chapter  1
42 Pages

An Overview of Adaptive Randomization Designs in Clinical Trials

WithOleksandr Sverdlov

Randomization Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.4 Response-Adaptive Randomization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

1.4.1 Response-Adaptive Randomized Urn Models . . . . . . . . . . . 14 1.4.2 Optimal Response-Adaptive Randomized Designs . . . . . 15 1.4.3 An Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.4.4 Bayesian Adaptive Randomization . . . . . . . . . . . . . . . . . . . . . . 19 1.4.5 Criticism of Response-Adaptive

Randomization Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 1.5 Covariate-Adjusted Response-Adaptive Randomization . . . . . . . . 22

1.5.1 Treatment Effect Mapping and Urn-Based CARA Randomization Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1.5.2 Target-Based CARA Randomization Designs . . . . . . . . . . . 23 1.5.3 Utility-Based CARA Randomization Designs . . . . . . . . . . . 24 1.5.4 Bayesian CARA Randomization . . . . . . . . . . . . . . . . . . . . . . . . 24

1.6 Other Designs with Elements of Adaptive Randomization . . . . . . 25 1.6.1 Randomized Phase I Trial Designs . . . . . . . . . . . . . . . . . . . . . . 25

1.6.2 Adaptive Optimal Dose-Finding Designs . . . . . . . . . . . . . . . . 25 1.6.3 Randomized Designs with Treatment Selection . . . . . . . . . 26 1.6.4 Group Sequential Adaptive Randomization . . . . . . . . . . . . . 27 1.6.5 Complex Adaptive Design Strategies . . . . . . . . . . . . . . . . . . . . 28

1.7 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

A randomized, placebo-controlled, double-masked, equal-allocation clinical trial can be viewed as an exemplary research design to obtain generalizable results on the treatment effect. However, modern clinical trials are increasingly complex and often require more elaborate designs. A competitive landscape of pharmaceutical research and development, an enormous number of molecules that are available as potential drugs, and limited patient resources call for clinical trial designs investigating the effects of multiple treatments within multiple patient subgroups. Such designs should, in addition, satisfy strict regulatory requirements such as controlling the chance of making a type I error.