Multiple Testing Problems in Pharmaceutical Statistics

The print quality of this book is terrible. I returned the first copy thinking I just got a bad copy. This was not the case. It looks like the book was printed as the ink was running out. Some letters are only partially there. This is a problem for a mathematical text that uses Greek as well as Roman letters. Chapman & Hall Books should be ashamed.

This is a brand new text with applications to the pharmaceutical industry. Mutiple Testing problems are of great importance to the FDA in reviewing new drug applications and also in genetics and bioinformatics. This is evident in the many recent texts on the topic. This is another one of the fine texts in the biostatistics series that Professor Shein Chow edits for Chapman & Hall/CRC. The book editors Dmitrienko, Tamhane and Bretz are well known for their work in this area and in addition to editing and organizing the text they are also major contributors to the chapters. Tamhane, Hsu and Westfall have each written recent texts on multiple testing that include the modern advances over the traditional methods of Scheffe, Tukey and Dunnett and the classical text Simultaneous Statistical Inference by Rupert Miller. Tamhane is an academic, while Dmitrienko and bretz are internationally knwo statisticians who work in the pharmaceutical industry. The first chapter of the book provides the FDA perspective and is coauthored by an FDA statistician Mohammad Huque Chapter 2 provides a broad picture of the various multiple comparison methods and is an extensive chapter written by the three editors and Brian Wiens, Peter Westfall, James Troendle and Jason Hsu. It includes at the end a section on software implementation in SAS and R. This chapter provides the foundation for the rest of the book. Chapter 3 specializes to dose response problems that are important in phase II of clinical trial research in the pharmaceutical industry. In recent years pharmaceutical companies strive to add more and more indications on the label for a drug. In some cases this requires special phase IV trials but well-planned phase III trials that incorporate hierarchical step-down procedures can often be used to get several of the most important endpoints accepted. These multiple endpoint trials are becoming more and more common. Many of these methods control the familywise type I error rate. More general concepts such as false discovery rate make it possible to control the number of mistakes while testing a very large number of statistical hypotheses as often comes up with microarrays. Gatekeeping procedures which generalize the hierarchical stepdown procedures are the subject of Chapter 5. Chapter 6 covers the adaptive designs that are becoming more and more common and popular. Case studies are also included. Chapter 7 is the final chapter and it covers microarray experiments and pharmacogenomics.

This is a well-written book by the masters in the field of multiplicity testing in clinical trials. However, the examples described throughout the book are only described, they are not solved. A few solutions appear on the accompanying website, but this is not what one would want to pay $80+ for, and it does not help support the pedagogical value of the text. Perhaps Alex Dmitrienko should consider writing a workbook that could accompany the text and provide a more comprehensive collection of worked examples using SAS and R. RonLev