeholders (including clinicians, developers, leaders, and patients) was conducted by researchers at the UNSW Sydney School of Optometry and Vision Science and School of Business (Supplementary file 1).\\u003c/p\\u003e\\u003c/div\\u003e\\n\\u003ch3\\u003ePARTICIPANT SELECTION\\u003c/h3\\u003e\\n\\u003cdiv id=\\\"Sec5\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eSampling\\u003c/h2\\u003e\\u003cp\\u003ePurposive sampling was used to identify key individuals from the professional contact list of authors to have a diverse representation across stakeholder groups. All participants were 18 years old or older and had an interest and/or experience in macular disease.\\u003c/p\\u003e\\u003c/div\\u003e\\n\\u003ch3\\u003eMethod of Approach (Recruitment) and Sample Size\\u003c/h3\\u003e\\n\\u003cp\\u003eFrom September 2022 to March 2023, 147 potential participants were contacted by email. Participation was voluntary, and no financial incentives were offered for the completion of the study.\\u003c/p\\u003e\\u003cp\\u003eInformed consent\\u003c/strong\\u003e was obtained from all participants in accordance with the Declaration of Helsinki and approved by the University of New South Wales, Sydney Human Research Ethics Advisory Committee (HC210986; February 2022).\\u003c/p\\u003e\\u003c/p\\u003e\\n\\u003ch3\\u003eDATA COLLECTION\\u003c/h3\\u003e\\n\\u003cp\\u003eDemographic data including age, sex, and occupation was obtained using an online questionnaire (Qualtrics, Utah, USA). All interviews were conducted in English via Zoom (Zoom Video Communications Inc, California, USA) by authors AL and SH (\\u003cspan citationid=\\\"CR17\\\" class=\\\"CitationRef\\\"\\u003e17\\u003c/span\\u003e).\\u003c/p\\u003e\\u003cdiv id=\\\"Sec8\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eInterview Guide\\u003c/h2\\u003e\\u003cp\\u003eA semi-structured interview guide was developed by research investigators following a literature review to explore key questions for each stakeholder group (\\u003cspan citationid=\\\"CR17\\\" class=\\\"CitationRef\\\"\\u003e17\\u003c/span\\u003e). The questions were designed to encourage open dialogue of AI and diagnostic technologies for AMD in Australia and were structured in sections: 1) experiences, 2) attitudes, 3) enablers, 4) barriers, 5) possible futures (anticipated personal and business use). Interviewers were encouraged to re-word, re-order or clarify the questions to facilitate the discussions.\\u003c/p\\u003e\\u003cp\\u003eParticipants were provided with a copy of the guide before their interview and were asked to only complete the interview once. All participants were informed that the purpose of the interview was to hear their experiences, views and attitudes toward digital diagnosis for macular disease, especially neovascular AMD.\\u003c/p\\u003e\\u003c/div\\u003e\\n\\u003ch3\\u003eAudio Recording and Transcription\\u003c/h3\\u003e\\n\\u003cp\\u003eAll interviews were audio-recorded, transcribed verbatim and analysed using a theory-driven approach based on the Updated CFIR (\\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e) in NVivo (Lumivero, Colorado, USA). The updated CFIR enabled systematic analysis of interview data and assessment of contextual factors influencing real-world implementation.\\u003c/p\\u003e\\n\\u003ch3\\u003eDefinitions\\u003c/h3\\u003e\\n\\u003cp\\u003eDefinitions have been adapted from the Updated CFIR (\\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e), with the elimination of sub-domains that were not identified (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e).\\u003c/p\\u003e\\u003cp\\u003e\\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab1\\\" border=\\\"1\\\"\\u003e\\u003ccaption language=\\\"En\\\"\\u003e\\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 1\\u003c/div\\u003e\\u003cdiv class=\\\"CaptionContent\\\"\\u003e\\u003cp\\u003eAdapted CFIR and definitions\\u003c/p\\u003e\\u003c/div\\u003e\\u003c/caption\\u003e\\u003ccolgroup cols=\\\"2\\\"\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e\\u003cthead\\u003e\\u003ctr\\u003e\\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eI. INNOVATION DOMAIN\\u003c/p\\u003e\\u003cp\\u003eThe \\u0026ldquo;thing\\u0026rdquo; being implemented\\u003c/p\\u003e\\u003cp\\u003e\\u003cem\\u003eProject Innovation: Artificial intelligence and related technologies for the digital diagnosis of AMD, especially neovascular AMD; which remain in practice when implementation is complete and is distinct from the implementation process and strategies used to implement the technology.\\u003c/em\\u003e\\u003c/p\\u003e\\u003c/th\\u003e\\u003c/tr\\u003e\\u003c/thead\\u003e\\u003ctbody\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Name\\u003c/b\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Definition\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003e\\u003cem\\u003eThe degree to which\\u003c/em\\u003e:\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eA. Innovation Source\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe group that developed and/or visibly sponsored use of the innovation is reputable, credible, and/or trustable.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eB. Innovation Evidence-Based\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe innovation has robust evidence supporting its effectiveness.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eC. Innovation Relative Advantage\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe innovation is better than other available innovations or current practice i.e. better than a human clinician\\u0026rsquo;s ability to diagnose AMD\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eD. Innovation Adaptability\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe innovation can be modified, tailored, or refined to fit local context or needs.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eE. Innovation Complexity\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe innovation is complicated, which may be reflected by its scope and/or the nature and number of connections and steps.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eF. Innovation Cost\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe innovation purchase and operating costs are affordable.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eII. \\u003cb\\u003eOUTER SETTING DOMAIN\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003eThe setting in which the Inner Setting exists. There may be multiple Outer Settings and/or multiple levels within the Outer Setting.\\u003c/p\\u003e\\u003cp\\u003e\\u0026nbsp;\\u003cb\\u003eProject Outer Setting(s)\\u003c/b\\u003e: \\u003cem\\u003eCommunity and local health district systems (public and private), state-wide health systems, national health schemes (within Australia), and other companies including affiliations, organisations and referral networks. Boundary between the outer and inner setting is the physical premises of healthcare clinics, bounded by brick and mortar.\\u003c/em\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Name\\u003c/b\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Definition\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003e\\u003cem\\u003eThe degree to which\\u003c/em\\u003e:\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eA. Critical Incidents\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eLarge-scale and/or unanticipated events disrupt implementation and/or delivery of the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eB. Local Attitudes\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eSociocultural values (e.g., shared responsibility in helping recipients) and beliefs (e.g., convictions about the worthiness of recipients) encourage the Outer Setting to support implementation and/or delivery of the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eC. Partnerships \\u0026amp; Connections\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eNetworks with external entities, including referral networks, academic affiliations, and professional organization networks.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eD. Financing\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eFunding from external entities (e.g., grants, reimbursement) is available to implement and/or deliver the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eE. External Pressure\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eExternal pressures drive implementation and/or delivery of the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eIII. \\u003cb\\u003eINNER SETTING DOMAIN\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003eThe setting in which the innovation is implemented. There may be multiple Inner Settings and/or multiple levels within the Inner Setting.\\u003c/p\\u003e\\u003cp\\u003e\\u003cb\\u003eProject Inner Setting(s)\\u003c/b\\u003e: \\u003cem\\u003eSingle or multi-site clinics united by a single brand.\\u003c/em\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Name\\u003c/b\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Definition\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003e\\u003cem\\u003eThe degree to which\\u003c/em\\u003e:\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eA. Structural Characteristics \\u0026ndash; Information Technology Infrastructure\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eTechnological systems for tele-communication, electronic documentation, and data storage, management, reporting, and analysis support functional performance of the Inner Setting.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eB. Culture\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThere are shared values, beliefs, and norms across the Inner Setting.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eC. Access to Knowledge \\u0026amp; Information\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eGuidance and/or training is accessible to implement and deliver the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eIV. \\u003cb\\u003eINDIVIDUALS DOMAIN\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003eThe roles and characteristics of individuals.\\u003c/p\\u003e\\u003cp\\u003e\\u003cb\\u003eProject Role(s)\\u003c/b\\u003e: \\u003cem\\u003eInnovation deliverers and recipients.\\u003c/em\\u003e\\u003c/p\\u003e\\u003cp\\u003e\\u003cb\\u003eProject Characteristic(s)\\u003c/b\\u003e: \\u003cem\\u003eNeed, capability and motivation of individuals\\u0026rsquo; roles.