Statistics Used In Bioequivalence

Bioequivalence is one of the important problems in pharmacokinetics and has a great relationship with statistics. The research on bioequivalence is meaningful for the guarantee on the same therapeutic efficacy and safety between test(generic) drug and reference(innovator) drug, as well as the quality of the test drug. It plays a significant role in developing and evaluating new medicines. Average bioequivalence is the main method, which is certificated by the government in bioequivalence research. Many scholars have engaged in the work on common situation of this method, however, few references are concerned on some special cases. This thesis is organized as follows: several mathematical models about pharmacokinetics and statistical methods relating to bioequivalence problem are introduced. Then some reliable methods concerning on average bioequivalence in special cases have been proposed. In chapter 3, the problem of over-variability, which is often overlooked during evaluating the average bioequivalence of high variability medicines and a method to test the variability of the medicines has been recommended by using the Schwarz information criterion in change point. Comparison between two examples is given to test the reliability of this method. In chapter 4, an approximate algorithm about the Fieller's problem, when the variation is not equal, is given by using small sample likelihood inference method. This method is used to evaluate the average bioequivalence when the variation of the pharmacokinetical data between two medicines is not equal and with the help of computer we have got the approximate confidence interval. At the end , this thesis presents the method to compare bioequivalence by direct concentration-time curve. As a whole, this method represent the difference between the concentration-time curves of two medicines. It matches the definition of bioequivalence better than other methods do, although it may be much strict and looks forward to be developed.