Developing efficient perpetual platform trials to study multiple treatments and multiple biomarkers

Lead Research Organisation: University of Cambridge
Department Name: MRC Biostatistics Unit


Clinical trials are used to test the effectiveness and safety of new treatments. In recent years they have become more and more expensive and have high rates of failure. New technology means that a large amount of biological information (summarised by measurements called biomarkers) can be measured. Often this information is likely to tell us how much a patient will benefit from being given a treatment. To make best use of this information, new approaches to clinical trials are needed.
A new type of clinical trial design allows testing of the effectiveness of several new treatments simultaneously while considering that the effect may depend on biomarker measurements. A number of real trials are using this approach. Some of these trials will allow new treatments to be included as they become available. This has some benefits and makes the cost of developing and testing treatments lower. However the effect of doing this on important characteristics of the trial, such as the chance of incorrectly recommending a poor treatment for use in practice, is not well understood.
We will work on ways of doing these trials in the most appropriate way. This will include getting a good understanding of how the trials work. It will also involve developing the best approaches to when new treatments should be added and when treatments in the trial should be recognised as being effective.
Our team includes a wide range of expertise that will mean we can develop suitable methods and support their use in real trials. The developed methods will help patients on trials have a better chance of getting the most suitable new treatment. It will also mean better quality information comes from the trial and that future patients will be treated more effectively.

Technical Summary

The availability of affordable biotechnology has meant that many biomarkers can be measured to predict the effect of different treatments. Novel trial designs have been proposed to test the effects of multiple treatments in multiple biomarker subgroups. Several of these are currently being implemented in practice. Many of these trials will allow new treatments and biomarkers to be added as the trial continues for logistical and administrative reasons. We refer to such trials, in which treatments are continually added and removed, as 'perpetual'. Statistical methodology has been lacking for such trials and in particular for when treatments are testing in distinct biomarker subgroups.
We aim to develop methodology for improving several aspects of perpetual trials with multiple treatments and biomarkers. This includes: 1) extending current to allow optimal planning of enriched phase III studies from the results of a perpetual phase II study; 2) developing optimal decision rules for when to drop a treatment from a perpetual trial; 3) understanding the effect of adding in a new arm to a perpetual trial; 4) determining when a new arm should be introduced based on factors internal and external to the trial.
We will use motivating examples of trials that our group has links with in order to inform simulation studies. These trials represent a range of trial designs. In addition to motivating the work, links with these trials provides additional routes for dissemination of the methodology.

Planned Impact

In addition to academic beneficiaries, described in the 'Academic beneficiaries' section, this work will benefit a number of other stakeholder groups.
The first stakeholder group is researchers working in clinical trials. The work in this grant will lead to useful methodology and open-source software which will mean better trials can be conducted and more information is available on the benefits and drawbacks of conducting perpetual biomarkers.
The second group is clinicians, both those working in clinical trials and those who are not. Clinicians who are running clinical trials will have the benefits described above: new designs which are more ethical and will provide better information about which subgroups of patients benefit from the treatment. New methodologies develop will mean clinical trials that provide more information on the effect of treatments in subgroups will be run. In the longer term (within 10 years), this may lead to improved quality of information for a clinician deciding how to treat a patient in the clinic.
A third stakeholder is the pharmaceutical industry, which is increasingly interested in funding and contributing drugs to large scale platform trials. This research will be to their benefit as it will help ensure robust evidence is provided from this investment.
A fourth stakeholder includes funding bodies and. This research will provide useful evidence on the effect of adding in arms which will allow these groups to decide whether a proposed trial design is appropriate.
Most of these outcomes will be realised within the length of the grant. Because we are focusing on novel and efficient approaches, longer term outcomes could be contributing towards making the UK become a more attractive place to run these types of trials, meaning more investment from pharmaceutical companies and better quality research.


10 25 50
Description A platform clinical trial approach to the management of Mycobacterium abscessus complex
Amount $2,100,000 (AUD)
Organisation National Health and Medical Research Council 
Sector Public
Country Australia
Start 11/2018 
End 10/2023
Description Improving the robustness of complex and structured study designd for clinical trials
Amount £158,598 (GBP)
Organisation Biometrika Trust 
Sector Private
Country United Kingdom
Start 12/2018 
End 12/2021
Description Australian platform trial collaboration 
Organisation University of Queensland
Country Australia 
Sector Academic/University 
PI Contribution I was invited to be a co-investigator on a clinical trial submission to the Australian Medical Research Council. I provided input to the application. My role will be to provide methodology input to the design and analysis of the trial.
Collaborator Contribution Professor Claire Wainwright led the application and will be the chief investigator of the trial.
Impact Successful application to Australian national health and medical research council, with just over $2m funding.
Start Year 2017
Description Basket of Baskets trial 
Organisation Fundacio Institut d'Investigacio Oncologica Vall Hebron (VHIO)
Country Spain 
Sector Hospitals 
PI Contribution VHIO are leading a European wide Basket trial which will investigate novel targeted oncology treatments for different tumour types. Cambridge Cancer Centre are involved and I am providing the statistical input to this trial.
Collaborator Contribution VHIO have led the development of the trial and have successfully attracted several million euros worth of funding from Roche.
Impact Trial has not yet started so no outcomes yet.
Start Year 2017
Description Australian Adaptive Designs Courses 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact We ran three 2-day workshops on Adaptive Clinical Trials in Australia (Melbourne, Brisbane and Sydney). This was organised by the Australian Clinical Trials Alliance.
Year(s) Of Engagement Activity 2019
Description Talk at NIHR statistics group meeting 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact I was invited to give a talk on methodology research to the annual meeting of the NIHR Statistics Group
Year(s) Of Engagement Activity 2017
Description Talk at patients event 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Patients, carers and/or patient groups
Results and Impact I was invited to give a talk about novel design in clinical trials to a event focused on patients with polycystic kidney disease.
Year(s) Of Engagement Activity 2017
Description Talk to pharmaceutical statisticians meeting 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Industry/Business
Results and Impact I was invited to give a talk to a PSI meeting on Bayesian methods in trials.
Year(s) Of Engagement Activity 2018