With the rapidly growing development in rare disease drug, the adaptation of sensible clinical trial designs becomes the most challenging task to both investigator/sponsors and regulatory. Given the sample size limit of rare disease patient population, the conventional statistical designs and rules of clinical trials are not suitable anymore. Frequentist method based on historical control or crossover patient self-control have been considered to perform hypothesis testing in controlling type I/II error. However, the underline assumptions for these designs can be easily violated. Bayesian adaptive design, as been used more and more in early phases in drug development, has now been considered in confirmatory trials for rare disease clinical trials. In this talk we will go over these statistical considerations and illustrate with real and simulated examples.