BHS-005

Section 1 - Basic information about you and your application:

Title of research project
Creation of a large dataset for the prediction of the recovery trajectory of patients with hip osteoarthritis based on their treadmill gait analysis, to optimize treatment and facilitate their return to work.

Grant Type
Collaborative – Multidisciplinary teams working to solve a clinical problem

Research area
Management

Duration
18

Start date
01/01/2024

Have you previously received funding from BHS or ORUK?
no

Your current job title/position
PDRA

Are you an early-career researcher (ECR)? (definition of ECR)
yes

Section 2 - Lay summary

Lay summary:

Problem We are Trying to Solve:

When it comes to surgical treatments of the hip, outcomes can differ greatly. Various surgical techniques exist, but which ones provide the best benefits and to which patients? Which methods ensure swift recovery, minimal pain, and long-term improved mobility? Our study seeks to answer these questions to ensure that patients undergoing hip surgery get the most effective treatment for them.

Estimated Incidence:

Musculoskeletal (MSK) problems, including hip conditions, are common. Millions suffer from hip-related issues every year. With an aging population and increasing lifestyle demands, the incidence is expected to rise. It is crucial that we refine our surgical approaches to make sure the right approach is taken for each patient.

Our Ambition and Impact on People with MSK:

Our goal is to redefine the standards for hip surgeries. By pooling together and examining movement data while walking from many different types of hip research studies through collaborating with different universities and their associated clinical partners, we aim to draw comprehensive insights in the management of hip pain. This isn’t just about the surgical techniques, rather it is about the tangible benefits to patients – faster recovery, reduced pain, and enhanced mobility. By determining the most effective surgical methods, we can ensure that people with hip problems can return to their normal lives more quickly and with fewer problems.

Involving People with Hip Conditions:

People suffering from hip conditions are at the heart of our research. Their experiences, feedback, and outcomes will be integral to our study and their thoughts have already informed the shaping of this study. We will continue to regularly consult a diverse patient and public engagement (PPIE) group, understanding their pre- and post-operative experiences, and ensuring their voices are central to our research process. This will establish a system where patients can provide feedback on their surgical outcomes, helping us to continuously refine our understanding. Once our research starts bearing results, we will reach back to the community, to seek how we can share this information so we can educate people about the most effective surgical techniques and what they can expect. This will be achieved through PPIE meetings where patients can interact with our researchers, fostering a two-way exchange of information.

In summary, this endeavour, generously supported by ORUK and BHS, seeks to bring clarity to the world of hip surgeries. By combining vast amounts of data, we hope to pinpoint the most effective techniques, ensuring that every patient benefits from the latest, most effective surgical advancements. Most importantly, we are making sure that the very people we aim to help – those with hip conditions – remain central to this exploration, from start to finish.

Section 3 - Purpose of research

Purpose of research:

Responding to the escalating challenge of MSK-related disorders and their impact on UK healthcare systems, this research focuses on technologies for detecting functional changes in hip osteoarthritis and understanding their impact on quality of life and subsequent care. This project will leverage a unique opportunity to collate cohorts of UK gait research data. It will bring together datasets capable of addressing vital questions in osteoarthritis hip patient care, specifically developing a predictive model of post-operative function using functional treadmill data and investigating the scope translating this predictive model into clinical care pathways and decision-making tools.

The proposed database draws together 1000 peri-operative cases encompassing various surgical procedures, approaches and implants along with healthy controls. Using this data, analytical approaches will develop robust methods for predicting functional outcomes and exploring management questions. This will permit data mining to evolve management strategies in hip patient care and significantly influencing patient selection, risk stratification, and overall healthcare quality.

The following questions will be explored:

• Does the type of implant, and operative procedure impact functional outcomes?
• Can the surgical approach be optimized to enhance patient recovery?
• Can a comprehensive functional assessment enhance patient care?
• Can postoperative function be foreseen by preoperative function?
• Should age or functional ability be used as a criterion for hip surgery?
• Can novel rapid and scalable technological solutions offer new approaches to data collection permitting wider clinical translation?

This grant explores kinematic outcomes sensitive to peri-operative changes to drive this process, creating a predictive model with the potential to predict functional outcomes using preoperative data. This work will refine patient selection and risk stratification for hip surgery, contributing to waiting list management and surgical outcomes including return to work.

