It is common in statistical analyses to perform two or more tests for a set of data. Then the problem of multiplicity arises. This article concisely describes the basic principles of multiple test procedures. Three fundamental methods for controlling the family-wise type I error rates are explained in a unified manner: 1) Bonferroni procedure, 2) simultaneous confidence intervals, 3) closed testing procedure.
This paper discusses various multiplicity issues arisen in clinical trials and possible statistical approaches to these issues. We especially stress the importance of the closed testing procedures (CTPs) in the setting of clinical trials: They include various modified Bonferroni procedures, e.g., step-down Dunnett procedure, hierarchical procedure, and Williams test. Moreover they can be even applied to adaptive designs in clinical trials. We illustrate the basic CTP procedures in detail.
Adaptive design is one of the most active research areas in clinical statistics for the last decade. This paper will give an introduction to the theory of two stage adaptive designs based on the pioneering articles by Bauer and Kohne (1994) and Proschan and Hunsberger (1995). As an application of Bauer and Kohne method, Bauer and Kiser method for two stage adaptive dose selection procedure in a dose response study is illustrated. Two actual examples of a confirmatory two treatments trial and a dose response study are briefly given.
This paper presents an application of an adaptive design to a randomized phase II clinical trial and discuss several practical issues that could occur when applying an adaptive design. The discussion is based on presentations with respect to adaptive designs, which were given at the 2006 Japanese Joint Statistical Meeting.
The ICH E14 guideline provides recommendation to assess QT interval prolongation and proarrhythmic potential of non-antiarrhythmic drugs in clinical studies. As there exist many statistical issues in the clinical evaluation of QT prolongation, electrocardiograms, background information of the guideline and QT interval correction methods are described for introduction. Because of the inverse relationship to heart rates, QT intervals are corrected for heart rates in order to obtain a variable which is independent of heart rate. Population-derived correction, subject-specific correction, and other correction methods are introduced. Assumptions of each correction method and its properties are discussed. Study design should be considered to collect appropriate data and estimate accurate heart rate correction formulae.
The thorough QT/QTc (TQT) study is a core part of the evaluation for effect of investigational drug on QT interval prolongation. The examples for the study similar to TQT study are presented in this article.
The ‘thorough QT/QTc study’ has a critical role in evaluating QT/QTc interval prolongation of a drug. In designing the ‘thorough QT/QTc study' it is important to consider cause of variation of QT intervals and risk factors of QT prolongation, such as reliability of QT interval measurement, diurnal variation, pharmacokinetics, pharmacokinetics-pharmacodynamics relationship, age, gender, electrolyte abnormalities, concomitant drugs, concomitant diseases, etc. Choice of parallel or cross-over design and sample size determination are also discussed.
Thorough QT/QTc study is suggested by the ICH E14 guideline as one of the core components in evaluating the effect of the study drug on QT interval. This article gives an explanation for the analysis of QT/QTc data from thorough QT/QTc studies, presents some statistical issues associated with the analysis and provides some recommendations on how to handle those issues.
In this short note, we discuss the body and its interpretation of evidence from confirmatory clinical trials. The crucial point is somewhat ambiguous and some different views seem to exist among investigators and even statisticians. Stress is given to reproducibility and two confirmatory trials.