Abstract:
|
In phase I oncology clinical trials, research has shown that in some instances, the traditional 3+3 design performs worse than the continual reassessment method (CRM) design or the modified toxicity probability interval (mTPI) design. This points to a lack of comprehensive evaluation for adaptive designs with flexible options, such as desired parameters setting. Thus, an interactive web application has been developed to find an appropriate adaptive design for conducting a real trial. The web application evaluates six different designs: 3+3 design, accelerated titration design (ATD), biased coin design (BCD), k-in-a-row (KIR) design, CRM design and mTPI design. Through simulation studies with a matched sample size, a comprehensive score is used to evaluate the performance of selected adaptive designs with desired parameters. The web tool provides interactive graphical user interface that allows users to easily conduct simulation and to assess the best design for meeting the primary objective of the proposed trials. The web can be accessed at http://cqs.mc.vanderbilt.edu/shiny/AdaptiveDesignS/
|