ORUK-VA-0001
Section 1 - Basic information about you and your application:
Title of research project
Implementation Science of AI Solutions in Trauma & Orthopaedic Surgery: Evaluating the Feasibility and Health Economic Implications of the Deployment of AI-driven fracture detection software to streamline fracture management
Duration
48
Start date
01/09/2023
Profession
Orthopaedic surgeon
Your current job title/position
ST5, NIHR Academic Clinical Fellow
Have you previously received funding from Orthopaedic Research UK or Versus Arthritis
No
Please provide project ref number for previous funding from Orthopaedic Research UK
Please provide project ref number for previous funding from Versus Arthritis
Did you attend the AI in Orthopaedic Conference organised by Orthopaedic Research UK in 2022?
Yes
Did you attend the Microsoft AI training courses organised by Orthopaedic Research UK?
No
What other AI training courses have you previously attended?
Grant: -Successful AOUK Grant Proposal: Artificial Intelligence in Orthopaedics – Using Machine Vision in Acute Paediatric Trauma to Streamline Fracture Management Courses: -University of Manchester/HEE, AI for Healthcare: Equipping the Workforce for Digital Transformation; January 2022 -Harvard Executive Education, Designing and Implementing AI Solutions for Healthcare; June 2023 Conferences: -AI in Orthopaedics, ORUK; Royal College of Surgeons England, London; September 2022 -Stanford University, Artificial Intelligence in Medical Imaging Symposium; June 2023
Are you an early-career researcher (ECR)? (definition of ECR)
yes
Section 2 - Lay summary
Lay summary:
Background
Artificial Intelligence (AI) is a field of technology that focuses on creating computer systems that can perform tasks that usually require human reasoning and intelligence. Fractures make up a large proportion of patients seen in A+E everyday, with a huge cost to the NHS. That is why I want to focus my studies on the potential impact of AI in Trauma & Orthopaedic Surgery, specifically launching a system in the NHS which uses AI to automatically detect fractures on x-rays.
Aims
Specifically, I want to look into systems that help doctors detect fractures quicker and more accurately from x-ray images of bones. By doing this research, I hope to understand what exactly is preventing the use of such AI-systems in the NHS, what the financial cost of using them would be and to make sure the fracture detection systems can actually be used in real-life healthcare situations. This will not only assist doctors in making quicker, better decisions but also improve patient care and our use of NHS resources.
Methods
The main factors I aim to study are:
Checking Availability: I will determine what fracture detection systems are currently available and approved for use in real-world healthcare settings.
Investigating barriers to deployment: Introducing new technology into healthcare can be challenging. I will explore the obstacles and issues that might arise when implementing the fracture detection software in the NHS. This will include things like integrating the software with existing healthcare systems, training healthcare professionals to use it effectively, and addressing any concerns or resistance from doctors and patients.
Considering the Costs and Benefits: I will explore the financial impact of using the fracture detection software in the NHS. This involves studying the costs involved in setting up, maintaining, and training people to use the software. I’ll also look at the potential benefits, such as improved patient care and outcomes. By looking at both sides, we can determine if it makes financial sense to use the software widely in the NHS.
Strategy: I will develop an understanding on the best methods to take in order to successfully launch an AI fracture detection system, in the NHS.
Potential Impact
By focusing on these objectives, my goal is to thoroughly evaluate the use of AI fracture detection systems. This research will help us determine what is available, the economic impact it may have, and whether it is possible to use in real healthcare settings. Ultimately, the aim is to make sure the software is cost-effective, and practical for the NHS. In turn, I hope for this work will enable us to safely start to use such AI technology in the NHS, which is going to improve patient care and outcomes.
Section 3 - Purpose of research
Purpose of research:
Aims:
The primary objective of this study is to scrutinize the feasibility and ramifications of implementing AI-driven fracture detection systems within the NHS.
Methodology:
- Evidence Synthesis: A comprehensive assessment of the availability of approved fracture detection AI systems will be undertaken to understand the options that are available for deployment in real-world healthcare environments.
