Resources hub

Response adaptive randomisation

The gold standard of any trial is to randomly allocate participants to different treatment options. This is known as randomisation. To understand the true effect of an intervention, participants must be randomised to reduce the risk of biases that may influence the results. For example, having unequal numbers of patients with severe disease symptoms in one group versus another can bias the results. Randomisation helps to balance out these influencing variables across the trial groups, ensuring the groups are comparable. Having similar groups when you begin the trial means that you can more accurately judge the effectiveness of a treatment or intervention.

In traditional trials, randomisation ratios are typically 1 to 1. For every participant you randomise to the intervention, you randomise one participant to the control group where they receive either nothing (placebo) or the standard treatment for that condition.

In clinical trials that use response adaptive randomisation, you can adapt these ratios over time based on the accumulating trial data. This means that if one or more treatments appear to have better health outcomes, the ratio can be modified so that future participants in the trial are more likely to be randomised to those better treatments. Alternatively, if some treatments appear to be performing poorly, fewer future participants will be allocated to these treatments to minimise exposure to harmful or ineffective treatments.