zbMATH — the first resource for mathematics

The design and analysis of sequential clinical trials. 2nd ed. (English) Zbl 0747.62109
The Ellis Horwood Series in Mathematics and its Applications. New York etc.: Ellis Horwood. 298 p. (1992).
This book is directed mainly to medical statisticians. There are some mathematical derivations, for example the properties of the test statistics, but the topics are generally developed by example. The majority of the book describes sequential designs for discriminating between two treatments. These are obtained by using a frequentist approach and assuming equal numbers on each treatment between inspections of the data.
The general approach, termed the boundary approach, includes group sequential up to fixed sample size designs and variable inspection times. It also allows a variety of response variables such as ordinal, binary and normal as well as survival data. Extension to more than two responses is discussed and alternative methods such as a Bayesian approach and repeated significance testing are described. Reference is made to the PEST 2 (Planning and Evaluation Sequential Trials) computer package, available from the author, which implements the designs. Some background is included on the planning of clinical trials, patient allocation and measurement of treatment difference. Details are given of several trials which have used the methodology.
Reviewer: P.W.Jones (Keele)

62P10 Applications of statistics to biology and medical sciences; meta analysis
62-01 Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics
62L05 Sequential statistical design
62L10 Sequential statistical analysis