Accelerating the use of efficient adaptive trial designs
NIHR (National Institute for Health and Care Research)
Funding, enabling and delivering world-class health and care research
Helen Parsons and Andy Metcalfe argue the case for researchers to design more adaptive trials, as part of a move towards more innovative and efficient research study designs.
The need for high quality clinical trials to be quick and efficient, as well as robust, has never been greater. During the pandemic there was an urgent worldwide need to find effective treatments quickly. As healthcare technology advances at pace, we need rapid answers to test innovations. We need to know if new treatments should be delivered widely or not. More than ever trials need to provide value for money.
Adaptive designs have the potential to make trials much more efficient, reducing costs and delivering benefits to patients faster than with traditional Randomised Controlled Trials (RCTs) and other research methods. They have been used in oncology and by industry for at least a decade and have become well accepted in those fields. There is a growing need for trialists to shorten timelines and bring new treatments to patients faster. The Covid-19 pandemic underlined this need and there have been several successful high profile adaptive designs, including the RECOVERY trial, that have highlighted the advantages of this approach. In another example, researchers behind the Covid-19 trial REMAP-CAP, have recently set up a new NIHR-funded adaptive platform trial to rapidly test treatments for flu.
However, adaptive designs are less common in other specialties such as orthopaedics, or surgery more generally. Follow-up periods can be long, and this brings particular challenges to delivering adaptive designs. Our study START:REACTS was a surgical trial funded by the Efficacy and Mechanism Evaluation (EME) programme - an NIHR and Medical Research Council (MRC) partnership, in response to a call for ‘Evaluation of new surgical procedures through the use of novel study designs’. We used an adaptive design that could be applied to the more difficult setting of a trial with a late primary outcome time point, in this case at 12 months.
Our experience of adaptive trials
Our trial found that an InSpace balloon, a device previously thought to be promising for use in shoulder surgery, was not beneficial for patients and might be worse for their care. We found a significant difference in the primary outcome in favour of the control group. We used a novel group sequential design (GSD) with two planned interim analyses, where we could assess if we had enough information from the data to make a decision about stopping the study early. The novel aspect of the study was that we used data at both early (before 12 months) and final (12 months) time points to strengthen decision making and improve efficiency.
We stopped the study at the first interim analysis, which shortened our study by around a year, even without accounting for delays due to the pandemic. The delivery of elective surgery in the UK has been strongly affected by the pandemic, so it is likely that without the adaptive design, these results would have been delayed by at least a further two years
We aimed to recruit 221 patients, but stopped after recruiting 117. Importantly, this means that over 50 patients avoided receiving non-beneficial treatment. Therefore, as well as shortening the time needed to evaluate new treatments, adaptive designs can also reduce the number of patients who are exposed to treatments which are not beneficial.
We also conducted a statistical study looking at the raw data from seven previous NIHR multi-centre trials in trauma and orthopaedics, to see if we could have used the same group-sequential approach. Five of the seven trials could have used our method directly, and with some simple changes, all seven could have done. Of the five studies we tested the method with, two of them would have stopped earlier (at least 12 months earlier) using our new approach.
Recommended by LinkedIn
Benefits of adaptive trials
If these innovative methods could be applied to many more trials, it could save costs. But more importantly, patients could benefit from earlier answers to important questions. In the setting of a pandemic and where elective services are prone to delays, adaptive designs may allow many more studies to complete and report than might otherwise have been possible. Where previously funding panels and clinicians may have lacked confidence in the acceptance of adaptive trials, they are becoming much more widely understood and accepted. Our studies have increased our confidence in this method greatly.
Increasing the use of innovative research designs is one of the key action areas in the vision for the Future of UK Clinical Research Delivery published earlier this year. This vision is being implemented by the UK-wide Clinical Research Recovery Resilience and Growth programme, which brings together partners from across the NHS, government, research organisations including NIHR, academia, regulators, medical research charities, industry, patients, and the public.
Innovative study designs have never been so widespread and so necessary throughout clinical research. Changing the way we design and deliver studies will allow us to build system resilience and increase system capacity, by freeing up NHS and research delivery staff to work on the activities that really need their involvement and skills.
We delivered START:REACTS with a team of clinicians, trialists, patient representatives and methodologists. By working closely together, we can change the way we design and deliver studies to benefit patients, healthcare workers and researchers. Patients will access better treatments faster, and methodologists will see their theories in action. Funders will see costs reduce and NHS and research delivery staff will be freed up to deliver more care and more research. This can only be a good thing.
Andy Metcalfe, Professor of Trauma and Orthopaedics, Warwick Clinical Trials Unit, Warwick University and University Hospitals Coventry and Warwickshire
Helen Parsons, Associate Professor of Medical Statistics, Warwick Clinical Trials Unit, Warwick University and University Hospitals Coventry and Warwickshire