Outcome-based adaptive randomization design

 

  • Description: The Bayesian Covariate Adjusted Response-Adaptive Randomization (BCARA) which combines a Bayesian, an adaptive and biomarker classification approach aims to match patients with the most efficacious treatments by utilizing patient’s biomarker information becoming available during the conduct of the clinical trial. It is also considered as a response-adaptive randomization strategy as the allocation of the study population depends on the responses of previous outcomes. A partial least square logistic regression approach is conducted to determine adaptively predictive biomarker-defined subsets.

  • Application: This strategy may be useful in the explanatory phase II setting of the drug development.

  • Alternative names: No alternative names found for this trial deisgn
  • Methodology:

    1. Randomly assign the first  patients to the different treatment arms where J the number of different treatment groups and K the number of biomarkers. At least one response should be observed in each of the different treatment groups before moving to the Bayesian response adaptive randomization

    2. After each new individual has been enrolled in the study, predictive biomarker-defined groups are determined by utilizing a partial least squares logistic regression strategy (PLSLR) which can predict whether the patient can benefit from the treatment. The biomarker status is determined before the randomization.

    3. After the establishment of the biomarker status and biomarker-defined groups of each new individual, the individual is then randomly assigned into one of the treatment arms using a BCARA randomization.

    4. According to the results of the BCARA randomization the trial either stops or continues based on decision rules proposed by Eickhoff et al. (2010) [53]. The Bayesian covariate adjusted response-adaptive trial design has the ability to identify the biomarker-defined groups likely to  respond to a treatment but it does not control the Type I error and in order to ensure that the identified result is true, a Phase III study should be conducted

  • Adavantages and limitations:
Advantages Limitations

Ability to incorporate prior knowledge from biomarkers into the design.

Identification of the subgroups for which a particular experimental treatment is more effective.

Can result in reduction of the number of patients required when compared to alternative designs (i.e, non-adaptive trial designs).

Solves the issue of the incorporation of information of multiple and possibly correlated biomarkers.

The Type I error is not controlled in the traditional sense.

An independent Phase III study focused on the selected biomarker-defined subgroups is required to show that the identified promising result is definitely true.

 

 

  • References:

Eickhoff JC, Kim K, Beach J, Kolesar JM, Gee JR. A Bayesian adaptive design with biomarkers for targeted therapies. Clinical trials (London, England). 2010;7(5):546–56. doi: 10.1177/1740774510372657. View Article PubMed/NCBI Google Scholar

Tajik P, Zwinderman AH, Mol BW, Bossuyt PM. Trial designs for personalizing cancer care: a systematic review and classification. Clinical cancer research: an official journal of the American Association for Cancer Research. 2013;19(17):4578–88. doi: 10.1158/1078-0432.CCR-12-3722. View Article PubMed/NCBI  Google Scholar