Grant: Discovery Grants Program - Individual

A randomized clinical trial is an efficient tool to draw a valid conclusion on a new medical intervention to benefit future patients outside the trial (collective ethics). Balanced randomization is considered as the gold standard in clinical trials. However, this randomization raises ethical concerns because half of the participants are allocated to the inferior intervention and individual ethics is sacrificed. According to the personal care principle, it is difficult or impossible to implement randomization in a clinical trial considering the well-being of patients in the trial (individual ethics). Response adaptive designs are proposed to alleviate the tension between the collective and individual ethics. With a response adaptive design, the probability of allocating the next patient to treatments is modified based on the cumulating information on previous treatment allocations and responses of previously treated patients in the trial. The purpose of the design is to deliberately bias the treatment allocation in order to assign more patients to the potentially better treatment. The modification of treatment allocation creates a dependency in the data and the traditional statistical analysis method for independent data cannot be applied without justification. The dependency in data presents a great challenge in searching for an optimal adaptive design. The proposed research will investigate how response adaptive designs affect the power of statistical tests and propose an optimal response adaptive design in terms of a better way to deal with the tension between individual and collective ethics. Statistical analysis methods will be developed for the optimal design. The potential significance of the research is that the proposed design may serve as a reference design for adaptive clinical trials and may help promote the use of response adaptive designs in practice.