 

Section 4 - Background to investigation

Background to investigation:

Hip osteoarthritis (OA) is a leading cause of pain and disability worldwide¹. Its prevalence increases with age, meta-analysis predicts that before retirement age of 65 20% of the population will have hip osteoarthritis (K-L grade ≥ 2)². This poses a substantial economic burden due to both direct medical costs (e.g., surgeries, hospitalizations, etc) and indirect costs. In the Dutch workforce it is estimated to cost €13.8 million annually³. Additionally, the personal burden on patients and their families is considerable.

The diagnosis of hip OA is based on a combination of clinical evaluation, patient-reported symptoms (pain), and imaging findings. However, symptoms and radiographic changes do not always align. As such, clinical judgment, in conjunction with imaging and patient history, is crucial for diagnosis.

There has been increasing interest in finding novel diagnostic markers⁴, both biochemical (e.g., blood and synovial biomarkers) and biomechanical (e.g., specific gait alterations⁵), to improve early diagnosis and assessment of hip OA. However, these remain research tools and are not suitable yet for translation into clinical practice.

Total hip arthroplasty (THA) is a transformative procedure for end-stage osteoarthritis, with significant improvements in patient-reported outcome measures. A notable percentage, however, remain dissatisfied and/ or experience persistent pain post-operation. While traditional manual THA shows promise, robotic-assisted THA offers superior precision in restoring hip kinematics, potentially reducing complications and improving outcomes. The cost-effectiveness and definitive benefits of robotic over manual THA remains to be established⁶.

Furthermore there is a need to consider persistent challenges for both younger patients and active older individuals including enhancing implant efficiency,  understanding safety issues of innovative implants, and formulating approaches to detect osteoarthritis at its onset thereby decelerating its advancement and reducing the need for extensive complex surgical procedures⁷.

Biomechanics provides a framework for understanding the onset, progression, and management of hip OA⁸. Recent research in osteoarthritis (OA) aetiology has shown promising advancements in biomechanics, with innovations in technology and machine learning (ML) enhancing its precision and applicability. This progress highlights the potential of biomechanics in informing targeted treatments, emphasizing the need for its inclusion in large-scale studies and future clinical decision-making tools⁸. Large scale biomechanical studies^9,10 have been conducted in knee of but no equivalent studies have been conducted in hip OA. Hall et al.’s study¹¹, indicated that the variations in hip joint movements during walking were not solely attributed to pain intensity or the severity of radiographic OA. This underscores the need for more research on hip studies to determine the factors influencing changes in hip OA-related kinematics, through which novel approaches to management could evolve.

Much of the biomechanical research require detailed gait and functional analysis with a wide variety of equipment¹², but these studies often struggle as the variability in gait across a population require large study numbers¹³. Markerless motion capture techniques offer greater data generation and adaptability, with the potential for clinical use. While markerless temporospatial measures approach marker-based accuracy, joint details lack clinical precision, and without a gold standard for comparison, their true accuracy remains uncertain¹⁴. We suggest that markerless motion capture will play an important role in the recruitment and management of people with osteoarthritis through mass data collection efforts, upon which more detailed analysis can be performed. However, to understand this key kinematic factors need to be identified and this can only be achieved through the creation of a large database as proposed here.

Artificial Intelligence (AI) offers transformative potential in healthcare, by enhancing care quality, personalizing medicine, streamlining research, and empowering both patients and professionals¹⁵. Integrating machine learning (ML) with human biomechanics simplifies gait analysis, with supervised ML techniques like support vector machine (SVM) showing over 90% accuracy and reinforcement learning benefiting gait rehabilitation and pain/loading adaptation strategies, making ML pivotal for clinical diagnosis and predicting outcomes¹⁶.

The 1000 participants have been scoped across three sites in the UK (Imperial College London (400)^17,18, University College London (UCL) (200)¹⁹ and Bournemouth University (400)²⁰). This will be added to through the continuation of recruitment and collect at these site in additional to publicly available databases²¹.

The research team has a well-established track record in this field. In a series of studies, Prof. McGregor highlighted the synergy between ML and biomechanical models. ML techniques, including random forest²² and probabilistic principal component analysis²³, were employed to discern variations in gait data associated with osteoarthritis. Further studies capitalized on deep learning to generate synthetic motion capture datasets, capturing marker trajectories and ground reaction forces²⁴. These demonstrate how musculoskeletal models can accurately determine internal forces during dynamic movements²⁵.

Dr Maslivec’s has an established database of gait function in hip arthroplasty in patients undergoing Ceramic Hip Resurfacing Arthroplasty (cHRA) and Total Hip Arthroplasty (THA), it was noted that both procedures improved gait function, but cHRA patients exhibited superior gait profiles post-operatively.

