Primer of epidemiology / Gary D. Friedman

The stated purpose of this book is to provide health professionals with a concise overview of the concepts of epidemiology and to "bridge the gap" in communication between epidemiologists and clinicians. That this is a fourth edition indicates ongoing success. Dr Friedman uses examples of studies and problems that apply the principles put forth in each chapter. As the author aptly states, "Epidemiology is not a rapidly changing discipline." However, a basic grasp of the appropriate use and interpretation of epidemiologic studies is becoming increasingly important as advances in computing and information technology make health-related information easier to collect and analyze and thus more widely available. This latest edition of his book provides rich examples of epidemiologic studies that not only teach and engage the reader but also provide an appreciation for the history, successes, and pitfalls of epidemiologic studies. The first nine chapters cover basic measurements in epidemiology, methods, and types of study design. The information in these chapters is more than sufficient to give the novice insight into terminology, concepts, and the strengths and weaknesses of various types of studies. In Chapter 4, Friedman refers to the often-heard criticism, "Of course, this study was retrospective, so we cannot be confident of the findings." It would have been useful to help debunk this myth by walking the reader through the ways to evaluate the quality of such studies. More specifically, readers should be encouraged to look at issues such as bias and determining whether the misclassification that may result from retrospective studies is differential or nondifferential. In many cases, the limitations of such studies result in conservative estimates of the effects of an exposure under study (because of nondifferential misclassification as described on page 50). eng

The treatment of statistical associations is adequate but could be improved by emphasizing that there is nothing magic about "p<.05" and that evidence of a strong association in the absence of this p value should not be dismissed! In view of this common flaw in the interpretation of data, a concise summary about the balance between statistical power, precision and strength of associations, and the probability of falsely rejecting the null hypothesis could improve the reader's ability to interpret the meaning of statistical analyses. The introduction to multivariate analysis is well done. The various types and uses of multivariate techniques will probably be understandable to most health professionals who read it, which, in itself, deserves praise. Chapter 9, entitled "How to Carry Out a Study" is simply wonderful! If every researcher were to follow these practical and systematic guidelines, the quality of both medical and epidemiologic research would take a major leap forward. This chapter is a golden nugget that should be read and reread until ingrained in the mind of anyone embarking on a research project. The book ends on two additional high notes. Chapter 15 makes the case for interdependence of the practices of medicine and epidemiology. The final chapter is a superb review of the most important aspects of each chapter. eng