- Health Economic Analysis: An economic evaluation will be conducted to appraise the financial implications of integrating the fracture detection software within the NHS. This analysis entails examining the costs associated with software implementation, maintenance, and personnel training. Additionally, the potential benefits, including enhanced patient care, improved clinical outcomes and optimised resource allocation, will be considered. A comprehensive examination of both costs and benefits will inform the financial viability of widespread deployment of the software within the NHS.
- Barriers to Implementation: The implementation of novel technology in healthcare necessitates addressing a range of challenges. This investigation will explore obstacles that may arise when deploying fracture detection software in the NHS. Factors to be examined include the seamless integration of the software with existing healthcare systems, effective training of healthcare professionals in its use, and addressing any concerns or resistance that may emerge from clinicians and patients.
- Strategising Implementation: A strategy will be devised to outline the optimal pathways for the successful launch and integration of an AI-driven fracture detection system within the NHS ecosystem. This entails a systematic blueprint encompassing technological, procedural, and logistical considerations.
Section 4 - Background to investigation
Background to investigation:
Background
Artificial intelligence (AI) is a rapidly advancing field that holds immense potential for various applications in healthcare. With its ability to recognize patterns and derive associations from large datasets, AI has emerged as a valuable tool for analysing complex medical data and improving diagnostic processes. In recent years, there has been a particular focus on leveraging AI techniques, such as machine learning, to enhance image interpretation and facilitate accurate diagnosis, ultimately leading to optimised patient care.
The significance of AI in healthcare and its potential to revolutionise clinical practice have been widely acknowledged. The NHS Long Term Plan (2019), a strategic vision for the National Health Service (NHS), recognizes the transformative role of AI and emphasises its integration into the digital transformation of healthcare services.
This recognition is also reflected in the National Institute for Health and Care Research (NIHR) Agenda, which highlights the importance of AI in advancing healthcare research and delivery. It states that the NIHR AI in Health and Care Award has allocated £123 million to “expedite the testing and evaluation of technologies that are most likely to align with the goals outlined in the NHS Long Term Plan”. By harnessing the potential of artificial intelligence, the AI in Health and Care Award seeks to improve health and care outcomes for patients. This initiative strives to augment the impact of AI systems in addressing various challenges faced by the NHS, such as reducing waiting times, facilitating early diagnosis, and optimizing staff efficiency. This ties in beautifully with the implementation of a national AI-driven fracture detection system, which would have a direct impact on all three of these measures, improving resource allocation and patient care.
As AI continues to evolve, its application in medical imaging has gained considerable attention. Medical imaging plays a crucial role in diagnosing various conditions, including fractures. Accurate and timely identification of fractures is essential for appropriate clinical decision-making and effective treatment planning. Traditionally, fracture detection has relied on the expertise of radiologists and orthopaedic specialists. However, the advent of AI-based image interpretation tools has opened new avenues for improving fracture detection and classification.
Research in this field has demonstrated promising results, highlighting the potential of AI algorithms in fracture detection. These algorithms utilize machine learning techniques to analyse X-ray images, identify fracture patterns, and differentiate them from normal anatomical structures. By leveraging the power of AI, these algorithms can process vast amounts of medical data, learn from patterns, and make accurate predictions, thereby assisting clinicians in their diagnostic tasks.
To fully capitalise on the potential of AI in fracture detection, it is essential to conduct a comprehensive evaluation of the existing evidence. A systematic review conducted by Jones et al. emphasizes the transformative impact of AI in healthcare and highlights the need for further exploration and understanding of its capabilities. Additionally, Langerhuizen et al.’s systematic review underscores the applications and limitations of AI for fracture detection in orthopaedic trauma imaging. Kuo et al’s meta-analysis concluded that “artificial intelligence (AI) and clinicians had comparable reported diagnostic performance in fracture detection, suggesting that AI technology holds promise as a diagnostic adjunct in future clinical practice”. These studies provide valuable insights into the current state of AI-based fracture detection and serve as a foundation for the proposed research project, however there is not a clear body of work which synthesises and compares the numerous AI-driven fracture detection systems which have passed regulatory testing and been launched with CE and FDA approval.
