Statistical Inference Of Bridge Study And Biosimilar Study In Clinical Trial

This dissertation is concerned with the bridging studies on extrapolating for-eign clinical data to a new region, and the assessing criterion and design of biosim-ilarity study in clinical trial.On the basis of Fundamental Bioequivalence/Biosimilarity Assumption, anoverview of statistical considerations including study design, criteria, and statisti-cal methods in terms of in vivo or in vitro bioequivalence testing will be provided.For in vivo bioequivalence testing, in addition to average bioequivalence, the con-cept of population bioequivalence and individual bioequivalence for addressingdrug interchangeability will also be discussed. For in vitro bioequivalence testing,an overview regarding some in vitro tests such as dose or spray content uniformitythrough container’s life, droplet and drug particle size distribution, spray pattern,plume geometry, priming and repriming, and tail of profile that are commonlyemployed for local action drug products. Recent development and future researchtopics will also be discussed.In clinical trial study, the variations of pharmaceutical products in efcacyand safety among diferent geographic regions due to ethic factors have become amatter of great concern for sponsors as well as for regulatory authorities. A bridg-ing study should be conducted in the new region to provide pharmacodynamic orclinical data on efcacy, safety, dosage, and dose regimen to allow extrapolationof the foreign clinical data to the population of the new region. On the basis ofsensitivity index, the shifts in population mean and population standard deviation can be classified into the three cases where (i) population mean is random andpopulation standard deviation is fixed,(ii) the mean is fixed and the standarddeviation is random,(iii) both of them are random. We derive the distribution ofthe clinical data in the new region, obtain the estimation of the mean, varianceand the sensitivity index in chapter2, and thus the similarity between the originalregion and the new region can be assessed by the estimated sensitivity index.Once more biosimilar products enter the market, the problem whether ap-proved biosimilar products can be used interchangeably and safely will be con-sidered. Most regulatory agencies require the evidence of average biosimilarityby conducting a bioavailability and bioequivalence trial in terms of the rate andextent of drug absorption. The assessment of interchangeability for biosimilars aregiven from two ways: switching and alternation.Based on the concept of switching and/or alternation several useful designs forassessing drug interchangeability are proposed, chapter3including: Balaam’s de-sign and two-stage design for the switching, two-squence dual design and William’sdesign for alternation, modified Balaam’s design, complete design, alternative de-sign and adaptive design for switching/alternation. On the biosimilarity index,the conceptions of switching index and alternation index are proposed to assessinterchangeability of approval biosimilars.The criterion of assessing biosimilarity is considered, we propose a conceptionof biosimilarity probability, obtain a frequency estimator and give its large sampleproperty. A test for biosimilarity based on the estimator of the biosimilarityprobability is derived. The power function and sample size calculation are shown. A simulation study is conducted to compare the proposed method and the twoexisting methods: the ratio estimator cretirion and the linearization cretirion.Finally we propose another scaled criterion to assess biosimilarity based onthe relative diference. The scaled relative diference criterion could be employeddue to the most of biosimilars have great variety, which is a common case in thebiological drug products. The power function and the calculation of sample sizeare derived in large sample. The large sample property is discussed. The powerand the empirical type I error probability are compared between the scaled relativediference criterion and the ratio estimator cretirion, the linearization cretirion bysimulation study.