Book Review

Textbook of Pharmacoepidemiology

N Engl J Med 2007; 357:1669-1670October 18, 2007

Article

Textbook of Pharmacoepidemiology
Edited by Brian L. Strom and Stephen E. Kimmel. 498 pp. Chichester, England, John Wiley & Sons, 2006. $160 (cloth); $80 (paper). ISBN: 978-0-470-02924-4 (cloth); 978-0-470-02925-1 (paper).

This nearly 500-page book is a reduced and simplified version of Brian Strom's reference book, Pharmacoepidemiology (4th edition. Chichester, England: John Wiley & Sons, 2005), which is to the field what Harrison's and Braunwald's are to internal medicine and cardiology, respectively. Any person involved in pharmacoepidemiology will immediately recognize this book and acknowledge its list of editors and authors as a who's who of the field.

For nonspecialists, it might be useful to explain what pharmacoepidemiology is and why it is important. Historically, concerns about medicines have evolved from toxicity (between 1906 and 1938) to teratology (in 1962) to clinical trials of the efficacy and safety of drugs (in the 1970s and 1980s) and culminated in the present regulations about the authorization to market drugs. These regulations include the Investigational New Drug Application (IND), the New Drug Application (NDA), Good Manufacturing Practice, and Good Clinical Practice — as well as the practice of evidence-based medicine. The worry now emerging is that a drug might not have the same risks and benefits in real life as it was demonstrated to have in highly constrained clinical trials. Patients who actually use the drugs may differ from participants in clinical trials because of genetics, patterns of drug usage, concomitant medications, or diseases. Errors can also be made by prescribers and patients. Because of the small number of patients (usually 5000 to 10,000) included in premarketing studies, only frequent risks (higher than 1 in 1000) can be identified in this type of study. All others can be discovered only after the drug has gone to market.

It is thus necessary to look at postmarketing data to identify patterns of drug use, new risks, and adverse reactions and possibly to confirm benefits. An armamentarium of methods, each tailored to a specific task, is available to pharmacoepidemiologists as they attempt to reach these goals. These methods, which collectively define pharmacoepidemiology, have been developed over the past few decades and are changing as technology, regulations, and concerns about privacy evolve. They include spontaneous reporting and data mining, to identify rare new adverse drug reactions and to generate hypotheses, and cohort and case–control studies — most of which use large automated databases that link insurance data to hospitalization or prescriber databases — to test these hypotheses. An example of this type of study was addressed in the recent literature on nonsteroidal antiinflammatory drugs and myocardial infarction. Randomized epidemiology, consisting of large, simple clinical trials, is practiced when there is a strong risk of confounding by indication. Pharmacogenetics is another promising branch of the field. The proper pharmacoepidemiologist should be able to master — or at least be aware of — all these elements, and this textbook will help considerably.

Textbook of Pharmacoepidemiology is lighter, much less expensive, and easier to use than Pharmacoepidemiology. It is an overview of pharmacoepidemiology that includes general principles, methods, and clear, practical examples that will allow the novice to understand why a certain method or study design is appropriate to the circumstances. It covers the field comprehensively, including the future of the field. It also gives details of certain methods, such as the numbers of subjects needed in various study designs, in a series of very useful tables. However, the book will be a starting point for most readers. It will neither replace method-specific textbooks nor reveal the nuances of logistic-regression analysis or teach students to program database extraction software. It will not replace expertise or hands-on training, but it provides an excellent foundation for those entering the field.

My only major complaint is that the book is very much “U.S.-centric” and as such will be much more useful to students in the United States than to those in Europe and other parts of the world. From reading it, one might imagine that there is not much in the field of pharmacoepidemiology outside the United States — a common misconception. For instance, there is no mention of the United Kingdom's Yellow Card Scheme for reporting suspected adverse drug reactions or the original regionalized French systems for spontaneous reporting. There is no mention of European regulatory systems at a time when European Risk Management Strategy plans are being implemented and European authorities have the power to enforce post-authorization studies, whereas the U.S. Food and Drug Administration is still toothless. In addition, the European editions of the International Conferences on Pharmacoepidemiology have more abstracts and participants than their North American counterparts. Despite this limitation (and I admit to having a conflict of interest in this matter), this book is a solid foundation for those who wish to start in pharmacoepidemiology.

Nicholas D. Moore, M.D., Ph.D.
Université Victor Segalen, 33076 Bordeaux, France
nicholas.moore@pharmaco.u-bordeaux2.fr