Files
Abstract
The conventional clinical trials are usually conducted in one region due to same or similar internal and external factors. However, it has many shortcomings and risks. In order to compensate for potential risks of the single region and broaden the application of clinical trials, Multi-Regional Clinical Trials (MRCT) are becoming more and more popular. In addition to that, due to the uncertainty of treatment effect, its not appropriate to arbitrarily set a fixed value. In practice, there is considerable uncertainty about the true effect, and power cannot reflect the true probability to demonstrate a positive outcome. Thus, assurance is an alternative to the power when the effect size is considered as an unconditional parameter which follows a prior distribution. In this thesis, we aim to incorporate the idea of assurance into MRCT, derive mathematical formulas for both the fixed effect model and random effect model, and simulate the models.