Dr. Banger has extensively researched the efficacy and outcomes of robotic arm-assisted arthroplasty RCTs. He also has a growing portfolio of industrial collaborations on using machine learning algorithms to deliver solutions for wearable product development.

Professor Cobb has extensive expertise in research and innovation within arthroplasty around surgical outcome measures³¹, surgical implant development, haptic-based robotic assistance³² and 3D planning³³. His proven track record in securing substantial funding and translating research into clinical practice makes his involvement a significant asset to the grant’s success.

Professor Fares Haddad, a renowned orthopaedic surgeon and academic, brings invaluable expertise in hip and knee surgery, as well as a strong track record in running arthroplasty RCTs and clinical trials. His extensive academic and clinical network will facilitate inter-institutional collaborations, potentially broadening the study’s scope and impact.

This grant’s dataset integrates multiple RCTs and clinical trials from several collaborating universities. Assembling this data offers a rare opportunity to juxtapose various surgical techniques and health metrics within a vast population. This model of gait assessment will be used to integrate with the whole body model of the disease³⁴ to develop of the findings of the best approached and lead into further large scale studies looking to review early osteoarthritis³⁵

Section 5 - Plan of investigation

Plan of investigation:

For the preparation and planning of the grant data holders will requiring bring together through a strategic data sharing agreement. Additional sites will also be sought to increase the numbers which will include internal collaborators.

Data amalgamation will begin at the MSK lab where a key group of researchers, clinicians, and technical experts are already assembled. Different protocols have been implemented at different sites. Amalgamating databases that have used different data collection protocols presents a range of challenges (collection, harmonisation, and translation).

The first collection task will be a review of data integrity & consistency. This will ensure that data is free from errors, duplicates, and outliers. Across 1000 participants this is expected to take around 3-6 months across the technical team at imperial. Missing data will initially be flagged where standard conservative imputation cannot complete datasets. It is anticipated that the machine learning model can cope with this missing data due to the dataset size.

Generation of a metadata review will collate equipment used, methodologies, variable mapping and any deviations from standard protocols. A finalised report will confirm the final expected outcomes from the dataset and how these will be calculated. This will include an internal review of the impact of different protocols on calculated variables.

It might be necessary to transform data from one or more datasets to make it comparable. This might involve mathematical transformations, re-categorization, or other techniques. An internal verification and validation review of the reported outcomes across the methodologies (eg marker protocol, walking speeds) will be reviewed to determine which translation maybe required. These will be finalised and report in for the final database.

The data will be used to build and refine the predictive models for post-operative function and other outcomes. Machine learning algorithms, along with the methods and techniques associated with them, will be used to determine the importance of factors in predicting post-operative outcomes. Based on evaluation results, models with be refined by revisiting feature engineering, adjusting hyperparameters, and ensemble methods. Upon plateauing of validation performances metrics, the results of the feature selection will interpreted to understand the model’s predictions by technique such as SHAP (SHapley Additive exPlanations) or LIME (Local Interpretable Model-agnostic Explanations). Through this the key research questions will be reviewed, and determination of key features required for larger scale data capture.

This grant also aims to tackle this challenge head-on by developing a rigorous and robust methodology that outlines best practices for hip function assessment in arthroplasty populations. The diversity of methods and technologies used in assessing hip function following arthroplasty presents both an opportunity and a challenge. An opportunity, because the range of tools and techniques, from motion capture to wearables, offers a rich landscape of data and insights. A challenge, because without standardization, comparing, and combining results across studies can be problematic, leading to discrepancies in findings and potentially hindering advancements in patient care. While the technologies and techniques may evolve, having a foundational methodology reported as part of this grant for best practice ensures that research remains grounded, relevant, and poised to make a real-world impact on the lives of patients undergoing hip arthroplasty.

The integration of research findings into clinical care will be a multistep, iterative process, across this collaborative network of surgeon, physiotherapist, GPs and engineers considering various factors like clinical workflows, training needs, and technological adaptations. One such development is the deployment of a markerless motion capture system for capturing of large number of people gait. Data capture at this scale represents an innovative approach to early detection and preventive care for conditions like osteoarthritis. This strategy can be instrumental in identifying the earliest signs of gait alterations which might indicate the onset of hip osteoarthritis. This expansion to this grant will allow the deployment to more locations or even integrating the system into workplace environments, gymnasiums, or recreational centres. This grant will lay groundwork to this application that has both clinical adoption and commercialisation opportunities. Further resources will be sought to expand on this future work.

