The randomization-induced risk of a trial failing to attain its target power: assessment and mitigation

摘要: Health researchers are familiar with the concept of trial power, a number that prior to start is intended describe probability results will correctly conclude intervention has an effect. Trial as calculated using standard software, expected power arises from averaging hypothetical over all possible treatment allocations could be generated by randomization algorithm. However, in ultimately conducted, only one allocation occur, and corresponding attained (conditional on occurred) not guaranteed equal may substantially lower. We provide examples illustrating this issue, discuss some circumstances when issue concern, define advocate examination pre-randomization distribution for evaluating risk obtaining unacceptably low suggest use restrictions reduce risk. In trials randomize modest units, we recommend designers evaluate getting and, if warranted, modify algorithm