\\u003c/em\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Name \\u0026ndash; Roles\\u003c/b\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Definition\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003e\\u003cem\\u003eThe degree to which\\u003c/em\\u003e:\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eA. Innovation Deliverers\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eIndividuals who are directly or indirectly delivering the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eB. Innovation Recipients\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eIndividuals who are directly or indirectly receiving the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Name - Characteristics\\u003c/b\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Definition\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003e\\u003cem\\u003eThe degree to which\\u003c/em\\u003e:\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e1. Need\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe individual(s) has deficits related to survival, well-being, or personal fulfillment, which will be addressed by implementation and/or delivery of the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e2. Capability\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe individual(s) has interpersonal competence, knowledge, and skills to fulfill Role.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e3. Motivation\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eThe individual(s) is committed to fulfilling Role.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eV. \\u003cb\\u003eIMPLEMENTAION PROCCESS DOMAIN\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003eThe activities and strategies used to implement the innovation.\\u0026nbsp;\\u003c/p\\u003e\\u003cp\\u003e\\u0026nbsp;\\u003cb\\u003eProject Implementation Process\\u003c/b\\u003e: \\u003cem\\u003eDynamic Sustainability Framework; activities and strategies.\\u003c/em\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Name \\u0026ndash; Roles\\u003c/b\\u003e\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003e\\u003cb\\u003eConstruct Definition\\u003c/b\\u003e\\u003c/p\\u003e\\u003cp\\u003e\\u003cem\\u003eThe degree to which\\u003c/em\\u003e:\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eA. Assessing Needs \\u0026ndash;\\u003c/p\\u003e\\u003cp\\u003eInnovation Recipients\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eCollects relevant? information about the priorities, preferences, and needs of recipients to guide implementation and delivery of the innovation.\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003c/tbody\\u003e\\u003c/colgroup\\u003e\\u003c/table\\u003e\\u003c/div\\u003e\\u003c/p\\u003e\\u003cdiv id=\\\"Sec11\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eDATA ANALYSIS\\u003c/h2\\u003e\\u003cdiv id=\\\"Sec12\\\" class=\\\"Section3\\\"\\u003e\\u003ch2\\u003eData Coding\\u003c/h2\\u003e\\u003cp\\u003eAL extracted participants\\u0026rsquo; responses to the question \\u003cem\\u003e\\u0026lsquo;What is the biggest barrier to digital diagnosis or AI for macular disease in Australia?\\u003c/em\\u003e\\u0026rsquo;. Other relevant remarks including the keyword \\u0026lsquo;barrier\\u0026rsquo; were also extracted for the analysis. Two authors AL and JN annotated the extracted text segments of the transcripts with the specific barrier to AI adoption using an inductive approach, with no predefined barrier list.\\u003c/p\\u003e\\u003cp\\u003eAlthough each participant was encouraged to describe the single biggest barrier to implementation, if more than one was mentioned, it was counted separately. If the same barrier was mentioned more than once by a participant during their interview, it was only counted once.\\u003c/p\\u003e\\u003cp\\u003eOnce all the barriers had been summarised, iterative meetings were performed to go through all the coded texts and crosscheck the results. Agreement was reached on 65 of all 72 identified barriers, resulting in inter-coder reliability of 90%. Disagreements were discussed by AL and JN until an agreement was made on the optimal concept to code a specific fragment of the text by checking the full transcript context and re-reading the section of interest. Where ambiguity remained, the CFIR framework definitions were consulted to guide coding decisions. The usual sources of disagreement were the scope of one concept (3/7) and the specific barriers and facilitators that one concept should encompass (4/7). In the latter cases, a single concept was split into two separate concepts until agreement was achieved.\\u003c/p\\u003e\\u003cp\\u003eThe derived index of specific barriers was then mapped to the constructs of the Updated CFIR. Any disagreements about which Updated CFIR construct was the most appropriate for the concept were resolved by discussing the possible options until an agreement was reached. All authors verified that the domains and sub-domains appropriately described the extracted data.\\u003c/p\\u003e\\u003c/div\\u003e\\u003c/div\\u003e\\u003cdiv id=\\\"Sec13\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eAssessment of Key Barriers\\u003c/h2\\u003e\\u003cp\\u003eTo quantify the relative importance of the barrier to implementation, the frequency of each cited barrier coded under each domain and sub-domain across stakeholders was compared (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003e).\\u003c/p\\u003e\\u003c/div\\u003e\"},{\"header\":\"RESULTS\",\"content\":\"\\u003cp\\u003eA total of 37 individuals across 4 stakeholder groups participated in the study. Of the participants, there were 12 clinicians (32%), 10 healthcare leaders including senior health managers and professional organisation representatives (27%), 8 patients with AMD (22%), and 7 developers including researchers or technologists (19%). Stakeholder groups were labelled as clinicians (C), developers (D), leaders (L), and patients (P).\\u003c/p\\u003e\\n\\u003cp\\u003eDespite the interview guide stating that the purpose is to explore AI and related technologies for the digital diagnosis of macular disease, specifically neovascular AMD; participants were more focused on the innovation aspect (i.e. AI) during the interview. The ensuing discussions were not limited to macular disease but covered the implementation of digital diagnosis in eye health care in general.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eTable 2. Barrier counts\\u0026nbsp;\\u003c/p\\u003e\\n\\u003ctable border=\\\"1\\\" cellspacing=\\\"0\\\" cellpadding=\\\"0\\\" width=\\\"605\\\"\\u003e\\n \\u003ctbody\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 78px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eCFIR Domain\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eCFIR Sub-Domain\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eSpecific Barriers\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eClinicians\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003en=12\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eDevelopers\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003en=7\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eLeaders\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003en=10\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003ePatients\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003en=8\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eTotals\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003en=37\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"6\\\" style=\\\"width: 78px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eInnovation\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eInnovation Source\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eLimited \\u003cstrong\\u003ecredibility\\u003c/strong\\u003e of the innovation source\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e1\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eInnovation Evidence-Base\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eLack of evidence and uncertainty whether sufficient resources are available to prove \\u003cstrong\\u003eclinical usefulness\\u003c/strong\\u003e.\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e5\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eInnovation Relative Advantage\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eGeneralisability\\u003c/strong\\u003e: The innovation seemed inaccurate when applied locally and does not improve \\u003cstrong\\u003eclinical workflow.\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e4\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e5\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eInnovation Adaptability\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eInteroperability\\u003c/strong\\u003e: Limited capacity to work with different imaging instruments, algorithms, and software.\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e4\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eInnovation Complexity\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eNature of ocular imaging being \\u003cstrong\\u003edifficult to acquire\\u003c/strong\\u003e clear images\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e1\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eInnovation Cost\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003ePurchase and operating \\u003cstrong\\u003ecosts\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e5\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e7\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"5\\\" style=\\\"width: 305px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e13\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e7\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e3\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e23\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"5\\\" style=\\\"width: 78px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eOuter Setting\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eCritical Incidents\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eHackers\\u003c/strong\\u003e obtaining personal health information\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e1\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eLocal Attitudes\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eUncertainty about whether it would be possible to maintain \\u003cstrong\\u003eequity\\u003c/strong\\u003e across different organisations and patient populations\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e3\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e4\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003ePartnerships \\u0026amp; Connections\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eLack of \\u003cstrong\\u003epartnerships and connections\\u003c/strong\\u003e to support utilisation\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e1\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eFinancing\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eUncertainty about \\u003cstrong\\u003efinancing; role of government subsidy\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e4\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eExternal Pressure\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eMisaligned