Adaptive Designs in Clinical Trials


Improving adaptive design use in practice

Adaptive designs aim to make clinical trials more efficient by changing aspects of ongoing studies whilst ensuring that the statistical validity of the study is preserved. Possible adaptations may include adding and dropping treatment arms, changing the randomisation ratio, altering the dose or level of intervention administered, or focusing on a subgroup of patients most likely to benefit based on past trial data.

Barriers to adaptive design use have been identified, including lack of training opportunities and statistical support, availability of user-friendly software, funding and reliance on traditional and established practices. This project encompasses research focussed on:

  • Creating novel user-friendly software for adaptive designs

  • Developing simpler approaches to adaptive designs that require fewer assumptions and are more likely to be adopted intro practice

  • Improving knowledge transfer between statistical and clinical communities

Model-based early-phase trial design​

As healthcare becomes more personalised, we are seeing higher demand for complex and innovative designs in clinical trials. Many novel approaches to conducting dose-finding trials use advanced statistical modelling, requiring many design parameters to be specified in advance and extensive simulation studies to evaluate how well-suited a particular design is for our trial.

We are working on approaches to generate operating characteristics for complex model-based trial designs that require fewer assumptions and minimal or no simulations. We aim to improve the speed and quality with which early-phase trials are designed, and therefore increase the appropriate use of novel dose-finding methods in practice.