The expansion of the database throughout and beyond the duration of this grant will align with other research initiatives by the team, including collaboration with the OATech+ network to create a centralized osteoarthritic gait database and platform for data sharing. This database will serve as an integral component of this broader data-sharing effort, marking the first instance of consolidating these types of datasets. Additionally, literature and scoping reviews of currently published databases will further contribute to this work. Centres will be identified and sought to contribute to this database where possible and work to establish future data collection points will be reviewed.

Regular updates will be sent to BHS and ORUK, as per grant requirements, to keep them informed about the research’s progress. Similarly, the group will disseminate results through publishing findings in reputable journals present results at conferences, workshops, or seminars.

Individuals with hip conditions are the core focus of our study. We will organize PPIE sessions, fostering a mutual exchange of insights between patients and our research team.

The amalgamation of datasets from various RCTs and clinical trials across collaborating universities, run the risk of poor integrity and consistency of data, given that different institutions might have employed varying data collection methods and standards. This disparity could compromise the reliability of the findings. With such a volume of information lies the challenge to ensure a comprehensive review of every subset and intersection. However, as data grows, so does the computational power and sophistication required to analyse it. Without the right infrastructural support, including advanced software and hardware capabilities, there is a risk of not harnessing the dataset’s potential, leaving behind essential insights or, worse, drawing incorrect conclusions due to inadequate analytical tools.

Section 6 - Research environment and resources

Research environment and resources:

This work will be hosted with the MSk Lab within the Department of Surgery and Cancer, Imperial College London. The MSk Lab focuses on advancing surgical technology and outcomes for patients with early osteoarthritis, collaborating with industry partners on new device development and post-operative clinical surveillance. Utilizing a multidisciplinary approach that incorporates computer science, imaging science, tribology, biomechanics, simulation, and surgical robotics, the lab is engaged in designing patient-specific tools and conservative surgical approaches, assessing the efficacy of new techniques, and developing technologies to improve surgical performance and outcomes.

The lab is located at our Sir Michael Uren Hub on our evolving White City campus. This campus draws together interdisciplinary expertise and research with the Uren Hub being a building cohabited by engineers, healthcare disciplines and clinicians and data scientists. This site provides a wealth of expertise and facilities for research with state-of-the-art motion capture suites including imaging resources, mechanical testing, and technical innovation. Imperial Hackspace resides on the campus and provides a range of workshop facilities and is linked to the College’s Innovation Rooms which allows regular interaction with, and hosting of events for our local community thereby enhancing our public engagement. Our local Imperial College NHS Trust is nearby, and we have links with several local GP practices.

Imperial College has an outstanding research environment with a variety of Centres and Networks of Excellence (e.g. Centre for Injury Studies, Musculoskeletal Medical Engineering Centre, Neurotechnology Centre, Rehabilitation Technologies Network, AI Network, and Trauma Bioengineering Network).

Imperial College offers high-performance computing clusters with specialized GPUs, ideal for the computational demands of AI algorithm development. The institution also provides robust data storage and management solutions, ensuring secure and compliant handling of sensitive medical data. Additional resources like specialized research labs, cloud computing, and high-speed networking facilitate efficient development and cross-institutional collaboration.

In order to create a comprehensive and robust dataset, we are leveraging a collaborative network of universities that specialize in musculoskeletal research and surgical techniques. As of now, University College London (UCL) has confirmed its involvement in principle. However, the scope of this initiative is to initiate conversations with several other leading universities in the field, both nationally and internationally, to participate in this project. The inclusion of multiple institutions will exponentially increase the diversity and scale of the data, offering a richer understanding of various surgical techniques and patient outcomes. Each participating university will contribute unique perspectives, specialized expertise, and subsets of data that will collectively amplify the quality and reach of our research. This network of universities is instrumental in ensuring the success of the grant, and we are committed to nurturing these relationships for the benefit of the broader scientific community and, most importantly, the patients themselves.

Section 7: Research impact

Who will benefit from this research?

This project aims to revolutionize healthcare by optimizing surgical practices and patient care. It seeks to develop predictive tools for pre-operative outcomes, thereby aiding in shared decision-making between surgeons and patients. This enhances not only clinical outcomes but also the healthcare experience, possibly leading to early interventions that improve patient quality of life and healthcare utilization. Additionally, the research will advance the realm of treadmill gait analysis. The introduction of new technologies and standardized protocols will not only enhance the quality of data but also make the research process more efficient, impacting power calculations and study design. This will benefit researchers in various disciplines, enabling them to conduct more rigorous studies. The project will disseminate its findings through targeted publications, serving as a resource for professionals across sectors. Overall, this research has the potential for far-reaching impact, from improving individual patient outcomes to influencing healthcare systems at large.