Building upon the existing evidence, this research aims to investigate systems specifically developed to accelerate and improve the precision of fracture detection in X-ray imaging, that are at a stage where they are ready for deployment. By identifying, evaluating, and assessing the feasibility of deploying such fracture detection systems in real-world situations, this study seeks to address the gaps in knowledge and contribute to the advancement of AI in fracture diagnosis and its implementation in the NHS.
Overview of Artificial Intelligence in Healthcare
Artificial intelligence (AI) has transformed various industries, including healthcare, by offering innovative approaches to data analysis and decision-making. In healthcare, AI techniques, such as machine learning, deep learning, and natural language processing, have shown significant potential in improving diagnostic accuracy, treatment planning, and patient management. AI algorithms can process and analyse vast amounts of medical data, learn patterns, and generate valuable insights to support healthcare professionals in making informed decisions.
Significance of AI in Fracture Detection
Fracture detection plays a crucial role in orthopaedic care, as it aids in accurate diagnosis, treatment planning, and monitoring of fractures. Traditional methods of fracture detection rely on manual interpretation of X-ray images, which can be subjective, time-consuming, and prone to variability among clinicians. AI-based fracture detection systems offer a promising solution to overcome these challenges.
By leveraging AI techniques, such as deep learning algorithms, these systems can learn from a vast number of labelled X-ray images to recognise fracture patterns and distinguish them from normal anatomical structures. The ability of AI algorithms to analyse images quickly and accurately has the potential to enhance diagnostic accuracy, reduce interpretation variability, and optimise workflow efficiency in fracture detection.
Section 5 - Plan of investigation
Plan of investigation:
I. Project Plan and Methodology
The proposed study will be executed in 4 phases:
- Evidence Synthesis:
-Conduct a systematic review of advanced stage fracture detection software, which have passed final trials and have FDA/CE-approval.
-Develop a comprehensive search strategy to identify relevant studies and literature.
-Screen and evaluate the selected studies based on predetermined inclusion and exclusion criteria.
-Extract and analyse data from the included studies to assess the performance, strengths, limitations, and advancements of fracture detection algorithms and software.
-Synthesise the findings to provide an unbiased and comprehensive assessment of the field, to understand what systems are ready for deployment.
- Health Economic Analysis:
-Evaluate the financial implications of integrating the fracture detection software into the NHS healthcare infrastructure.
-Identify and quantify the costs associated with software implementation, maintenance, and personnel training.
-Examine the potential benefits, including enhanced patient care, improved clinical outcomes and optimised resource allocation.
-Conduct a comprehensive cost-benefit analysis to determine the economic viability of widespread adoption of the software in the NHS.
- Barriers to Deployment:
-Investigate the challenges and obstacles in deploying fracture detection software in the NHS.
-Assess the compatibility and integration of the software with existing healthcare systems.
-Develop effective training programs for healthcare professionals to ensure proficiency in using the software.
-Evaluate concerns or resistance that may arise from clinicians and patients regarding the implementation of the software with qualitative analysis.
-Explore strategies to overcome barriers and ensure the smooth adoption and acceptance of the software in real healthcare settings.
- Strategizing Implementation:
-Devise a comprehensive strategy for the seamless integration and deployment of AI-driven fracture detection software within the multifaceted NHS ecosystem.
-Pioneer a meticulous blueprint that encapsulates technological, procedural, and logistical considerations, ensuring harmonious assimilation.
-Synchronise the software with prevailing healthcare systems, formulating tailored training regimens, and adroitly addressing potential barriers and concerns.
II. Dissemination and Outputs
The research outcomes will be disseminated through peer-reviewed publications, conference presentations, and knowledge exchange activities. The findings of the systematic review will contribute to the existing literature on AI-based fracture detection and provide insights into the performance, limitations, and advancements in this field. The validation of existing algorithms, health economic analysis, and feasibility investigation will inform healthcare professionals, policymakers, and technology developers, enabling evidence-based decision-making and promoting the effective utilisation of AI technologies for improved fracture diagnosis and patient care.
