Background: When using adaptive randomisation for randomised controlled trials (RCTs) of hyperacute stroke interventions, it is important to achieve balance in size of treatment groups to maximise statistical power, and in prognostic baseline covariates such as age and NIHSS score to avoid potential confounding. It is also important to achieve these balances while preserving as much randomness as possible in treatment allocation. Currently, common-scale minimal sufficient balance (CS-MSB) adaptive randomisation effectively controls for covariate imbalance between treatment groups while preserving allocation randomness but does not balance group sizes.
Aims: To extend the existing CS-MSB adaptive randomisation method to achieve both group size and covariate balance in hyperacute stroke trials.
Methods: Data from four hyperacute stroke trials were used to investigate the performance of the proposed adaptive CSSize-MSB algorithm under various conditions. A full factorial in silico simulation study evaluated the performance of CSSize-MSB in achieving group size balance, covariate balance and allocation randomness compared to the original CS-MSB method. A discrete-event simulation (DES) model created with AnyLogic was used to dynamically visualise the decision logic CSSize-MSB randomisation process.
Results: The proposed CSSize-MSB algorithm uniformly outperformed the CS-MSB algorithm in controlling for group size imbalance whilst maintaining comparable levels of covariate balance and allocation randomness in hyperacute stroke trials. This improvement was consistent across a distribution of simulated trials with varying levels of imbalance but was increasingly more pronounced for trials with extreme cases of imbalance. These results were consistent across different trial datasets that include a range of covariates and covariate types.
Conclusion: The proposed adaptive CSSize-MSB algorithm successfully controls for group size imbalance in hyperacute stroke trials under various settings and its complex logic can be readily explained to stroke clinicians using dynamic visualisation.