Biostatistics (2nd ed.).

Biostatistics (2nd ed.), Gerald VAN BELLE, Lloyd D. FISHER, Patrick J. HEAGERTY, and Thomas LUMLEY, Hoboken, NJ: Wiley, 2004, ISBN 0-471-03185-2, xi + 871 pp., $110.00.

The first edition (1E) of this book (Fisher and van Belle 1993) was a massive 1,000-page comprehensive introduction

to the statistical methods useful for biostatistical applications. Despite the 1E's very traditional style of presentation and its lack of any focus on statistical computing, my report (Ziegel 1994) declared it to be the perfect book for statisticians who might be considering a career change from heavy industry to the world of pharmaceuticals, health, and medicine. I certainly would consider updating a 1,000-page book a daunting task. The authors apparently agreed, so two new coauthors, experts in "all things modern and statistical," were found to give "a twenty-first-century thrust to the book" (Preface, p. xi).

The book was decidedly difficult for the usual examination of the contents listing for the assessment of the changes. The table of contents provides only the 20 chapter titles. These reveal new chapters on longitudinal data analysis and randomized clinical trials (RCTs), very appropriate additions for a book subtitled "A Methodology for the Health Sciences." Sections for the longitudinal data chapter are exploratory data analysis, derived variable analysis, impact of correlation on inference, regression methods, and missing data. Topics for the RCT chapter include studies in humans, planning for an RCT, and designing an RCT. There also is a new final chapter, "Personal Postscript," whose topics vary from "Is There Too Much Coronary Artery Surgery?" to "Are Technicians as Good as Physicians?". The back cover also lists Bayesian statistics, generalized estimating equations, and the bootstrap method as additional new topics for the book.

This second edition of the book has fewer pages than the 1E. This results in part from a new (8 X 10)-inch format. The book is still very heavy. In addition, most of the tables from the 1E plus some other material have been shifted from the book to the website for the book. I described the full content for the 1E in my earlier report (Ziegel 1994). Despite the absence of any consideration for statistical computing, this comprehensive book should continue to merit consideration by anyone looking for a desktop reference in biostatistics.

REFERENCES

Fisher, L., and van Belle, G. (1993), Biostatistics, New York: Wiley.

Ziegel, E. (1994), Editor's Report on Biostatistics, by L. Fisher and G. van Belle, Technometrics, 36, 231.