priorities\\u003c/strong\\u003e of different stakeholders\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e1\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"5\\\" style=\\\"width: 305px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e3\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e5\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e11\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"4\\\" style=\\\"width: 78px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eInner Setting\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" rowspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eStructural Characteristics\\u003c/p\\u003e\\n \\u003cp\\u003e\\u0026ndash; Information Technology Infrastructure\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eLack of infrastructure to ensure appropriate \\u003cstrong\\u003edata security and privacy\\u003c/strong\\u003e; fear of data misuse\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e3\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eLarge \\u003cstrong\\u003edata storage\\u0026nbsp;\\u003c/strong\\u003erequirements\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e2\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eCulture\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eResistance to change\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e5\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eAccess To Knowledge \\u0026amp; Information\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eOptometrists\\u0026rsquo; \\u003cstrong\\u003elack of training\\u0026nbsp;\\u003c/strong\\u003ein being able to use the AI recommendations\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e4\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"5\\\" style=\\\"width: 305px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e7\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e4\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e14\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"6\\\" style=\\\"width: 78px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eIndividuals\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd rowspan=\\\"4\\\" style=\\\"width: 66px;\\\"\\u003e\\n \\u003cp\\u003eInnovation Deliverers\\u003c/p\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 66px;\\\"\\u003e\\n \\u003cp\\u003eNeed\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eOptometrists\\u0026rsquo; fear of job loss\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e1\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"2\\\" style=\\\"width: 66px;\\\"\\u003e\\n \\u003cp\\u003eCapability\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eRisk of \\u003cstrong\\u003eover-reliance\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e3\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e5\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eFear that \\u003cstrong\\u003eprofessional judgement of clinicians\\u003c/strong\\u003e will be diminished\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e1\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 66px;\\\"\\u003e\\n \\u003cp\\u003eMotivation\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eLack of trust\\u003c/strong\\u003e in the technology itself\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e3\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"2\\\" style=\\\"width: 66px;\\\"\\u003e\\n \\u003cp\\u003eInnovation Recipients\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 66px;\\\"\\u003e\\n \\u003cp\\u003eNeed\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003ePatient need for \\u003cstrong\\u003ehuman empathy\\u003c/strong\\u003e and oversight\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e4\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e7\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 66px;\\\"\\u003e\\n \\u003cp\\u003eMotivation\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003eLack of individual \\u003cstrong\\u003epatient awareness\\u0026nbsp;\\u003c/strong\\u003eregarding the need for eye care\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e3\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"5\\\" style=\\\"width: 305px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e5\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e3\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e7\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e5\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e21\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 78px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eImplementation Process\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 132px;\\\"\\u003e\\n \\u003cp\\u003eAssessing Needs\\u003cbr\\u003e\\u0026nbsp;\\u0026ndash; Innovation Deliverers\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 94px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eTraining and knowledge\\u003c/strong\\u003e of innovation deliverers, persistent need to retrain\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e3\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"5\\\" style=\\\"width: 305px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e0\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd colspan=\\\"2\\\" style=\\\"width: 60px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e3\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 119px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 76px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 74px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 128px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 76px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 86px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 62px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 62px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 1px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 48px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 0px;\\\"\\u003e\\u003cbr\\u003e\\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003c/tbody\\u003e\\n\\u003c/table\\u003e\\n\\u003cp\\u003eTable 3 summarises the domain count across the stakeholder groups in order of highest to lowest frequency. For clinicians and developers, \\u0026lsquo;innovation\\u0026rsquo; was the most frequently quoted domain. For leaders and patients \\u0026lsquo;individuals\\u0026rsquo; were the most cited; however, leaders focussed on innovation deliverers, whereas patients were more concerned about innovation recipients. Implementation processes were mentioned the least across the stakeholders as an identified barrier.\\u003c/p\\u003e\\n\\u003cp\\u003eTable 3. Key domains identified by stakeholder groups.\\u003c/p\\u003e\\n\\u003ctable border=\\\"1\\\" cellspacing=\\\"0\\\" cellpadding=\\\"0\\\"\\u003e\\n \\u003ctbody\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eClinicians\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eDevelopers\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eLeaders\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003ePatients\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eInnovation (13)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eInnovation (7)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eIndividuals (7)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eIndividuals (5)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eInner setting (7)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd rowspan=\\\"2\\\" style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eInner setting (4), Individuals (4)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eOuter setting (5)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd rowspan=\\\"2\\\" style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eOuter setting (1), Inner setting (1)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eIndividuals (5)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eInnovation (4)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eOuter setting (2)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eOuter setting (3)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd rowspan=\\\"2\\\" style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eInner setting (2), Implementation processes (2)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003eImplementation process (1)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 150px;\\\"\\u003e\\n \\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003c/tbody\\u003e\\n\\u003c/table\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003e1. \\u0026nbsp;Clinicians\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe costs of purchasing and operating an innovation was expressed by many clinicians as a key barrier to its implementation (Figure 1).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;The only barrier for practitioners to acquire it is cost\\u0026hellip;The only barrier for every practice to have it\\u0026hellip;is cost\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; C4\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;To be honest, the biggest factor\\u0026hellip;is cost\\u0026hellip;assuming it works reasonably well, I think, it\\u0026rsquo;s mainly a cost thing.\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; C12\\u003c/p\\u003e\\n\\u003cp\\u003eClinicians also cited concerns of generalisability and the innovations\\u0026rsquo; lack of relative advantage over their own ability to make accurate diagnoses (Figure 1).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;If the AI incorrectly does, like pick up an artefact\\u0026hellip;an error in judgement can be made\\u0026hellip;it highlights that there\\u0026rsquo;s something severely wrong and they\\u0026rsquo;re unnecessarily referred [false positive] or the opposite where it misses something, and they\\u0026rsquo;re told\\u0026hellip;there\\u0026rsquo;s nothing wrong [false negative].\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; C8\\u003c/p\\u003e\\n\\u003cp\\u003eThe structural limitations within their clinical practices were addressed (Figure 1), with the lack of information technology infrastructure to support the functionality of the innovation, including data storage and cyber security (digital health privacy).\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;We are still some ways away from having the infrastructure to do so\\u0026hellip;hackers managed to obtain personal information\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; C1\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;Other barriers\\u0026hellip;data storage, privacy issues\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; C12\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003e2. \\u0026nbsp;Developers\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eDevelopers identified similar barriers to clinicians correlating to the innovation domain.