ELSEVIER
BOOK REVIEWS
Statistics in Medical Research: Developments in Clinical Trials by Edmund A. Gehan and Noreen A. Lemak, Plenum Medical Book Company, New York, 1994,214pages, $39.50 Although not conveyed by the title, this book is devoted to the history of statistics as it applies to the development of the clinical trial. In it, the authors have set out to “describe the milestones in this development and how each generation of scientists built on the heritage and foundations laid by their predecessors.” It brings together the separate historical development of public health epidemiology and mathematical probability theory, culminating in the 20th century development of the academic field of statistics, and the growth of the application of formal statistical methods in clinical research. Stylistically, the book is organized as a long sequence of short biographies of influential scientists, beginning with early pioneers such as John Graunt, John Lind, and Thomas Bayes and ending with the mid-20th century leaders such as Fisher, Neyman, and Bradford Hill. Despite this approach the authors reject the “great man theory” in their preface and state their view that these prominent figures are essentially emblematic products of the collective scientific advances of their contemporaries. The first chapter deals with the very early figures, from Ancient Greece to the Enlightenment. It includes some fascinating examples of early clinical experiments and demographic studies that led to the development of actuarial methods. Chapter 2 covers primarily figures of the 18th century, a period in which scientific ideas were receiving increasingly greater acceptance. This chapter culminates in detailed sections on Pearson and Gosset (Student), heralding the origins of modern mathematical statistics. Chapter 3 focuses exclusively on Fisher, Bradford Hill, and the popularization of randomization. Chapter 4 catalogs the luminaries who developed biostatistics in the United States. The biographical approach is dropped in the last two chapters. Chapter 5 is a rather odd chapter in which historical details about the cancer cooperative groups are related and some more technical issues are discussed, including an extended defense of the authors’ unconventional views about the relative merits of historical versus randomized controls.The final chapter provides examples of clinical trials and of some landmark epidemiologic studies. The book would appear to be most suitable for students of medical statistics. It is a sign of the academic maturity of a field when historical texts are undertaken, and this one follows a few recent predecessors, including books on the early development of statistics by Stigler (The History of Statistics, Belknap Press, Harvard, 1986) and Porter (The Rise of Statistical Thinking, Princeton University Press, 1986), and also one by Silverman (Human Experimentation, Oxford University Press, 1985) that focused similarly on medical experimentation. Controlled Clinical Trials 17~176-179 (1996) 0 Elsevier Science Inc. 1996 655 Avenue of the Americas, New York, NY 10010
0197-2456/%/$15.00
Book Reviews
177
Overall, I found the biographical material of the early chapters to be entertaining and informative. Those chapters contain a wealth of useful quotes and anecdotes, and the style of presentation is quite engaging. The later chapters seem somewhat incomplete and overly specialized. These chapters would have benefited from some speculation about the current trends that may prove to be of historical importance. For example, is the development of academic departments of biostatistics separate from statistics departments a positive trend? Some of the most influential forces in medical statistics today derive from clinical epidemiology groups, whose faculty do not have classical training in statistics. What does this trend bode for the traditionally trained statistician, and does it represent an indictment of current training in statistics? The computer revolution has increasingly placed statistical analytic tools in the hands of nonexperts. What is the proper response of the statistics profession to this trend? In clinical trials specifically, the vast industry that has evolved in drug regulatory activities and the influence of the Food and Drug Administration would also have been worthy of extended discussion. Finally, an entertaining aspect of the book is the catalog of photographs of the “great men” (and two great women, Florence Nightingale and Gertrude Cox). These include one of an elderly Wilcoxon on a motorbike. The book makes easy and informative reading, and I enjoyed it.
Memorial Sloan-Kettering
Colin B. Begg Cancer Center New York
SSDI 0197-2456(95)00100-U
Exploring Collaborative Research in Primary Care, edited by Benjamin F. Crabtree, William L. Miller, Richard B. Addison, Valerie J. Gilchrist, and Anton J. Kuzel, Sage Publications, 1994,326 pages, $55.00 Exploring Collaborative Research in Primary Care describes the topics and proceedings of a 1993 conference on multimethod research, which aimed to develop and further “a collaborative, cross-disciplinary research community in primary care. ” The central tenet of the book is that no single discipline possesses the methods needed to answer the questions of greatest importance in primary care. Support was provided by the Agency for Health Care Policy and Research, whose mission is to promote better clinical practice and improved patient outcomes, and to improve health care financing, organization, delivery, and access. The conference focused on “changing provider behavior,” a topic laid out by Schwartz and Cohen in a 1990 article of the same name [I]. For those who may not be aware, the publication of clinical trial results is neither necessary nor sufficient for achieving better provider behaviors and improved patient outcomes. As the book sensibly points out (p. 55), you shouldn’t seek to ask “Is primary care beneficial for patients and the medical field” without first asking “What is primary care?“ The book identifies four related but distinct meanings: (1)a person trained in one of three generalist “specialties”- family medicine, general pediatrics, or general internal medicine; (2) anyone who self-identifies as a primary care provider, (3) a setting, such as a community health center, where a particular type of medicine is practiced; or, (4) a specific level of care that is part of a larger