The Imperial Clinical Trials Unit (ICTU) team had the pleasure of attending the International Clinical Trials Methodology Conference (ICTMC) 2024 in Edinburgh from 30th September to 3rd October. As one of the largest gatherings of clinical trials professionals, held every two years, this event provided an exceptional opportunity to connect with experts from around the world to exchange ideas and explore the latest developments in clinical trials methodology.
On the backdrop of a beautiful early Autumn Edinburgh, our team presented a series of talks and poster sessions [more details on our LinkedIn post]. It was an invaluable experience to showcase our work, receive feedback from other researchers, and gain fresh perspectives on emerging trends in the field.
With over 30 of us attending the conference from the Statistics, Clinical Data Systems, and Operations teams, we were likely one of the largest teams at the event. Following the event, we held a reflective session, discussing our key takeaways from the conference and exchanging our individual perspectives. We highlight our favourite takeaways here, providing a collective reflection on what we learned and how it may shape our future work.
Decentralised Clinical Trials
The spotlight session on decentralised clinical trials (DCT) hosted by Prof. Sofia Villar [University of Cambridge] was a favourite among the ICTU team. The structure of the session saw the opinions of the audience elicited via Slido on their perceptions around DCTs. Questions included DCTs ability to enhance trial recruitment and improve the quality of outcome measures. The session then addressed each question individually with fantastic talks from Dr Mia Tackney [University of Cambridge], Dr Mark Toshner [University of Cambridge], Dr Bethan Copsey [University of Leeds] and Prof. Marion Mafham [University of Oxford].
Exampled trials highlighting the benefits of decentralised methods including direct-to-patient invitation, recruitment and follow-up model to achieve high recruitment numbers and wide geographic reach were ACSEND (>15,000 randomised) and ASCEND Plus trials (targeting >20,000 patients). Critical to this is patient and public involvement (PPI), and what their experience can teach us in the design of DCTs. This has been documented in their Trials publication: Patient and public involvement and engagement in the ASCEND PLUS trial: reflections from the design of a streamlined and decentralised clinical trial.
Dr Mia Tackney’s talk regarding her paper on digital endpoints in clinical trials and eight key methodological challenges surrounding using them to their full potential was another standout. So much so that their BMC Medicine paper was chosen for our November Journal Club 2024 (Unleashing the full potential of digital outcome measures in clinical trials: eight questions that need attention | BMC Medicine | Full Text (biomedcentral.com)).
At the end of the session, opinions were re-evaluated and showed a remarkable improvement in the audience’s perception on the capacity of DCTs. The format highlighted the value of conferences in both stimulating discussions and helping the community share knowledge. Undoubtedly, we will be hearing more of DCTs in years to come.
Adaptive designs
Many trials at ICTU use the adaptive group sequential framework; so, Prof. James Wason’s [Newcastle University] talk on robust interim analysis for adaptive designs (ROBIN) was eagerly anticipated by many of the team. The talk summarised practical and accessible guidance on how to approach interim analyses in the setting of adaptive trials. Prof. Wason’s background work into this framework was rigorous. Development included a review of trials that utilised interim analysis, and qualitative research to understand what is required for the most effective implementation. Further, an excellent point was also made about the importance of involving members of the public in plans for interim analysis and ensuring these concepts are clearly understood. Plans are already in place to implement some of our learnings from this talk into several of adaptive trials currently in set-up!
Prof. Stephen Senn [University of Sheffield] presented a talk on the need for caution when using adaptive trial designs. Drawing from his experience as a Data Monitoring Committee statistician, Prof. Senn provided a critical perspective, particularly on response-adaptive randomisation (RAR). He highlighted the statistical risks of unrestricted changes to allocation ratios and emphasised the importance of critically evaluating new methods that alter the well-established trial designs.
A presentation by Prof. Ian C. Marschner [University of Sydney] on confidence-adaptive trials was insightful, based on his Statistics in Medicine paper: Confidence Distributions for Treatment Effects in Clinical Trials: Posteriors Without Priors. Prof. Marschner proposed a ‘confidence distribution’ that offers an alternative to the Bayesian posterior in a frequentist setting. The key takeaway is that confidence distributions offer a powerful method to represent the uncertainty of treatment effects in clinical trials without relying on prior distributions. This approach bridges the gap between frequentist and Bayesian statistics by providing posterior-like distributions based solely on observed data. Subsequently, we can achieve more informative interpretations of trial results, enhancing decision-making and inference in clinical research while maintaining the objectivity inherent in frequentist methods.
JAMA Summit Update Spotlight Session
“Is the Clinical Trials Enterprise Broken? How Can it be Fixed?”. This session facilitated engaging discussions on the general direction of clinical trials, and what improvements need to be made. One topic was choice of trial outcomes, and whether selected trial outcomes are suitable for determining the wellness of a patient. It is not always the case that outcomes are selected on this basis; rather, they are often chosen due to their relative simplicity, for example surrogate outcomes. However, these may not be outcomes that a patient would deem important to their overall well-being. A key takeaway from the session was a need for greater patient and public involvement during the set-up of the trial to help identify suitable outcomes to measure, and more complex study design and methodologies may be required to handle this.