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u0026ldquo;\\u003cem\\u003eI think people have issues with AI and decision support\\u0026hellip;how can I trust that system has done it the same way that I would have as a clinician?\\u003c/em\\u003e\\u0026rdquo; \\u0026ndash; D6\\u003c/p\\u003e\\n\\u003cp\\u003e\\u0026ldquo;\\u003cem\\u003eIt comes back to knowledge, right?...Like how do they know this is legit? That whatever technology that is put forward to me is safe, is accurate, is relevant?\\u0026rdquo;\\u0026nbsp;\\u003c/em\\u003e\\u0026ndash; D5\\u003c/p\\u003e\\n\\u003cp\\u003eHowever, their views were more generalised across subdomains (Figure 1); with the only barrier with repeated citations being the lack of evidence for the innovation and hence the uncertainty of its clinical usefulness in the real world.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;I think it\\u0026apos;s definitely\\u0026hellip;a near term issue in terms of getting enough data to make that really work\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; D1\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;So that\\u0026apos;s a real barrier for us, conducting the research itself in a setting\\u0026hellip;out into the world to do, you know, pragmatic studies, but that\\u0026apos;s not currently what we\\u0026apos;re doing yet.\\u0026rdquo;\\u0026nbsp;\\u003c/em\\u003e\\u0026ndash; D2\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003e3. \\u0026nbsp;Leaders\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eLeaders on the other hand focussed more on individuals (innovation deliverers) (Figure 1), highlighting the potential risk of over-reliance on the innovation and the consequent deskilling of clinicians.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;AI lends itself to sort of protocol-based care\\u0026hellip;you feed this information in, you get a green light or a red light\\u0026hellip;but I think the risk is over-reliance on that\\u0026hellip;people just sort of blindly trusting it \\u0026ldquo;\\u003c/em\\u003e \\u0026ndash; L1\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;People get\\u0026hellip;a bit blas\\u0026eacute; about re-checking because it\\u0026rsquo;s quite good for a while and then there\\u0026rsquo;s a false negative and they\\u0026rsquo;re not doing the adequate other tests\\u0026hellip;but it doesn\\u0026rsquo;t have the same sensitivity or specificity\\u0026rdquo;.\\u003c/em\\u003e \\u0026ndash; L3\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;It\\u0026rsquo;s an algorithm, it\\u0026rsquo;s a model\\u0026hellip;.but people in the system have got to remember that that\\u0026rsquo;s what it is. It\\u0026rsquo;s a tool. It\\u0026rsquo;s an aid\\u0026hellip;So people have to be aware that you can\\u0026rsquo;t over-rely on a model\\u0026hellip;Practitioners need to appreciate both the strength of an AI diagnosis as well as the shortcomings and where the tool might be at its weakest. So I think that worries me because people have a tendency to over-rely on technology these days.\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; L5\\u003c/p\\u003e\\n\\u003cp\\u003eAnother commonly expressed barrier was the uncertainty of being able to maintain equity across different organisations and patient populations (Figure 1).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;It\\u0026rsquo;s also about how you roll out the accessibility and availability of these tools to the broader community\\u0026rdquo;\\u0026nbsp;\\u003c/em\\u003e\\u0026ndash; L5\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;One of the problems across the health area is consistency. And so you want everybody to get the same level of care\\u0026hellip;You\\u0026rsquo;re building up a new silo with people who are for digital health\\u0026hellip; and people who are not so knowledgeable\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; L7\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003e4. \\u0026nbsp;Patients\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003ePatients posed the most individual-centric outlook (innovation recipients) (Figure 1), with the key barrier being the need and the lack thereof human empathy with the implementation AI.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;I think there\\u0026rsquo;s sort of an added element in the personal connection\\u0026hellip;the patient if they feel safe and secure, is much more able to speak about their concerns\\u0026hellip;.if you just leave them up to a computer, the computer can only do what it\\u0026rsquo;s set up to do.\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; P5\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;I\\u0026rsquo;m reassured by the fact there is human intervention.\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; P6\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cem\\u003e\\u0026ldquo;I like the personal touch\\u0026hellip;maybe I\\u0026rsquo;m a bit old fashioned still\\u0026hellip;whereby I\\u0026rsquo;d like to actually be able to ask the person whatever questions\\u0026hellip;it\\u0026rsquo;s about having the person in the room with you\\u0026rdquo;\\u003c/em\\u003e \\u0026ndash; P8\\u003c/p\\u003e\\n\\u003cp\\u003eSupplementary file 2 summarises all the other barriers.\\u003c/p\\u003e\"},{\"header\":\"DISCUSSION\",\"content\":\"\\u003cp\\u003eAlthough there has been a revolution of AI applications, it is well-recognised that its integration remains limited in healthcare (\\u003cspan citationid=\\\"CR12\\\" class=\\\"CitationRef\\\"\\u003e12\\u003c/span\\u003e). Our study addressed this by exploring barriers to the implementation of AI for digital diagnosis in eye care and found that clinicians are concerned about the cost and financing of AI, developers face challenges proving its usefulness in the clinical workflow, leaders foreboded the consequence of clinicians deskilling due to potential over-reliance on AI and patients had apprehension for the absence of human interaction.\\u003c/p\\u003e\\u003cp\\u003eMost importantly, our results highlight the differing focus of each stakeholder group and their perceived barriers, whereby the misalignment of perspectives itself can be an inherent barrier to the implementation of the innovation. Barriers to AI in healthcare are typically explored from the perspective of clinicians (\\u003cspan citationid=\\\"CR19\\\" class=\\\"CitationRef\\\"\\u003e19\\u003c/span\\u003e), leaders (\\u003cspan citationid=\\\"CR20\\\" class=\\\"CitationRef\\\"\\u003e20\\u003c/span\\u003e) and patients (\\u003cspan citationid=\\\"CR21\\\" class=\\\"CitationRef\\\"\\u003e21\\u003c/span\\u003e) or an incomplete combination of stakeholders (\\u003cspan citationid=\\\"CR22\\\" class=\\\"CitationRef\\\"\\u003e22\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR23\\\" class=\\\"CitationRef\\\"\\u003e23\\u003c/span\\u003e). To our knowledge, there is currently no framework-based analysis of all the stakeholders\\u0026rsquo; perspectives. Stakeholder co-design is crucial from the initial development phase (\\u003cspan citationid=\\\"CR24\\\" class=\\\"CitationRef\\\"\\u003e24\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR25\\\" class=\\\"CitationRef\\\"\\u003e25\\u003c/span\\u003e) of digital health innovations to successfully deliver patient-centred care (\\u003cspan citationid=\\\"CR25\\\" class=\\\"CitationRef\\\"\\u003e25\\u003c/span\\u003e). Thus, a synthesised evaluation (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab4\\\" class=\\\"InternalRef\\\"\\u003e4\\u003c/span\\u003e) combining multiple stakeholder perspectives is fundamental to achieving a more holistic understanding of barriers to AI in digital diagnosis.\\u003c/p\\u003e\\u003cp\\u003e\\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab4\\\" border=\\\"1\\\"\\u003e\\u003ccaption language=\\\"En\\\"\\u003e\\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 4\\u003c/div\\u003e\\u003cdiv class=\\\"CaptionContent\\\"\\u003e\\u003cp\\u003eSummary of barriers with a high count* (frequency of 5 or more) across stakeholders. *\\u003cem\\u003eThe count of \\u0026lsquo;Costs\\u0026rsquo; and \\u0026lsquo;Financing\\u0026rsquo; were added as commonly cited together.\\u003c/em\\u003e\\u003c/p\\u003e\\u003c/div\\u003e\\u003c/caption\\u003e\\u003ccolgroup cols=\\\"10\\\"\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c6\\\" colnum=\\\"6\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c7\\\" colnum=\\\"7\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c8\\\" colnum=\\\"8\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c9\\\" colnum=\\\"9\\\"\\u003e\\u003c/div\\u003e\\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c10\\\" colnum=\\\"10\\\"\\u003e\\u003c/div\\u003e\\u003cthead\\u003e\\u003ctr\\u003e\\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u0026nbsp;\\u003c/th\\u003e\\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c4\\\" namest=\\\"c3\\\"\\u003e\\u003cp\\u003eClinicians\\u003c/p\\u003e\\u003c/th\\u003e\\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c6\\\" namest=\\\"c5\\\"\\u003e\\u003cp\\u003eDevelopers\\u003c/p\\u003e\\u003c/th\\u003e\\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c8\\\" namest=\\\"c7\\\"\\u003e\\u003cp\\u003eLeaders\\u003c/p\\u003e\\u003c/th\\u003e\\u003cth align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c10\\\" namest=\\\"c9\\\"\\u003e\\u003cp\\u003ePatients\\u003c/p\\u003e\\u003c/th\\u003e\\u003c/tr\\u003e\\u003c/thead\\u003e\\u003ctbody\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u003cp\\u003eCosts\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c2\\\"\\u003e\\u003cp\\u003eFinancing\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c3\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c4\\\"\\u003e\\u0026nbsp;\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c5\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c6\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c8\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c9\\\"\\u003e\\u0026nbsp;\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colname=\\\"c10\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eLack of Clinical Usefulness\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c4\\\" namest=\\\"c3\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c6\\\" namest=\\\"c5\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c8\\\" namest=\\\"c7\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c10\\\" namest=\\\"c9\\\"\\u003e\\u0026nbsp;\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eNot Improving Workflow\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c4\\\" namest=\\\"c3\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c6\\\" namest=\\\"c5\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c8\\\" namest=\\\"c7\\\"\\u003e\\u0026nbsp;\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c10\\\" namest=\\\"c9\\\"\\u003e\\u0026nbsp;\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eDeskilling