Platform Trials Workshop Summary

Director Adam Meehan was pleased to attend the Platform Trials Workshop hosted by Monash University and ANZIC CTG on Saturday May 7th. The workshop covered the key components and concepts of a platform trial, current platform trials, as well as when and how to integrate adaptive trial designs into your research.

What is an adaptive trial?

Unlike traditional randomised controlled trials, adaptive trials use available data at various points to revise allocation of patients to treatment arms which have demonstrated to be more effective and likewise minimise allocation of patients to less effective treatment arms. This is achieved using an adaptive algorithm and supported by data management software with adaptive randomisation capabilities to revise treatment arm allocation probabilities. The protocol must stipulate which data, and at what thresholds and frequency such interim analysis would be performed to revise allocation probabilities, and by how much the probabilities will change for a given observed effect.

What is a platform trial?

The platform trial is an adaptive trial which adds the possibility of the introduction of new intervention arms to the study as they become available and dropping existing intervention arms that are shown to be inferior. Interventions are grouped into independent domains affecting a specific target outcome (e.g. a domain of goal guided practice vs. standard care, plus a domain of competing antibiotic therapies). Multiple domains may be combined to form a composite treatment regime having within it multiple independent randomised interventions.

There also exists the potential to run platform trials continuously if there are sufficient competing therapies or interventions within a domain that could be cycled through as inferior interventions are dropped and replaced by others.

When and why to do adaptive platform clinical trials

Platform trials are best suited to comparative effectiveness trials with a large number of possible treatments, studies addressing conditions with high mortality or risk, study designs with high levels of uncertainty and/or complexity for optimal treatment in experimental arms (particularly those requiring multiple interventions), and high cost drugs / devices. The goal is a shorter time to a definitive result that can reduce mortality and increase effectiveness of care of patients entering the trial by reducing likelyhood being allocated to ineffective intervention arms.

The benefit of adaptive trials is most pronounced in trials where one treatment, or a treatment combination, proves to be significantly more effective.

Risks, concerns, issues

Some queries were raised by attendees on the bias of research using this method, particularly with the reduced sample size compared with traditional trials. Dr. Scott Berry of Berry Consultants reminded the audience that strategies to overcome this include explicit, pre-defined rules which mandate the acceptable power and accuracy for estimating treatment effects / effectiveness at each stage, as well as comprehensive modelling and simulations of trial performance during the trial design phase.

Reflection and where to from here

Reflecting on the public and professional acceptance of Platform Trials, Adam reflected that randomisation is a concept which is already challenging to explain to study participants, and this style of trial may more closely meet the expectations of participants joining at any stage of the trial to be receiving the best ‘known’ treatment. How to communicate this and build research to support the analysis of this novel trial design is a body of work we eagerly await.

We’re excited to currently be developing an adaptive randomisation module to support emerging Platform Trials in Australia and abroad. The potential gains for this style of trial design, particular in delivery of best possible patient care, will be a win for everyone.

Further reading:

Berry SM et al. (2016). A response adaptive randomisation platform trial for efficient evaluation of Ebola virus treatments: A model for pandemic response. Clinical Trials, 1740774515621721.

Berry SM, et al. (2015). The platform trial: an efficient strategy for evaluating multiple treatments. JAMA. 313(16):1619-1620

Angus, D. (2015). Fusing Randomized Trials With Big Data: The Key to Self-learning Health Care Systems? JAMA. 2015;314(8):767-768

Research Path develops fully customised and bespoke modules, including adaptive and novel randomisation methods, to compliment our clinical trial management & electronic data capture platform. If you are running an adaptive or novel trial please get in touch for a no obligation estimate.

Return to News