A Markov decision process for response-adaptive randomization in clinical trials

In clinical trials, response-adaptive randomization (RAR) has the appealing ability to assign more subjects to better-performing treatments based on interim results. Traditional RAR strategies alter the randomization ratio on a patient-by-patient basis. An alternate approach is blocked RAR, which groups patients together in blocks and recomputes the randomization ratio in a block-wise fashion; past works show that this provides robustness against time-trend bias. However, blocked RAR poses additional questions: how many blocks should there be, and how many patients should each block contain? TRIALMDP is an algorithm that designs two-armed blocked RAR clinical trials. It differs from other trial design approaches in that it optimizes the size and number of blocks in addition to their treatment allocations. More precisely, the algorithm yields an adaptive policy that chooses the size and allocation ratio of the next block, based on results seen up to that point in the trial. TRIALMDP is related to past works that compute optimal trial designs via dynamic programming. The algorithm maximizes a utility function balancing (i) statistical power, (ii) patient outcomes, and (iii) the number of blocks. It attains significant improvements in utility over a suite of baseline designs, and gives useful control over the tradeoff between statistical power and patient outcomes. It is well suited for small trials that assign high cost to patient failures.(c) 2022 Elsevier B.V. All rights reserved.