of Clinicians\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c4\\\" namest=\\\"c3\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c6\\\" namest=\\\"c5\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c8\\\" namest=\\\"c7\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c10\\\" namest=\\\"c9\\\"\\u003e\\u0026nbsp;\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eLack of Human Interaction for Patients\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c4\\\" namest=\\\"c3\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c6\\\" namest=\\\"c5\\\"\\u003e\\u0026nbsp;\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c8\\\" namest=\\\"c7\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c10\\\" namest=\\\"c9\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003ctr\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c2\\\" namest=\\\"c1\\\"\\u003e\\u003cp\\u003eResistance to Change\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c4\\\" namest=\\\"c3\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c6\\\" namest=\\\"c5\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c8\\\" namest=\\\"c7\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003ctd align=\\\"left\\\" colspan=\\\"2\\\" nameend=\\\"c10\\\" namest=\\\"c9\\\"\\u003e\\u003cp\\u003e●\\u003c/p\\u003e\\u003c/td\\u003e\\u003c/tr\\u003e\\u003c/tbody\\u003e\\u003c/colgroup\\u003e\\u003c/table\\u003e\\u003c/div\\u003e\\u003c/p\\u003e\\u003cp\\u003eBy default in the design of our study, all the stakeholders had at least a fundamental awareness of AI. However, the level of knowledge in its potential role and use case of digital diagnosis innovations varied across individuals. This further highlighted the importance of co-design, and the need for establishing a common language and understanding of the technological innovation as a foundation for successful implementation.\\u003c/p\\u003e\\u003cdiv id=\\\"Sec16\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eCosts and Financing\\u003c/h2\\u003e\\u003cp\\u003eThe multi-factorial challenge of costs and financing involved in implementing AI digital diagnosis systems was unanimously identified as a key barrier across stakeholder groups, including high initial start-up costs, operating expenses and the uncertainty of financing the implementation of the innovation. These issues are widely cited in literature (\\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e, \\u003cspan additionalcitationids=\\\"CR27\\\" citationid=\\\"CR26\\\" class=\\\"CitationRef\\\"\\u003e26\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR28\\\" class=\\\"CitationRef\\\"\\u003e28\\u003c/span\\u003e), and highlight the importance of favourable financial models and reimbursement policies in the adoption of innovative technologies such as AI, despite its potential benefits. Further exploration of stakeholders\\u0026rsquo; willingness to pay (\\u003cspan citationid=\\\"CR29\\\" class=\\\"CitationRef\\\"\\u003e29\\u003c/span\\u003e) can elicit their perceived value and preferences regarding the implementation of AI and contribute towards building a sustainable financial model to overcome the barrier of costs and financing.\\u003c/p\\u003e\\u003c/div\\u003e\\u003cdiv id=\\\"Sec17\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eLack of Clinical Usefulness and Not Improving Workflow\\u003c/h2\\u003e\\u003cp\\u003eAnother inherent barrier arises from the uncertainty of the usefulness of the AI innovation, and whether it ultimately improves clinical workflow. This barrier combines notions of needing robust evidence supporting AI effectiveness, accuracy, and ability to outperform clinicians, and its lack thereof being a barrier for all innovation deliverers (clinicians, developers, and leaders). There has been a shift in literature, with initial concerns arising from the limited number of randomised controlled trials in the clinical context (\\u003cspan citationid=\\\"CR30\\\" class=\\\"CitationRef\\\"\\u003e30\\u003c/span\\u003e) to its persistent lack of generalisability and practicality (\\u003cspan citationid=\\\"CR31\\\" class=\\\"CitationRef\\\"\\u003e31\\u003c/span\\u003e) at present, despite there being an expanding interest in AI. The need for more comprehensive research to minimise publication bias and to prioritise patient-relevant outcomes when evaluating AI in clinical practice has been recognised (\\u003cspan citationid=\\\"CR31\\\" class=\\\"CitationRef\\\"\\u003e31\\u003c/span\\u003e). Identifying clinical \\u0026ldquo;pain points\\u0026rdquo; for more solution-driven AI innovations and enhancing collaborations between stakeholders (\\u003cspan citationid=\\\"CR30\\\" class=\\\"CitationRef\\\"\\u003e30\\u003c/span\\u003e) are also acknowledged as potential solutions to improving its usefulness and the overall workflow.\\u003c/p\\u003e\\u003c/div\\u003e\\u003cdiv id=\\\"Sec18\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eDeskilling of Clinicians (Innovation Deliverers)\\u003c/h2\\u003e\\u003cp\\u003eAll innovation deliverers expressed the potential risk of over-reliance by clinicians as a barrier as well as the consequence of deskilling; compromising one\\u0026rsquo;s competence, knowledge, and skills. Over-reliance is identified as a precursor to deskilling (\\u003cspan citationid=\\\"CR26\\\" class=\\\"CitationRef\\\"\\u003e26\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR32\\\" class=\\\"CitationRef\\\"\\u003e32\\u003c/span\\u003e), with concerns at present regarding increased dependence on the capabilities of automation (\\u003cspan citationid=\\\"CR26\\\" class=\\\"CitationRef\\\"\\u003e26\\u003c/span\\u003e). Deskilling in the clinical context can be defined as a \\u003cem\\u003e\\u0026ldquo;situation where clinicians experience reduced opportunities to exercise particular skills\\u0026rdquo;\\u003c/em\\u003e (\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e) as an unintended consequence of AI implementation in the medical profession. This is considered an inevitable process in improving efficiency and reducing costs, as exemplified by the now widespread adoption of electronic health records and clinical practice guidelines (\\u003cspan citationid=\\\"CR34\\\" class=\\\"CitationRef\\\"\\u003e34\\u003c/span\\u003e). The alternative phenomenon of upskilling or reskilling (\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR35\\\" class=\\\"CitationRef\\\"\\u003e35\\u003c/span\\u003e) involving enhancing one\\u0026rsquo;s existing skill set or learning new skills, respectively was not discussed by any of the stakeholders. However, it is to acknowledge that there is a growing consensus that \\u0026ldquo;AI is not going to replace humans, but humans with AI are going to replace humans without AI\\u0026rdquo;(\\u003cspan citationid=\\\"CR36\\\" class=\\\"CitationRef\\\"\\u003e36\\u003c/span\\u003e). Thus, focusing on the positive potentials of AI, reskilling efforts should centre on preparing eye care professionals to effectively collaborate with AI systems (\\u003cspan citationid=\\\"CR35\\\" class=\\\"CitationRef\\\"\\u003e35\\u003c/span\\u003e) whilst maintaining the clinician\\u0026rsquo;s critical role in patient outcomes.\\u003c/p\\u003e\\u003c/div\\u003e\\u003cdiv id=\\\"Sec19\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eLack of Human Interaction for Patients (Innovation Recipients)\\u003c/h2\\u003e\\u003cp\\u003eMoreover, the lack of human interaction for patients was identified as a key barrier by both innovation deliverers and recipients, but most commonly by the patients. This highlights the significance of the patient-clinician relationship whereby building rapport is crucial for the patient to entrust the clinician to act on behalf of their health and have their best interests at heart (\\u003cspan citationid=\\\"CR33\\\" class=\\\"CitationRef\\\"\\u003e33\\u003c/span\\u003e). Thus, this barrier arises from the fear that healthcare is shifting away from patient interactions and towards data analytics with a loss of the holistic approach (\\u003cspan citationid=\\\"CR32\\\" class=\\\"CitationRef\\\"\\u003e32\\u003c/span\\u003e). It is commonly cited in literature that patients prefer human interaction over automation (\\u003cspan citationid=\\\"CR32\\\" class=\\\"CitationRef\\\"\\u003e32\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR37\\\" class=\\\"CitationRef\\\"\\u003e37\\u003c/span\\u003e), despite the capability of AI to outperform humans (\\u003cspan citationid=\\\"CR38\\\" class=\\\"CitationRef\\\"\\u003e38\\u003c/span\\u003e). The strive for patient-centred care, which involves patients as active participants in their care, is considered an essential part of high-quality healthcare systems (\\u003cspan citationid=\\\"CR39\\\" class=\\\"CitationRef\\\"\\u003e39\\u003c/span\\u003e). Thus co-design of the implementation of AI should be considered such that it can instead foster the connection between clinicians and patients(\\u003cspan citationid=\\\"CR40\\\" class=\\\"CitationRef\\\"\\u003e40\\u003c/span\\u003e).\\u003c/p\\u003e\\u003c/div\\u003e\\u003cdiv id=\\\"Sec20\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eResistance to Change\\u003c/h2\\u003e\\u003cp\\u003eFinally, similar views emerged from all stakeholders with resistance to change, being cited as a barrier by at least one stakeholder in each group (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab4\\\" class=\\\"InternalRef\\\"\\u003e4\\u003c/span\\u003e). Change resistance (\\u003cspan citationid=\\\"CR41\\\" class=\\\"CitationRef\\\"\\u003e41\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR42\\\" class=\\\"CitationRef\\\"\\u003e42\\u003c/span\\u003e) being an intrinsic and complex problem, is amplified by the sociotechnical implications of AI (\\u003cspan citationid=\\\"CR43\\\" class=\\\"CitationRef\\\"\\u003e43\\u003c/span\\u003e); with apprehensions raised across healthcare organisations by clinicians (\\u003cspan citationid=\\\"CR44\\\" class=\\\"CitationRef\\\"\\u003e44\\u003c/span\\u003e), patients (\\u003cspan citationid=\\\"CR45\\\" class=\\\"CitationRef\\\"\\u003e45\\u003c/span\\u003e) and healthcare leaders (\\u003cspan citationid=\\\"CR27\\\" class=\\\"CitationRef\\\"\\u003e27\\u003c/span\\u003e) alike. However, resistance to change can be expressed differently amongst the stakeholders (\\u003cspan citationid=\\\"CR43\\\" class=\\\"CitationRef\\\"\\u003e43\\u003c/span\\u003e). This can take the form of: 1. rejection (disinclination to adopt the innovation, often due to a reluctance to change the status quo), 2. postponement (accepting the innovation in principle but deciding not to adopt it) and/or 3. opposition (actively engaging in strategies to prevent the innovation\\u0026rsquo;s success) (\\u003cspan citationid=\\\"CR46\\\" class=\\\"CitationRef\\\"\\u003e46\\u003c/span\\u003e). Given the intricate nuances of change resistance, successful implementation of innovations requires effective change management strategies (\\u003cspan citationid=\\\"CR47\\\" class=\\\"CitationRef\\\"\\u003e47\\u003c/span\\u003e) whereby leaders play a crucial role (\\u003cspan citationid=\\\"CR48\\\" class=\\\"CitationRef\\\"\\u003e48\\u003c/span\\u003e). Successful navigation of change resistance can be exemplified by the national implementation of eye health electronic patient records in Scotland and Wales, with increasing adoption across England (\\u003cspan additionalcitationids=\\\"CR50\\\" citationid=\\\"CR49\\\" class=\\\"CitationRef\\\"\\u003e49\\u003c/span\\u003e\\u0026ndash;\\u003cspan citationid=\\\"CR51\\\" class=\\\"CitationRef\\\"\\u003e51\\u003c/span\\u003e).