Epidemiology: An Introduction

Rating:
Summary: An excellent introduction
Review: "Epidemiology: An Introduction" is a pleasure to read. My introductory epidemiology class at UNC-Chapel Hill used an advanced printing for our textbook. I find Rothman's writing to be excellent: clear and concise, but not "dumbed down." He explains complex concepts in a straightforward and accessible way. If you are taking an introductory epidemiology course, purchase this book and read it, even if it is not your required text! A layperson with some background in the sciences and an interest in epidemiology would find this a good read as well.

Rating:
Summary: simple and straightforward
Review: A great cut-and-dry introduction to epidemiology. Although I would prefer to see some of the mathematical justification, for the less interested it's a good start. Examples are plentiful and clear, although some topics are handled as asides which may break up the flow of reading. Topics such as study design and measurements of association and occurence are clear-cut. Issues such as causation and confounding were sparse - helpful to supplement with Rothman & Greenland's "Modern Epidemiology".

Rating:
Summary: A graduate student's salvation!
Review: I disagree with the previous reviewer. I do not find the author to be anti-statistics, rather anti-mindless use of statistics.

The previous reviewer missed the point on page 20: generalization in epidemiology. The point is that generalization is not purely within the statistical realm, but context dependent. In fact, the previous reviewer makes the author's point: if our basic understanding of the biology is that all humans are the same and that the differences between Houston and Chicago are neglible, then the study can be generalized to Houston. This conditional statement is the context and underlying biology. Changing our assumptions of the underlying biology (i.e., ethnic, age, gender composition matters and differ between the two cities) means that our belief in the generalization is different.

The point is that statistics is not interpreted in a vacuum. The author distinguishes between this concept of generalization, as opposed to a random sample of larger population, much like a sample survey. In a sample survey, it is a violation of random sampling to limit those humans to Chicago. For example, if I want to know who is going to win the next US Presidential election, I would not just survey humans in Chicago, because in this context I think that they would vote differently than humans in Houston.

Anyway, I think it is wrong to take the philosophy with a grain of salt. Attempting to train epidemiologists to think and to clarify Modern Epidemiology concepts are foci of the text.

Rating:
Summary: A good layman's introduction, but wrong on stats
Review: Rothman has an easy to read writing style. Additionally, he typically sticks to verbal - not mathematical - explanations, again making the book easy to read.

Be forewarned, though. Reviewers of some of his other books described him as anti-statistics and I'd have to agree with this characterization. Repeatedly he seeks to demonstrate that epidemiology is so much more than just statistics applied to medicine and that epidemiology often is not bound by the silly restrictions in statistics.

Example: He states that in statistics, inference is limited to the population from which the sample was drawn. If it's determined from a sample in Chicago, for example, that ionizing radiation causes cancer, there is no need to limit the inference to people just in Chicago - there is no need to conduct another study in Houston. Hence, inference in epidemiology can step beyond the bounds to which inference in statistics is limited.

Like most of his examples attempting to demonstrate his point, this one is faulty. The population we're interested in here is a biological population - humans. There is no difference between humans living in Chicago and those living in Houston, other than environment. If there are no environmental causes or iteractions, then he's right. No additional study is needed - but only because we believe there is no biological difference between humans in Chicago and Houston. Both are part of the same population.

Inference is always limited to those type of humans studied. How do we answer the questions: do men and women have the same cancer rates? or is the effect the same for different races? We get answers to these questions by sampling from these populations.

And, as an additional point, if we do believe there are environmental effects, then an additional study in Houston would be warranted; the populations would be different.

These types of errors are typical in this book. Overall, though, I'd recommend the book as an easy-going introduction. Just take his philosophy with a grain of salt.

Rating:
Summary: Pretty bad book if you ask me
Review: This is my far the most boring epidemiological text I have ever read. The author fails to correctly explain terms that he uses and displays a lot of arrogance on his part (i.e. "obviously this is blah blah blah" occurs 2-4 times in every page. This is supposed to be "an introductory text" and should help/facilitate learning of epidemiology for the beginning student. The author does not explain new terms in complete sentences but in a philosophical manner, has too many poor as well as good examples, and tries to elucidate mathematical logic with overly wordy descriptions. C'mon, just show us how it is used without making us read a whole page about it! After reading chapter 3 and 4, I can honestly say that I will NEVER read another book by Rothman again. That's my 2 cents.

Rating:
Summary: Thought provoking, direct, and well written
Review: This is the best epid primer based on my growing experience with epid texts. It is concise and uses good examples. I refer to it regularly. The text also is a helpful companion to Modern Epidemiology (Rothman and Greenland).

Rating:
Summary: Thought provoking, direct, and well written
Review: This is the best epid primer based on my growing experience with epid texts. It is concise and uses good examples. I refer to it regularly. The text also is a helpful companion to Modern Epidemiology (Rothman and Greenland).

Rating:
Summary: Best conceptual epi book published to date
Review: To my knowledge, this is the best and simplest textbook in epidemiology that has been published to date. I could compare it with practically all the 'mainstream' textbooks that I have been using, more or less recent, and more or less detailed textbooks.
If a potential reader has already some familiarity with epi this is not the best reference in terms of details (R&G "Modern Epi" is far better), but it can definitely be considered the best reference available for epi concepts. Ideal to refresh a tired epi mind.
If a potential reader had no prime in epi, this is an excellent prime. A warning: if you are looking for a text 'for dummies' filled with smart tips and mechanical shortcuts, this book is probably not what you are looking for. Actually the main 'side effect' of the book is that it forces you to 'think,' furthermore with your own head.
The book is simple but never poor; it focuses on concepts through examples, what a reader needs to begin mastering the subject. One of the main ideas that I found useful is that formulas are not the goal of epidemiology: they serve the concepts & the context that frames the problem or question.
When you read a chapter a second or a third time, later in your studies, you realize how carefully thought and written this book is: I believe it is difficult nowadays (in any field) to find a textbook so sound and well constructed. I would definitely recommend it to anyone interested in epidemiology at any level.