Robust Response-Adaptive Randomization Design
share
In clinical trials, patients are randomized with equal probability among treatments to obtain an unbiased estimate of the treatment effect. However, response-adaptive randomization has been proposed due to ethical reasons, especially in rare diseases where randomization ratio is tilted to favor the better performing treatment. The substantial disagreement regarding time-trends in adaptive randomization is not fully addressed. The type I error is inflated in the traditional Bayesian adaptive randomization approach when time-trend is present. In our approach, patients are assigned in blocks and the randomization ratio is recomputed for blocks rather than traditional adaptive randomization where it is done per patient. We further investigate the design with a range of scenario for both frequentist and Bayesian design. We compare our method with equal randomization and with different number of blocks including traditional response-adaptive randomization design where randomization ratio is altered patient by patient basis. The analysis is done in stratification if there is two or more patient in each block. Having a large number of blocks or randomizing a few subjects into a block should be avoided due to the possibility of not acquiring any information from the block(s). On the other hand, response-adaptive randomization with a small number of blocks has a good balance between efficiency and treating more subjects to the better-performing treatment.
READ FULL TEXT