How can your research be translated in real-life?

The use of detailed functional assessments are not new, but in the past such assessment are limited due to duration of assessment, access to complex and expensive resources and studies limited by patients numbers. Through the collaborations realised in this work large datasets will be drawn together enabling complex analytics and data interrogation. The evolution of simple markerless motion capture using cameras and associated algorithms has the potential to translate into clinical decision-making tools in orthopaedic and other healthcare clinics. Once the markerless solution is proven effective, it will be leveraged to deploy predictive models via rapid and scalable gait evaluations. This will facilitate the concept of the right surgery and at the right time for the right patient. It has the potential to reshape how we think about joint health and its maintenance.  

How will your research be beneficial for ORUK and its purpose?

This grant aims to innovate hip patient care in the UK by refining surgical methods, risk stratification, and management of waiting lists, thereby accelerating patients’ return to work which directly aligns with the British Hip Society (BHS) and Orthopaedic Research UK (ORUK). We aspire to advance surgical technologies for early-stage osteoarthritis, fulfilling BHS’s goal of improving clinical practice. Our clinicians can then offer more personalized treatments, enhancing patient outcomes. In parallel, our comprehensive dataset, comprising multiple RCTs and clinical studies, aligns with ORUK’s mission to foster groundbreaking orthopaedic research. This database will be a robust platform for comparative studies, enriching understanding of surgical techniques and healthcare outcomes, and potentially fuelling further ORUK-backed initiatives. Moreover, our commitment to PPIE ensures a patient-centric-approach, congruent with both BHS’s and ORUK’s focus on patient-centred care. This multidimensional approach will contribute to healthcare policies, clinical practices, and patient well-being, furthering the aims of both organizations.

Section 8: Outreach and engagement

Outreach and engagement:

Individuals with hip conditions are the core focus of our study. Their feedback, experiences, and results will play a pivotal role in guiding our research. We plan to engage consistently with a diverse PPIE group, delving into their experiences before and after surgery, ensuring that their perspectives remain central to our investigative journey. This collaboration will create a platform for patients to relay feedback on their surgical experiences, enhancing our ongoing comprehension. As we begin to draw insights from our research, we will reconnect with the community to educate them about the most promising surgical approaches and the outcomes they can anticipate. We will organize PPIE sessions, fostering a mutual exchange of insights between patients and our research team.
Many patients are currently frustrated by slow routes to surgery for hip osteoarthritis and long waiting lists, particularly in recent years following the pandemic and doctors strikes. This work seeks to understand the impact of hip osteoarthritis on health and well-being seeking to identify the right care pathway, and right interventions at the right time. By working with our established PPIE group (currently 50 patients with a history of osteoarthritis) we will seek to identify appropriate routes of communication for messaging our work to ensure the questions we seek of our databases are important to our key stakeholders and that we seek to communicate these findings appropriately to a wide range of audiences from diverse social and ethnic backgrounds.
Dissemination will include an interactive study website designed with our PPIE that will serve to garner public interest in the project and its findings. This will be supported by public engagement outreach events such as the annual Great Exhibition Road Festival, which will seek to engage with our local populations to explore their understanding of joint health and the role of physical activity, and surgery in maintaining joint health. This event will also be used to grow our PPIE group to inform future work in this area. We will also explore how this information could be useful to the public in understanding approaches and timings of surgery with a view to understanding the viability of a clinical decision making tool used in partnership with their managing clinician, and as well as exploring their personal role in maintaining their joint health.

Section 9: Research budget

Requested funding from ORUK

University fees (if any)
£0

Salary
£101731.24

Consumables
£12000

Publications
£0

Conference attendance
£0

Other items
£

Total 'requested fund'
£113731.24



Other items




Other secured funds

Internal funding
£0

Partner (University)
£0

Partner (Commercial)
£0

Partner (Charity)
£0

Partner (Angel investors)
£0

Other sources
£0

Total 'other funds)
£0



List all the 'other funders' and explain how their funds are used to cover the costs of your research.
At present, no additional funding exists for this project. While efforts are underway to secure resources for developing a markerless motion capture system, this grant is envisioned to yield an independent predictive model with multiple applications. The allocated funds will enable Dr. Maslivec and Dr. Banger to focus on building the database. Additional budget is needed for consumables associated with the platforms that will host the machine learning algorithms.

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?