New Statistical Methods
Dr Matthew Burnell [University College London] presented on power calculations in a large late-phase MAMS platform trial for Parkinson disease (EJS-ACT PD). They expanded the existing methods for the analytic calculation of power for mixed models to handle their primary outcome: the Movement Disorder Society-Unified Parkinson’s Disease Rating Scale (MDS-UPDRS), which is a comprehensive assessment consisting of 65 items divided into four components: I) Non-motor Experiences of Daily Living, II) Motor Experiences of Daily Living, III) Motor Examination, and IV) Motor Complications. In addition to this novel approach, we were introduced to insightful new arguments in favour of analytical power calculations over simulation given the obvious advantage of speed and accuracy when feasible.
Another interesting talk in this session was Dose-Response Analysis in Complex Interventions by Mollie Payne [PhD student at King’s College London]. Mollie walked us through her work on the AVATAR trial in which participants were randomised to either therapy sessions or supportive counselling. Treatment adherence was only assessed in one arm and the number of therapy sessions were treated as the dose. She then applied current methods from the literature to answer the question “how many sessions of therapy do patients need to reach optimal outcomes?” and introduced a novel alternative called the Principal Treatment Effects method. This allows for causal inference while avoiding the need for distributional assumptions for the dose-response function.
Estimands
It was fantastic to see a whole session of the conference dedicated to estimands. Within this session, we heard from Dr Brennan Kahan [University College London], highlighting the importance of using estimands to clarify precisely what treatment effect is of interest in a trial. He gave an excellent overview of material that is now freely available to all researchers that provides guidance on different estimators that can be used to implement an appropriate method of analysis for their chosen estimand; accessible from MRC-NIHR-TMRP Workshop: Methods of Analysis for Different Estimands (MADE).
Dr Philip Pallman [Cardiff University] discussed the importance of using estimands in adaptive trials to clearly formulate analysis objectives for interim analysis, as well as final analysis. We then heard from Prof. David Dunn [University College London] who presented a new estimand for non-inferiority active-control trials. This estimand enables much smaller sample size by using averted events (events that did not occur because of the treatment) instead of observed events when we have an active control. This was demonstrated to be highly effective and resulted in a more clinically intuitive interpretation, in addition to a 7-fold reduction in sample size.
We also enjoyed learning about estimands in other sessions of the conference. One memorable presentation stood out for both its creativity and content: Dongquan Bi [PhD student at University College London] electrified the lightning round by presenting their ideas in the form of a rap. You can listen to it here.
Patient and Public Involvement (PPI)
We saw a great poster presented by Dr Annabelle South [University College London] on best practice for sharing trial results with participants. The work was part of the SHOW RESPECT study conducted by Annabelle and her team at MRC CTU at UCL.
The SHOW RESPECT team aimed to investigate how results should be shared with participants. They conducted a Study Within A Trial (SWAT) to assess different ways this could be carried out, for example, a basic or enhanced webpage. The SHOW REPSECT framework was presented, which included a list of considerations to help trials teams plan for sharing results with participants. The considerations cover a wide range of aspects, from how to make the message accessible, to anticipating whether the news will be positive or negative. This was a valuable and comprehensive piece of research that we were delighted to see.
And finally, throughout ICTMC 2024, we found that nearly every session mentioned the importance of PPI. While applied trialists are familiar and experienced with how to involve patient and public members in applied trials research, it was interesting to see this interweaved throughout the conference (as well as having several dedicated sessions). The focus on involving PPI in the development of trial methodology represents a significant shift in how we are starting to undertake research in this area.
Additional highlights
We also enjoyed a thought-provoking Keynote presentation by Prof. Linda M. Collins [New York University]. The keynote described a new framework called the multiphase optimisation strategy (MOST) which utilises the clinical trial environment to better understand complex behavioural interventions. The objective is to use MOST to build, optimise, and evaluate digital health interventions within a randomised controlled trial. This is pertinent in the field of behavioural change, as many interventions are simply packaged together with little regards to how the components act in themselves or interact with other.
Another insightful session was the Lightning Talks, which featured around 20 different topics. Each presenter gave a brief, informative presentation of their work, making it a concise and engaging way to discover new ideas to explore further if you're interested.
We were also intrigued to learn about a fantastic resource called LIGHTS that has generated a collection of all methodology guides and so is a useful resource to know about for methods research in clinical trials.
Last but not least, the conference party was truly memorable, capping off an inspiring few days. This year’s celebration was held at the iconic National Museum of Scotland where we enjoyed an authentic Scottish experience, including bagpipes, traditional Scottish dancers, ceilidh dancing, and even a taste of haggis!
Thank you to all the organisers who made ICTMC 2024 such a thought providing and enjoyable event. As we look forward to the next ICTMC in Birmingham in 2026, we’re excited to bring back what we’ve learned to shape our own work and contribute to the future of clinical trials.
Authors: Rayka Malek, Martin Orr, Dr Suzie Cro and Dr Rachel Phillips
Acknowledgements: We gratefully acknowledge the valuable contributions of ICTU members in the development of this work, especially Prof. Victoria Cornelius, Annie Wright, Dr Leila Janani, Jack Elkes, Jack Message, Louise Cherrill and Dr Daphne Babalis.
Comments