摘要
The clinical trial based on the likelihood ratio test has many preferred statistical properties such as asymptotic and self-adaptive properties. Along with the constant MCID （minimum clinical important difference）, the conventional trial outcomes can be expanded into the strict, aggressive and conservative outcomes; it makes the interpretations of trial conclusion more flexible. Besides, it will show that the trial goal and trial outcome are different concept. The trial goal switching is not a statistical issue, but a clinical decision.
