Sample Size Determination for Response‐Adaptive Randomization With Recurrent Event Responses and Unequal Follow‐Up Time

Trials with recurrent event responses are frequently encountered during clinical trials. Benefits have been demonstrated in previous studies when trials were conducted with a doubly biased coin design (DBCD). However, there is a lack of discussion about sample size determination. In this paper, we provide the methods to compute the required sample size with a given test power when DBCD is used. More importantly, our proposed sample size determination procedure also allows unequal follow-up time due mainly to possible patient dropout. Without considering patient dropout, it is shown that the required sample size to reach the prespecified test power is underestimated. Theoretical results are derived to measure the impact of dropouts on the required sample size and treatment allocation proportions. A simulation study is conducted to illustrate the usefulness of our proposed procedure. Finally, a clinical example is used to illustrate the advantages of our proposed procedure in practice.