\\u003c/p\\u003e\\u003c/div\\u003e\\u003cdiv id=\\\"Sec21\\\" class=\\\"Section2\\\"\\u003e\\u003ch2\\u003eStrengths and Limitations\\u003c/h2\\u003e\\u003cp\\u003e The main strength of our analysis lies in using purposive sampling to identify key stakeholders, which enabled us to capture diverse and informed perspectives on the barriers to implementing AI in eye care diagnostics. Selecting individuals who play a critical role in the implementation process ensured the insights gathered are comprehensive and reflective of real-world challenges. While the sample distribution across stakeholder groups was uneven, this was by design and reflects real-world variability in end-users and their associated willingness to participate. The use of purposive sampling however ensured the inclusion of key informants with relevant experience in AI and AMD, which thereby influenced the specific number and role of the participants (\\u003cspan citationid=\\\"CR17\\\" class=\\\"CitationRef\\\"\\u003e17\\u003c/span\\u003e).\\u003c/p\\u003e\\u003cp\\u003eThis study was conducted within the Australian healthcare system and focused on AMD. While the findings offer insights into stakeholder perspectives on AI implementation in this context, the transferability of these insights to other diseases, demographics or international settings may be limited by differences in healthcare infrastructure, funding models, and disease-specific workflows.\\u003c/p\\u003e\\u003cp\\u003e However, the open dialogue format during the semi-structured interviews naturally allowed responses to explore a wider perspective on AI's role in eye care diagnostics. Notions around cost and financing and deskilling of clinicians are likely to be applicable across ophthalmic diseases and digital diagnostics.\\u003c/p\\u003e\\u003cp\\u003eSustainable reimbursement is well-recognised as a pivotal enabler for the adoption of novel technologies within healthcare systems (\\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR52\\\" class=\\\"CitationRef\\\"\\u003e52\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR53\\\" class=\\\"CitationRef\\\"\\u003e53\\u003c/span\\u003e). Thus, it is critical to understand the stakeholders that determine reimbursement and consider which models, including fee-for-service and value-based care, may best address the financial sustainability of AI implementation in eye care (\\u003cspan citationid=\\\"CR52\\\" class=\\\"CitationRef\\\"\\u003e52\\u003c/span\\u003e). Recent evaluations of cost-effectiveness and real-world implementation feasibility highlight the importance of aligning AI integration with performance-based reimbursement frameworks; therefore, advocating for hybrid models that incentivise innovation whilst ensuring fiscal sustainability (\\u003cspan citationid=\\\"CR52\\\" class=\\\"CitationRef\\\"\\u003e52\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR53\\\" class=\\\"CitationRef\\\"\\u003e53\\u003c/span\\u003e).\\u003c/p\\u003e\\u003cp\\u003eAnother similarity that is echoed in literature is the need for clinicians to be AI competent. Beyond the imperative for eye care professionals to reskill to utilise AI systems (\\u003cspan citationid=\\\"CR35\\\" class=\\\"CitationRef\\\"\\u003e35\\u003c/span\\u003e), emerging perspective suggest that AI needs to be integrated at the education level, across different aspects of the clinical curriculum to combine technical skills with ethical reasoning (\\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR54\\\" class=\\\"CitationRef\\\"\\u003e54\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR55\\\" class=\\\"CitationRef\\\"\\u003e55\\u003c/span\\u003e). Thus, current and future clinicians require adequate training to be equipped to navigate the complexities of AI-driven healthcare (\\u003cspan citationid=\\\"CR55\\\" class=\\\"CitationRef\\\"\\u003e55\\u003c/span\\u003e).\\u003c/p\\u003e\\u003cdiv id=\\\"Sec22\\\" class=\\\"Section3\\\"\\u003e\\u003ch2\\u003eRecommendations for Future Directions\\u003c/h2\\u003e\\u003cp\\u003eTo reduce barriers to adoption, financing models could be adapted through national procurement guidelines, targeted grants, and expanded tax incentives, particularly for small and medium-sized enterprises developing AI tools, as recommended in the Australian Alliance for AI in Healthcare\\u0026rsquo;s National Policy Roadmap (\\u003cspan citationid=\\\"CR56\\\" class=\\\"CitationRef\\\"\\u003e56\\u003c/span\\u003e). Successful examples, such as Singapore\\u0026rsquo;s semi-automated diabetic retinopathy screening program, demonstrate how cost savings and clinical efficiency can be achieved (\\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e).\\u003c/p\\u003e\\u003cp\\u003eHowever, cost-effectiveness varies by geography, deployment strategy, and healthcare funding model. High-income countries may benefit more due to higher human grader costs, but studies from lower-income settings show mixed results (\\u003cspan citationid=\\\"CR57\\\" class=\\\"CitationRef\\\"\\u003e57\\u003c/span\\u003e). Fee-for-service, bundled payments and subscription-based deployment strategies are likely to influence uptake. Moreover, tailoring financial strategies to the structure of healthcare systems, whether public, private, or hybrid, can enhance relevance and sustainability. In Australia, corporate-led rollout models have been described as a strong catalyst for uptake, with even a small number of providers able to shift practice norms across the profession (\\u003cspan citationid=\\\"CR17\\\" class=\\\"CitationRef\\\"\\u003e17\\u003c/span\\u003e).\\u003c/p\\u003e\\u003cp\\u003eTo mitigate deskilling and promote upskilling or reskilling, leaders should proactively redesign training frameworks to preserve core clinical competencies while fostering human\\u0026ndash;AI collaboration. Strategies include maintaining hands-on learning opportunities, integrating AI literacy into curricula, and using AI as a tool for feedback and simulation-based training. Successful examples include double-reading workflows where clinicians assess cases independently before reviewing AI outputs, and structured validation sessions that reinforce clinical reasoning (\\u003cspan citationid=\\\"CR58\\\" class=\\\"CitationRef\\\"\\u003e58\\u003c/span\\u003e, \\u003cspan citationid=\\\"CR59\\\" class=\\\"CitationRef\\\"\\u003e59\\u003c/span\\u003e). These approaches align with broader calls for safeguards against skill erosion and support a shift toward hybrid intelligence, where clinicians and AI systems co-evolve to enhance diagnostic quality and professional resilience.\\u003c/p\\u003e\\u003cp\\u003eConcrete examples of successful co-design in healthcare highlight its potential to address both resistance to change and concerns about loss of human connection. A statewide initiative in Australia co-designed AI workflows to analyse patient-reported experience measures, enabling clinicians to integrate narrative feedback into quality improvement cycles while maintaining human oversight and accountability (\\u003cspan citationid=\\\"CR60\\\" class=\\\"CitationRef\\\"\\u003e60\\u003c/span\\u003e). Similarly, the development of a deep learning skin lesion classifier incorporated a multidisciplinary co-design framework to address ethical, legal, and technical risks\\u0026mdash;such as overdiagnosis and bias\\u0026mdash;ensuring the system supported clinician decision-making rather than replacing it (\\u003cspan citationid=\\\"CR60\\\" class=\\\"CitationRef\\\"\\u003e60\\u003c/span\\u003e). These examples demonstrate how co-design can preserve the human elements of care, enhance trust, and tailor AI systems to clinical realities.\\u003c/p\\u003e\\u003c/div\\u003e\\u003c/div\\u003e\"},{\"header\":\"CONCLUSION\",\"content\":\"\\u003cp\\u003eThis framework-based analysis of barriers to implementation of AI in digital diagnosis provides a more complete overview of stakeholder perspectives, highlighting differences in the views of clinicians, healthcare leaders, patients and developers including researchers or technologists. Five key barriers were identified as costs and financing, lack of perceived clinical usefulness or not improving workflow, deskilling of clinicians, lack of human interaction for patients and resistance to change.\\u003c/p\\u003e\"},{\"header\":\"Abbreviations\",\"content\":\"\\u003cdiv class=\\\"DefinitionList\\\"\\u003e\\u003cdiv class=\\\"DefinitionListEntry\\\"\\u003e\\u003cdiv class=\\\"Term\\\"\\u003e\\u003cb\\u003eAI\\u003c/b\\u003e\\u003c/div\\u003e\\u003cdiv class=\\\"Description\\\"\\u003e\\u003cp\\u003eArtificial Intelligence\\u003c/p\\u003e\\u003c/div\\u003e\\u003c/div\\u003e\\u003cdiv class=\\\"DefinitionListEntry\\\"\\u003e\\u003cdiv class=\\\"Term\\\"\\u003e\\u003cb\\u003eAMD\\u003c/b\\u003e\\u003c/div\\u003e\\u003cdiv class=\\\"Description\\\"\\u003e\\u003cp\\u003eAge-related Macular Degeneration\\u003c/p\\u003e\\u003c/div\\u003e\\u003c/div\\u003e\\u003cdiv class=\\\"DefinitionListEntry\\\"\\u003e\\u003cdiv class=\\\"Term\\\"\\u003e\\u003cb\\u003eCFIR\\u003c/b\\u003e\\u003c/div\\u003e\\u003cdiv class=\\\"Description\\\"\\u003e\\u003cp\\u003eConsolidated Framework for Implementation Research\\u003c/p\\u003e\\u003c/div\\u003e\\u003c/div\\u003e\\u003c/div\\u003e\"},{\"header\":\"Declarations\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eEthics approval and consent to participate:\\u003c/strong\\u003e\\u003cp\\u003e Written informed consent was obtained from all participants in accordance with the tenets of the Declaration of Helsinki and approved by the UNSW Sydney Human Research Ethics Advisory Committee (HC210986; February 2022).\\u003c/p\\u003e\\u003c/p\\u003e\\u003cp\\u003e\\u003cstrong\\u003e\\u003cem\\u003eConsent for publication\\u003c/em\\u003e:\\u003c/strong\\u003e\\u003cp\\u003eNot applicable.\\u003c/p\\u003e\\u003c/p\\u003e\\u003ch2\\u003eFunding:\\u003c/h2\\u003e\\u003cp\\u003eThis work was carried out by the UNSW-Roche Digital Diagnostics Project and funded by Roche Products, Australia. The funder had no role in study design, data collection, data analysis, data interpretation, writing of the report, and decision to submit the paper for publication.