III. Anticipated Impact
With such rapidly evolving technology, AI is already being introduced into many aspects of healthcare. With regards image interpretation, there has been a drive for the more routine use of AI[1].
Given the high-pressure and very chaotic nature of Emergency Departments all over the world, there is massive potential to facilitate the detection of fractures with a machine vision tool utilising AI. This would effectively streamline the diagnosis and guide the management of these fractures, preventing misdiagnosis and improving patient outcomes[6,7]. Such a tool has the scope to majorly benefit the NHS as a whole in terms of enhancing patient safety, improving staff workload and optimising resource allocation. This will move us into a position whereby we can aim to make strides towards deployment, taking the technology from the bench to the bedside.
IV. Project Management
The proposed research project will be managed efficiently by establishing clear milestones, timelines, and deliverables. Regular project meetings will be conducted to ensure effective communication and collaboration among the research team members. Data management protocols will be established to ensure the security, integrity, and confidentiality of the collected data.
V. Ethics
The research project will adhere to ethical principles and guidelines to protect the rights, privacy, and confidentiality of the study participants and the integrity of the research. Ethical approval will be sought from the relevant ethics committees or institutional review boards before conducting any data collection activities. Informed consent will be obtained from participants involved in the validation and feasibility phases of the study.
VI. Success Criteria
The success of the research project will be evaluated based on the achievement of the following criteria:
-Completion of a comprehensive systematic review of the literature on advanced stage AI-based fracture detection in X-ray imaging.
-Conducting a robust health economic analysis to evaluate the financial implications of implementing AI-based fracture detection software in healthcare settings.
-Investigation of the feasibility of deploying AI-based fracture detection software in real-world healthcare environments.
-Develop a strategy to optimise the deployment and implementation of AI-driven fracture detection software in the NHS.
-Dissemination of research outcomes through peer-reviewed publications, conference presentations, and knowledge exchange activities.
By addressing these success criteria, the proposed research project will contribute to advancing the field of AI in fracture detection, enhancing the accuracy and efficiency of fracture diagnosis, and ultimately improving patient care outcomes.
Section 6 - Research environment and resources
Research environment and resources:
Project Support
The project will form the basis of a PhD which I aim to complete in a prestigious academic institution, in collaboration with a group out of AIMI (Artificial Intelligence in Medicine and Imaging) Center, Stanford University, USA.
Professor Dan Perry (Orthopaedic Consultant, Alder Hey Hospital; NIHR Research Professor) will be lead supervisor and provide Orthopaedic research expertise, guidance and direction. Dr Susan Shelmerdine (Radiology Consultant, Great Ormond Street Hospital; Chief Investigator FRACTURE Study) will be a senior collaborator and provide AI in Radiology expertise. Dr Alaa Yousseff (PhD Implementation Science; AIMI Research Fellow) will be a collaborator and provide implementation science expertise.
Project Management
The proposed research project will be managed efficiently by establishing clear milestones, timelines, and deliverables. Regular project meetings will be conducted to ensure effective communication and collaboration among the research team members. Data management protocols will be established to ensure the security, integrity, and confidentiality of the collected data.
Section 7: Research impact
Who will benefit from this research?
Anticipated Impact
With such rapidly evolving technology, AI is already being introduced into many aspects of healthcare. With regards image interpretation, there has been a drive for the more routine use of AI.
Given the high-pressure and very chaotic nature of Emergency Departments all over the world, there is massive potential to facilitate the detection of fractures with a machine vision tool utilising AI. This would effectively streamline the diagnosis and guide the management of these fractures, preventing misdiagnosis and improving patient outcomes. Such a tool has the scope to majorly benefit the NHS as a whole in terms of enhancing patient safety, improving staff workload and optimising resource allocation.
How can your research be translated in real-life?
