TitleA study of stagewise phase II and phase II/III designs for clinical trials
NameDong, Gaohong (author), Shih, Weichung Joe (chair), Moore, Dirk (internal member), Marcella, Stephen (internal member), Quan, Hui (outside member), Rutgers University, Graduate School - New Brunswick,
Bayesian statistical decision theory,
Decision making--Mathematical models
DescriptionClinical trials play vital roles in drug development. Traditionally, phase II and phase III studies are conducted separately. However, in the pharmaceutical industry there is a recent trend toward combining phase II and phase III in a seamless fashion (a so-called phase II/III clinical trial). To first understand traditional phase II clinical trial designs, we develop two two-stage single-arm phase II clinical trial designs: a Bayesian-frequentist design and a Bayes factor-based design. Both designs control frequentist Type I and Type II error rates. Then we develop a varying-stage adaptive phase II/III clinical trial design. In this design, in addition to traditional initial learning stage (phase II) and final confirmatory stage (phase III), we also consider whether there is a need to have an intermediate stage to obtain more data, so that a more informative decision can be made to advance the trial to the final confirmatory stage. With respect to adaptations, we consider dropping dose arm(s), changing the primary study endpoint, determining sample size, and early stopping for futility. We use an adaptive combination test to perform final statistical analyses. Under conditional distribution of p-values or combined p-values, we derive Type I error rate and statistical power for each decision path. By applying closed testing procedure, we control family-wise Type I error rate at nominal level of α in the strong sense.
NoteIncludes bibliographical references
Noteby Gaohong Dong
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.