\\u003c/p\\u003e\\u003ch2\\u003eAuthor Contribution\\u003c/h2\\u003e\\u003cp\\u003eFS and MAW conceived the idea and supervised the study. Interviews were conducted by AL and SH. Data analysis was performed by JN and AL. JN wrote the manuscript aided by AL. All authors read and approved the final manuscript.\\u003c/p\\u003e\\u003ch2\\u003eAcknowledgements:\\u003c/h2\\u003e\\u003cp\\u003eNot applicable.\\u003c/p\\u003e\\u003ch2\\u003eAvailability of data and material:\\u003c/h2\\u003e\\u003cp\\u003eThe data and material generated and analysed during the study are not publicly available to protect participant privacy.\\u003c/p\\u003e\"},{\"header\":\"References\",\"content\":\"\\u003col\\u003e\\u003cli\\u003e\\u003cspan\\u003eOxford English Dictionary. Oxford University Press; 2023. artificial intelligence, n.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eFui-Hoon Nah F, Zheng R, Cai J, Siau K, Chen L. Generative AI and ChatGPT: Applications, challenges, and AI-human collaboration. Journal of Information Technology Case and Application Research. 2023;25(3):277\\u0026ndash;304.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eAlowais SA, Alghamdi SS, Alsuhebany N, Alqahtani T, Alshaya AI, Almohareb SN, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med Educ. 2023;23(1):689.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eOuanes K, Farhah N. Effectiveness of Artificial Intelligence (AI) in Clinical Decision Support Systems and Care Delivery. J Med Syst. 2024;48(1):74.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eSim I, Gorman P, Greenes RA, Haynes RB, Kaplan B, Lehmann H, et al. Clinical decision support systems for the practice of evidence-based medicine. J Am Med Inform Assoc. 2001;8(6):527\\u0026ndash;34.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eSutton RT, Pincock D, Baumgart DC, Sadowski DC, Fedorak RN, Kroeker KI. An overview of clinical decision support systems: benefits, risks, and strategies for success. NPJ Digit Med. 2020;3:17.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eWong WL, Su X, Li X, Cheung CM, Klein R, Cheng CY, et al. Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis. Lancet Glob Health. 2014;2(2):e106-16.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLy A, Nivison-Smith L, Zangerl B, Assaad N, Kalloniatis M. Advanced imaging for the diagnosis of age-related macular degeneration: a case vignettes study. Clin Exp Optom. 2018;101(2):243\\u0026ndash;54.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eNeely DC, Bray KJ, Huisingh CE, Clark ME, McGwin G, Jr., Owsley C. Prevalence of Undiagnosed Age-Related Macular Degeneration in Primary Eye Care. JAMA Ophthalmol. 2017;135(6):570\\u0026ndash;5.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eFerrara D, Newton EM, Lee AY. Artificial intelligence-based predictions in neovascular age-related macular degeneration. Curr Opin Ophthalmol. 2021;32(5):389\\u0026ndash;96.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eKumar H, Goh KL, Guymer RH, Wu Z. A clinical perspective on the expanding role of artificial intelligence in age-related macular degeneration. Clin Exp Optom. 2022;105(7):674\\u0026ndash;9.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eWu K, Wu E, Theodorou B, Liang W, Mack C, Glass L, et al. Characterizing the Clinical Adoption of Medical AI Devices through U.S. Insurance Claims. NEJM AI. 2024;1(1):AIoa2300030.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eTseng R, Gunasekeran DV, Tan SSH, Rim TH, Lum E, Tan GSW, et al. Considerations for Artificial Intelligence Real-World Implementation in Ophthalmology: Providers' and Patients' Perspectives. Asia Pac J Ophthalmol (Phila). 2021;10(3):299\\u0026ndash;306.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLi JO, Liu H, Ting DSJ, Jeon S, Chan RVP, Kim JE, et al. Digital technology, tele-medicine and artificial intelligence in ophthalmology: A global perspective. Prog Retin Eye Res. 2021;82:100900.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eSingh RP, Hom GL, Abramoff MD, Campbell JP, Chiang MF, Intelligence AAOTFoA. Current Challenges and Barriers to Real-World Artificial Intelligence Adoption for the Healthcare System, Provider, and the Patient. Transl Vis Sci Technol. 2020;9(2):45.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eDamschroder LJ, Reardon CM, Widerquist MAO, Lowery J. The updated Consolidated Framework for Implementation Research based on user feedback. Implement Sci. 2022;17(1):75.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLy A, Herse S, Williams M-A, Stapleton F. Artificial Intelligence for Age-Related Macular Degeneration Diagnosis in Australia: A Novel Qualitative Interview Study. Ophthalmic Physiol Opt. 2025.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eThe Center for Implementation. CFIR 2.0. V2025.01 ed2025. p. Adapted from Damschroder, L. J., Reardon, C. M., Widerquist, M. A. O., et al. (2022). The updated consolidated framework for implementation research based on user feedback. Implementation Science, 17, 75. \\u003cspan class=\\\"ExternalRef\\\"\\u003e\\u003cspan class=\\\"RefSource\\\"\\u003ehttps://doi.org/10.1186/s13012-022-01245-0\\u003c/span\\u003e\\u003cspan address=\\\"10.1186/s13012-022-01245-0\\\" targettype=\\\"DOI\\\" class=\\\"RefTarget\\\"\\u003e\\u003c/span\\u003e\\u003c/span\\u003e.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eFujimori R, Liu K, Soeno S, Naraba H, Ogura K, Hara K, et al. Acceptance, Barriers, and Facilitators to Implementing Artificial Intelligence-Based Decision Support Systems in Emergency Departments: Quantitative and Qualitative Evaluation. JMIR Form Res. 2022;6(6):e36501.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eNeher M, Petersson L, Nygren JM, Svedberg P, Larsson I, Nilsen P. Innovation in healthcare: leadership perceptions about the innovation characteristics of artificial intelligence-a qualitative interview study with healthcare leaders in Sweden. Implement Sci Commun. 2023;4(1):81.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eRobinson R, Liday C, Lee S, Williams IC, Wright M, An S, et al. Artificial Intelligence in Health Care-Understanding Patient Information Needs and Designing Comprehensible Transparency: Qualitative Study. JMIR AI. 2023;2:e46487.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLiao X, Yao C, Jin F, Zhang J, Liu L. Barriers and facilitators to implementing imaging-based diagnostic artificial intelligence-assisted decision-making software in hospitals in China: a qualitative study using the updated Consolidated Framework for Implementation Research. BMJ Open. 2024;14(9):e084398.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eHo V, Brown Johnson C, Ghanzouri I, Amal S, Asch S, Ross E. Physician- and Patient-Elicited Barriers and Facilitators to Implementation of a Machine Learning-Based Screening Tool for Peripheral Arterial Disease: Preimplementation Study With Physician and Patient Stakeholders. JMIR Cardio. 2023;7:e44732.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003evan de Sande D, Chung EFF, Oosterhoff J, van Bommel J, Gommers D, van Genderen ME. To warrant clinical adoption AI models require a multi-faceted implementation evaluation. NPJ Digit Med. 2024;7(1):58.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eSanz MF, Acha BV, Garcia MF. Co-Design for People-Centred Care Digital Solutions: A Literature Review. Int J Integr Care. 2021;21(2):16.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eAhmed MI, Spooner B, Isherwood J, Lane M, Orrock E, Dennison A. A Systematic Review of the Barriers to the Implementation of Artificial Intelligence in Healthcare. Cureus. 2023;15(10):e46454.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eHe J, Baxter SL, Xu J, Xu J, Zhou X, Zhang K. The practical implementation of artificial intelligence technologies in medicine. Nat Med. 2019;25(1):30\\u0026ndash;6.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eParikh RB, Helmchen LA. Paying for artificial intelligence in medicine. NPJ Digit Med. 2022;5(1):63.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eAbbas S, Usmani A, Imran M. Willingness To Pay And Its Role In Health Economics. Journal of Bahria University Medical and Dental College. 2018;09:62\\u0026ndash;6.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eGrant K, McParland A, Mehta S, Ackery AD. Artificial Intelligence in Emergency Medicine: Surmountable Barriers With Revolutionary Potential. Ann Emerg Med. 2020;75(6):721\\u0026ndash;6.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eHan R, Acosta JN, Shakeri Z, Ioannidis JPA, Topol EJ, Rajpurkar P. Randomised controlled trials evaluating artificial intelligence in clinical practice: a scoping review. Lancet Digit Health. 2024;6(5):e367-e73.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLoncaric F, Camara O, Piella G, Bijnens B. Integration of artificial intelligence into clinical patient management: focus on cardiac imaging. Rev Esp Cardiol (Engl Ed). 2021;74(1):72\\u0026ndash;80.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLeslye Denisse Dias D. Deskilling of medical professionals: an unintended consequence of AI implementation? Giornale di Filosofia. 2021;2(2).\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eHoff T. Deskilling and adaptation among primary care physicians using two work innovations. Health Care Manage Rev. 2011;36(4):338\\u0026ndash;48.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLi L. Reskilling and Upskilling the Future-ready Workforce for Industry 4.0 and Beyond. Inf Syst Front. 2022:1\\u0026ndash;16.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eIgnatius A, Lakhani, K. AI Won\\u0026rsquo;t Replace Humans \\u0026mdash; But Humans With AI Will Replace Humans Without AI Harvard Business Review2023 [Available from: \\u003cspan class=\\\"ExternalRef\\\"\\u003e\\u003cspan class=\\\"RefSource\\\"\\u003ehttps://hbr.org/2023/08/ai-wont-replace-humans-but-humans-with-ai-will-replace-humans-without-ai\\u003c/span\\u003e\\u003cspan address=\\\"https://hbr.org/2023/08/ai-wont-replace-humans-but-humans-with-ai-will-replace-humans-without-ai\\\" targettype=\\\"URL\\\" class=\\\"RefTarget\\\"\\u003e\\u003c/span\\u003e\\u003c/span\\u003e.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLongoni C, Bonezzi A, Morewedge CK. Resistance to Medical Artificial Intelligence. Journal of Consumer Research. 2019;46(4):629\\u0026ndash;50.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eRiedl R, Hogeterp SA, Reuter M. Do patients prefer a human doctor, artificial intelligence, or a blend, and is this preference dependent on medical discipline? Empirical evidence and implications for medical practice. Front Psychol. 2024;15:1422177.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eGreene SM, Tuzzio L, Cherkin D. A framework for making patient-centered care front and center. Perm J. 2012;16(3):49\\u0026ndash;53.