Approaches To The Adaptive Design Of Early Clinical Trials

In recent years,the research and development of innovative drugs in China is booming.In order to avoid the resource cost caused by ineffective treatment and minimize the sample size on the basis of ensuring the test efficiency,multi-stage adaptation design in early clinical trials arises at the historic moment.Based on ethical considerations,the trial can be terminated in time when the effect of the drug is not satisfactory,so as to protect the subjects from being exposed to the invalid test.In this paper,several single-group multi-stage designs suitable for early exploratory studies(phase I,IIa,II)are introduced in detail,which are used to preliminarily evaluate whether the experimental drugs meet the pre-set effective standards,so as to decide whether to continue the next confirmatory research.For the early clinical trial of binary endpoints,the Simon’s two-stage design,Lin and Shih’s adaptation design methods are applied to a single group and multi-stage phase I clinical trial.The results show that the type II error estimated by Lin and Shih’s adaptive design method is lower than the standard level,so the actual test efficiency is higher.At the same time,when the real type I error is close,the sample size of Lin and Shih is smaller,which is more practical than the two-stage design of Simon.In addition,in order to deal with complex endpoints at the same time,another Bayesian design BOP2 method is introduced,and the BOP2 method is applied to a nested end point phase II clinical trial.Through the simulation study,it is found that,compared with the ordinary Bayesian design method,the BOP2 square method can achieve higher testing efficiency,and when the drug is actually effective,the BOP2 method will have a lower probability of termination of the test and more practical.