Translation
The ultimate goal of this research project is to conduct a comprehensive evaluation of the feasibility of deployment of AI-driven fracture detection systems, within the NHS environment. This research will provide insights into the software’s economic implications, the clinician, patient, structural and workflow barriers to its implementation within real healthcare environments, as well as the optimal strategy to take to ensure successful deployment of such AI solutions in the NHS. This project aims to ensure the system’s readiness, cost-effectiveness, and practicality for utilisation within the NHS and to confirm its positive impact on patient care. This will move us into a position whereby we can aim to make strides towards deployment, taking the technology from the bench to the bedside.
How will your research be beneficial for Orthopaedic Research UK, Versus Arthritis and their purpose?
According to the ORUK Impact Report, 2022: “ORUK funds innovative research projects in the UK that expand knowledge, improve patient outcomes and pioneer new forms of MSK diagnosis and treatment.”.
This project does exactly that. It is a study truly focused on IMPACT, by way of evaluating the barriers to implementation of key AI Solutions (fracture detection software) which will transform the healthcare landscape as we know it. My aim is to develop a deep understanding of critical strategies which will enable us to take AI-driven orthopaedic software from the bench to the bedside. This ties in seamlessly with the “NHS Long Term plan” and the “NIHR Agenda”.
Understanding these key factors will move us closer to a position whereby we are able to deploy AI technology in the NHS, improving patient outcomes, staff workload and resource allocation.
Section 8: Outreach and engagement
Outreach and engagement
Dissemination and Outputs
The research outcomes will be disseminated through peer-reviewed publications, conference presentations, and knowledge exchange activities. The findings of the systematic review will contribute to the existing literature on AI-based fracture detection and provide insights into the performance, limitations, and advancements in this field. The health economic analysis, feasibility investigations and development of implementation strategy will inform healthcare professionals, policymakers, and technology developers, enabling evidence-based decision-making and promoting the effective utilisation of AI technologies for improved fracture diagnosis and patient care.
Furthermore, there will be public engagement events organised as part of the “Barriers to Implementation” phase in order to understand any reservations the broader patient-base have as well as to involve them as part of the public steering committee. This is critical to ensure a holistic approach to the implementation strategy and subsequent public outreach.
I have been involved in steering meetings for the FRACTURE Study, alongside Dr Susan Shelmerdine and Prof Dan Perry. Public engagement events have been arranged in order to understand public perception and opinion on the potential value of AI in paediatric fracture detection.
Section 9: Research budget
Requested funding from Orthopaedic Research UK and Versus Arthritis
University fees (if any)
£200000
Salary
£150000
Consumables
£50000
Publications
£0
Conference attendance
£0
Other items
£
Total 'requested fund'
£400000
Other items
Other secured funds
Internal funding
£0
Partner (University)
£0
Partner (Commercial)
£0
Partner (Charity)
£0
Other sources
£3500
Total 'other funds'
£3500
List all the 'other sources' and explain how their funds are used to cover the costs of your research.
Section 10: Intellectual property and testing on animal
Is there an IP linked to this research?
No
Who owns and maintains this patent?
Does your research include procedures to be carried out on animals in the UK under the Animals (Scientific Procedures) Act?
No
If yes, have the following necessary approvals been given by:
The Home office(in relation to personal, project and establishment licences)?
Animal Welfare and Ethical Review Body?
Does your research involve the use of animals or animal tissue outside the UK?
No
Does the proposed research involve a protected species? (If yes, state which)
Does the proposed research involve genetically modified animals?
Include details of sample size calculations and statistical advice sought. Please use the ARRIVE guidelines when designing and describing your experiments.
There should be sufficient information to allow for a robust review of any applications involving animals. Further guidance is available from the National Centre for the Replacement, Refinement and Reduction of Animals in Research (NC3Rs), including an online experimental design assistant to guide researchers through the design of animal experiments.
Please provide details of any moderate or severe procedures
Why is animal use necessary, are there any other possible approaches?
Why is the species/model to be used the most appropriate?