Sample size calculations in clinical research.

The specific goal of this (2nd edition) book continues to be to provide a comprehensive and unified presentation regarding sample size calculations in a variety of clinical research and development phases and to serve as a solid reference for researchers in academia, industry, and government. The book has been expanded, by around 100 pages, from its original 12 chapters to its current 15 chapters. All original material from the 1st edition exists (and appears to be completely unchanged) in this 2nd edition. New material has been added as follows: (i) Two entirely new chapters have been added, Chapters 12—Microarray Studies, and Chapter 13—Bayesian Sample Size Calculation; (ii) a section from the 1st edition has been expanded into its own chapter, Chapter 11—Dose Response Studies; and (iii) two new sections have been added to the last chapter, Chapter 15—Sample Size Calculations in Other Areas. It’s two new sections are entitled, respectively, “QT/QTc Studies with Time-Dependent Replicates” and “Propensity Analysis in Nonrandomized Studies”. Hence, for a review of Chapters 1 through 10, 14 (11 in 1st edition), and Sections 3 through 6 of Chapter 15 (Chapter 12, Sections 2–5 in 1st edition), one should refer to the original book review (Filloon, 2004). A review of the new material in this 2nd edition follows, including summary comments of which some are carried over from the original book review. Chapter 11 provides sample size calculations for dose-response studies under a variety of scenarios. Whereas the 1st edition discusses only the two topics of linear contrast testing and minimum effective dose (MED) estimation for continuous responses, the 2nd edition extends linear contrast testing to include binary and timeto-event (i.e., survival) responses also. Furthermore, Cochran-Armitage trend testing for proportions is now addressed as well as extensive discussion of several types of phase 1 dose-escalation trials for toxicity estimation (i.e., maximum tolerable dose, MTD). In Chapter 12, microarray studies are addressed with emphasis on the issues associated with multiplicity. Specific direction per sample size is given on how to deal appropriately with false detection rate (FDR) and family-wise error rate (FWER). Hence, guidance is given on how to effectively deal with multiplicity adjustment at the design (i.e., sample sizing) stage. Chapter 13 shows how to determine sample size requirements when addressing oneand two-sample problems from a Bayesian perspective. Both confidence