Implementing Optimal Allocation in Sequential Binary Response Experiments Export

@article{Tymofyeyev2007, abstract = {For sequential experiments with K treatments, we establish two formal optimization criteria to find optimal allocation strategies. Both criteria involve the sample sizes on each treatment and a concave noncentrality parameter from a multivariate test. We show that these two criteria are equivalent.We apply this result to specific questions: (1) How do we maximize power of a multivariate test of homogeneity with binary response?, and (2) for fixed power, how do we minimize expected treatment failures? Because the solutions depend on unknown parameters, we describe a response-adaptive randomization procedure that ” targets” the optimal allocation and provides increases in power along the lines of 2–4\\ over complete randomization for equal allocation. The increase in power contradicts the conclusions of other authors who have explored other randomization procedures for K = 2 and have found that the variability induced by randomization negates any benefit of targeting an optimal allocation.}, author = {Tymofyeyev, Yevgen and Rosenberger, William F. and Hu, Feifang}, citeulike-article-id = {3462412}, citeulike-linkout-0 = {http://dx.doi.org/10.1198/016214506000000906}, doi = {10.1198/016214506000000906}, journal = {The American Statistician}, keywords = {adaptive, allocation, biased, clinical, coin, design, doubly, file-import-08-10-29, neyman, power, randomization, response-adaptive, trials}, month = {March}, number = {477}, pages = {224--234}, posted-at = {2008-10-29 15:37:15}, priority = {0}, title = {Implementing Optimal Allocation in Sequential Binary Response Experiments}, url = {http://dx.doi.org/10.1198/016214506000000906}, volume = {102}, year = {2007} }

TY - JOUR ID - Tymofyeyev2007 L3 - citeulike-article-id:3462412 N2 - For sequential experiments with K treatments, we establish two formal optimization criteria to find optimal allocation strategies. Both criteria involve the sample sizes on each treatment and a concave noncentrality parameter from a multivariate test. We show that these two criteria are equivalent.We apply this result to specific questions: (1) How do we maximize power of a multivariate test of homogeneity with binary response?, and (2) for fixed power, how do we minimize expected treatment failures? Because the solutions depend on unknown parameters, we describe a response-adaptive randomization procedure that “targets” the optimal allocation and provides increases in power along the lines of 2–4\\ over complete randomization for equal allocation. The increase in power contradicts the conclusions of other authors who have explored other randomization procedures for K = 2 and have found that the variability induced by randomization negates any benefit of targeting an optimal allocation. IS - 477 JF - The American Statistician EP - 234 TI - Implementing Optimal Allocation in Sequential Binary Response Experiments VL - 102 SP - 224 KW - adaptive KW - allocation KW - biased KW - clinical KW - coin KW - design KW - doubly KW - file-import-08-10-29 KW - neyman KW - power KW - randomization KW - response-adaptive KW - trials AU - Tymofyeyev, Yevgen AU - Rosenberger, William AU - Hu, Feifang PY - 2007/March// UR - http://dx.doi.org/10.1198/016214506000000906 ER -