Multiple analyses in clinical trials. Fundamentals for investigators (Q1407109)

This monograph of 436 pages, organized in 14 chapters and 5 appendices, is an essentially nonmathematical discussion of the problems posed by the execution of multiple analyses in clinical trials written for advanced medical students, clinical investigators and biostatisticians. Only a basic background in health care and introductory statistics is required. Chapter 1, Fundamentals of Clinical Trial Design, reviews fundamentals of clinical trial design and analysis required for the multiple analysis issue. Chapter 2, Multiple Analyses and the Random Experiment, describes the difficulty of interpreting clinical trial results when the prospective analysis plan has been altered. Chapter 3, The Lure and Complexity of Multiple Analyses, presents the motivations for, and consequences of, executing multiple statistical analyses in clinical trials. The need to minimize the type I error rate in a clinical trial is developed. Chapter 4, Multiple Analyses and Multiple Endpoints, develops how one chooses and analyzes multiple endpoints in a two-armed clinical trial using the Bonferroni multiple comparison procedure. Chapters 5 and 6, on Multiple Dependent Analyses, develop the concept of multiple hypotheses testing in a clinical trial when the primary analyses are related to each other. This includes allocating a type I error level across a collection of dependent hypothesis tests principles. Chapter 7, Introduction to Composite Endpoints, applies the multiple analyses methodology to specific, complex circumstances of the composite or combined endpoint as a primary analysis variable. Chapter 8, Multiple Analyses and Composite Endpoints, provides examples of the use of combined endpoints, unequal allocation of type I error and measures of dependency between statistical hypothesis tests. Chapters 9 to 11, on Subgroup Analyses, discuss their role and interpretation in clinical trials. A distinction is drawn between a subgroup analysis that compares the effect of a randomly allocated intervention in a clinical trial across subgroup strata on the one hand, and the analysis which only seeks to confirm the efficacy of therapy within a single subgroup stratum on the other. Confirmatory analyses procedures are described that allow subgroup evaluations to make determinations of endpoints, endpoint event rates, efficacy levels, and the precision of the endpoint measurement for the analysis of the effect of therapy within the subgroup stratum of interest. Chapter 12, Multiple Analyses and Multiple Treatment Arms, elaborates upon the ability of investigators to draw confirmatory conclusions from a clinical trial that consists of a control group and multiple treatment arms. Chapter 13, Combining Multiple Analyses, shows how the use of differential allocation, combined endpoints, prospective requirements of different levels of efficacy, and confirmatory subgroup analyses can be prospectively blended into effective combinations. The volume ends with Chapter 14 on Conclusions: The Two-Front War, five appendices (case reports and causality, estimation in random research, Federal Regulations of the US, sample size and advanced results) and an Index.