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eTopol E. Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again: Basic Books; 2019.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eTalwar S, Dhir A, Islam N, Kaur P, Almusharraf A. Resistance of multiple stakeholders to e-health innovations: Integration of fundamental insights and guiding research paths. Journal of Business Research. 2023;166:114135.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eDesveaux L, Soobiah C, Bhatia RS, Shaw J. Identifying and Overcoming Policy-Level Barriers to the Implementation of Digital Health Innovation: Qualitative Study. J Med Internet Res. 2019;21(12):e14994.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eYang Y, Ngai EWT, Wang L. Resistance to artificial intelligence in health care: Literature review, conceptual framework, and research agenda. Information \\u0026amp; Management. 2024;61(4):103961.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003ePrakash AV, Das S. Medical practitioner's adoption of intelligent clinical diagnostic decision support systems: A mixed-methods study. Information \\u0026amp; Management. 2021;58(7):103524.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003ePark EH, Werder K, Cao L, Ramesh B. Why do Family Members Reject AI in Health Care? Competing Effects of Emotions. Journal of Management Information Systems. 2022;39(3):765\\u0026ndash;92.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eKleijnen M, Lee N, Wetzels M. An exploration of consumer resistance to innovation and its antecedents. Journal of Economic Psychology. 2009;30(3):344\\u0026ndash;57.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eGrol R, Wensing M. Effective Implementation of Change in Healthcare. Improving Patient Care2020. p. 45\\u0026ndash;71.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eLemak CH, Pena D, Jones DA, Kim DH, Guptill J. Leadership to Accelerate Healthcare's Digital Transformation: Evidence From 33 Health Systems. J Healthc Manag. 2024;69(4):267\\u0026ndash;79.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eWales launches first national digital eye care patient record system [press release].\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eOpenEyes [press release]. 2022.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eOpenEyes Platform Proves Concept in Glasgow [press release]. 2024.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eAbramoff MD, Dai T, Zou J. Scaling Adoption of Medical AI \\u0026mdash; Reimbursement from Value-Based Care and Fee-for-Service Perspectives. NEJM AI. 2024;1(5):AIpc2400083.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eEl Arab RA, Al Moosa OA, Sagbakken M. Economic, ethical, and regulatory dimensions of artificial intelligence in healthcare: an integrative review. Front Public Health. 2025;13:1617138.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eParanjape K, Schinkel M, Nannan Panday R, Car J, Nanayakkara P. Introducing Artificial Intelligence Training in Medical Education. JMIR Med Educ. 2019;5(2):e16048.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eSyeda LH, Batool Z, Hayder Z, Ali S. Medical undergraduate students' awareness and perspectives on artificial intelligence: A developing nation's context. BMC Med Educ. 2025;25(1):1060.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eAustralian Alliance for Artificial Intelligence in Healthcare. A National Policy Roadmap for Artificial Intelligence in Healthcare.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eRajesh AE, Davidson OQ, Lee CS, Lee AY. Artificial Intelligence and Diabetic Retinopathy: AI Framework, Prospective Studies, Head-to-head Validation, and Cost-effectiveness. Diabetes Care. 2023;46(10):1728\\u0026ndash;39.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eNilsson C. The artificial intelligence (AI) competence paradox: how AI reshapes clinical expertise. Transforming Government: People, Process and Policy. 2025.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eNatali C, Marconi L, Dias Duran LD, Cabitza F. AI-induced Deskilling in Medicine: A Mixed-Method Review and Research Agenda for Healthcare and Beyond. Artificial Intelligence Review. 2025;58(11):356.\\u003c/span\\u003e\\u003c/li\\u003e\\u003cli\\u003e\\u003cspan\\u003eCanfell OJ, Chan W, Pole JD, Engstrom T, Saul T, Daly J, et al. Artificial intelligence after the bedside: co-design of AI-based clinical informatics workflows to routinely analyse patient-reported experience measures in hospitals. BMJ Health Care Inform. 2024;31(1).\\u003c/span\\u003e\\u003c/li\\u003e\\u003c/ol\\u003e\"}],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":true,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":false,\"hideJournal\":false,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":true,\"isAuthorSuppliedPdf\":false,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":false,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"ophthalmic-and-physiological-optics\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":false,\"externalIdentity\":\"\",\"sideBox\":\"Learn more about [Ophthalmic and Physiological Optics](https://link.springer.com/journal/44402)\",\"snPcode\":\"44402\",\"submissionUrl\":\"https://submission.springernature.com/new-submission/44402/3?\",\"title\":\"Ophthalmic and Physiological Optics\",\"twitterHandle\":\"\",\"acdcEnabled\":true,\"dfaEnabled\":true,\"editorialSystem\":\"stoa\",\"reportingPortfolio\":\"Springer Open\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":true},\"keywords\":\"artificial intelligence, barriers, implementation, stakeholders\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-7911455/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-7911455/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"\\u003ch2\\u003ePurpose\\u003c/h2\\u003e\\u003cp\\u003eDespite the revolution of artificial intelligence (AI), its integration remains limited in healthcare. A comprehensive understanding of the barriers to implementation is crucial to enhance the utilisation of AI. This study applies a conceptual framework-based analysis, to explore stakeholder perspectives of implementation barriers of AI in digital diagnosis in eye care.\\u003c/p\\u003e\\u003ch2\\u003eMethods\\u003c/h2\\u003e\\u003cp\\u003ePurposive sampling was used to identify key individuals across stakeholder groups, including technology developers, clinicians, patients, and healthcare leaders. Semi-structured interviews were conducted with 37 stakeholders. Using the Updated Consolidated Framework for Implementation Research (CFIR), responses to the question: \\u0026lsquo;What is the biggest barrier to digital diagnosis or AI, specifically age-related macular degeneration (AMD) in Australia?\\u0026rsquo; were analysed. Barriers identified by stakeholders were mapped to thematic constructs of Updated CFIR and the relative importance of each implementation barrier was measured.\\u003c/p\\u003e\\u003ch2\\u003eResults\\u003c/h2\\u003e\\u003cp\\u003eFor clinicians and developers, \\u0026lsquo;innovation\\u0026rsquo; domain was the most frequently cited. Clinicians were most concerned of the costs involved; whereas for developers the lack of evidence of the innovation in real world applications was the main challenge. For leaders and patients, \\u0026lsquo;individuals\\u0026rsquo; domain was the most frequently cited. Leaders were focused on the innovation deliverers: expressing the potential risk of over-reliance on the innovation, and the subsequent consequence of clinician deskilling. Patients were more concerned about innovation recipients: emphasising the perceived lack of human empathy with the implementation of AI.\\u003c/p\\u003e\\u003ch2\\u003eConclusions\\u003c/h2\\u003e\\u003cp\\u003eDifferences were revealed in the identified barriers to the implementation of AI across stakeholder groups. A co-design approach to address the misalignment in key barriers may be essential to successful implementation of AI in digital health innovations.\\u003c/p\\u003e\",\"manuscriptTitle\":\"Stakeholder Perspectives of Implementation Barriers of Artificial Intelligence in Eye Care: A qualitative framework-based study\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2025-11-04 10:22:58\",\"doi\":\"10.21203/rs.3.rs-7911455/v1\",\"editorialEvents\":[{\"type\":\"communityComments\",\"content\":0},{\"type\":\"decision\",\"content\":\"Revision requested\",\"date\":\"2025-12-23T18:07:40+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"editorInvitedReview\",\"content\":\"\",\"date\":\"2025-12-20T19:57:53+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"editorInvitedReview\",\"content\":\"\",\"date\":\"2025-12-07T22:30:04+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewerAgreed\",\"content\":\"199710532045021101643689378225287890488\",\"date\":\"2025-11-30T22:53:20+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewerAgreed\",\"content\":\"275010921006241828733394270628896872828\",\"date\":\"2025-11-30T20:41:26+00:00\",\"index\":\"hide\",\"fulltext\":\"\"},{\"type\":\"reviewersInvited\",\"content\":\"\",\"date\":\"2025-10-23T16:43:52+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"editorAssigned\",\"content\":\"\",\"date\":\"2025-10-23T16:42:44+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"checksComplete\",\"content\":\"\",\"date\":\"2025-10-22T00:04:51+00:00\",\"index\":\"\",\"fulltext\":\"\"},{\"type\":\"submitted\",\"content\":\"Ophthalmic and Physiological Optics\",\"date\":\"2025-10-21T07:03:02+00:00\",\"index\":\"\",\"fulltext\":\"\"}],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"ophthalmic-and-physiological-optics\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":false,\"externalIdentity\":\"\",\"sideBox\":\"Learn more about [Ophthalmic and Physiological Optics](https://link.springer.com/journal/44402)\",\"snPcode\":\"44402\",\"submissionUrl\":\"https://submission.springernature.com/new-submission/44402/3?\",\"title\":\"Ophthalmic and Physiological Optics\",\"twitterHandle\":\"\",\"acdcEnabled\":true,\"dfaEnabled\":true,\"editorialSystem\":\"stoa\",\"reportingPortfolio\":\"Springer Open\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":true}}],\"origin\":\"\",\"ownerIdentity\":\"2ae2a8f4-f08f-4f86-9a56-b55b8b193d09\",\"owner\":[],\"postedDate\":\"November 4th, 2025\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"under-review\",\"subjectAreas\":[],\"tags\":[],\"updatedAt\":\"2026-03-26T13:55:54+00:00\",\"versionOfRecord\":[],\"versionCreatedAt\":\"2025-11-04 10:22:58\",\"video\":\"\",\"vorDoi\":\"\",\"vorDoiUrl\":\"\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-7911455\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-7911455\",\"identity\":\"rs-7911455\",\"version\":[\"v1\"]},\"buildId\":\"XKTyCvWXoU3ODBz1xrDgd\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}