The Updated Modern Epidemiologist's Library
One of the first posts on this blog was a description of the core texts I felt were necessary in the library of the modern epidemiologist. I wrote this in 2015. [Side note: I cannot believe I have been writing this blog for so long.] Now that I am in my mid-career as a faculty member, I have further reflected on the books that have been seminal in my career and training, and those that resonate with my students as well.
For those just starting out in the discipline - especially those contemplating a career in academia - reading these texts provides both historical and contemporary accounts of the methods, frameworks, and theories of the discipline. For those seasoned in the field, these books remind us of the context and development of critical methodologies, many of which we take for granted (e.g., odds ratios to approximate relative risks in case-control studies).
I also find that as I continue to learn about newer and innovative approaches in epidemiology, I return to the core traditions of the field. After all, what separates epidemiologists from data scientists or statisticians, is our knowledge of study design, measurement, and causal inference, not to mention substantive expertise. Data tell what happened but not why: the epidemiologist looks for the why. Thus, re-visiting some of these texts helps ground our work in well-established principles.
Contemporary Texts
- Modern Epidemiology. Now in its 4th edition this is still considered by many to be the gold standard in the field with primary authorship passed to a contemporary generation of educators. It is by no means an introductory text, despite covering foundations in epidemiology. Rather this text is best suited to those individuals, especially academic, seeking advanced study in the subtleties, nuances, and complexities of rigorous epidemiological investigation.
- Epidemiology: Beyond the Basics. This book offers an excellent intermediate-to-advanced treatment of epidemiology. It is much more approachable than Modern Epidemiology, but consequently may give short shrift to important concepts, especially as it relates to bias. Nevertheless, if there is one book that I recommend to graduate students or professionals more often than any other, it is this one.
- Applying Quantitative Bias Analysis to Epidemiologic Data. I teach a course in bias and validity to my program's doctoral students and this has become my go-to text for treatment of these issues. Not only does this text define the traditional sources of bias in an efficient manner, but it also provides the methodological guidance for quantifying their impact on epidemiological analysis. If more researchers adhered to their recommendations, our science would benefit greatly.
- Causal Inference: What If. This book is the newest entrant to the advanced study of epidemiology, but based on the reputation of the two authors, is likely to be a classic in time. The authors provide clear explanations, sound advice, and guidance on methodology. As opposed to the other generalist epidemiology books (vis a viz. Modern Epidemiology), this book focuses on the science of causality. Thus, epidemiologists not engaged in causal inference, such as those in health departments who work descriptively, may not benefit from this book as much as etiologists.
Historical Texts
- Epidemiologic Methods. This book is important for numerous reasons. First, it was the first book on the methods of epidemiology. Second, it was written by a well-known epidemiologist, Professor Brian MacMahon, who chaired the department at Harvard University for many years, yet is also remembered for the controversial and flawed coffee and pancreatic cancer study. Third, it is still relevant even today and makes the reader appreciate the advanced thinking and foresight of the book's two authors. Sadly, it also makes you realize how difficult the science of epidemiology is, as we still grapple with the same issues brought up now 60 years ago.
- Causal Thinking in the Health Sciences. This is another one of those seminal books that has stood the test of time. As with Epidemiologic Methods, this book remains highly readable and relevant today. It is less concerned about the quantitative methods than Epidemiology Methods and is more concerned about the framework and theory behind epidemiology, for the purposes of identify etiologic pathways. In this way, it may be considered the historical predecessor to the contemporary Causal Inference: What If text.
- Evolution of Epidemiologic Ideas. This is an annotated collection of historic papers from the 1950s through the late 1970s. The papers included are considered by many to be the seminal papers in causal inference and epidemiologic methods from these decades. Many of the papers are as relevant and applicable to today's epidemiology as they were when written, many decades ago. The short commentary accompanying each paper is equally as interesting as the papers themselves.
- Eras in Epidemiology. This book is a both a summary and commentary on the history of epidemiology. As opposed to other historical texts that examine the development of epidemiologic theory and methods in the mid-20th century, this text takes a longer view of the history of epidemiology, with references from the Age of Enlightenment. In that respect is more like the classic text by George Rosen, A History of Public Health. But this text also tackles more contemporary (up to the 21st century) epidemiologic issues, such as moving beyond the risk-factor epidemiology paradigm.
- Theoretical Epidemiology. This text introduced many important advancements in case-control study design, as well as thinking about study sampling and measurement. Unfortunately, the author - a noted and some might say infamous epidemiologist - used his own language to define concepts and thus this book is quite challenging for modern readers.
- A History of Epidemiologic Methods and Concepts. Divided into two parts, this contemporary book provides a nice summary, as the name suggests, on the history of our methods. The first part is a summary of the evolution of the field by the primary author, while the second part is contributed chapters by leading epidemiologists reflecting on the history of epidemiologic issues, such as confounding, bias, case-control study design, and so on.
- Statistical Methods in Cancer Research. This two-volume set put out by the World Health Organization circa 1980 is an approachable while still exhaustive examination of methods in observational epidemiology for the case-control (Volume I) and cohort (Volume II) study designs. It was edited by the eminent Breslow & Day. These monographs illustrate how to carry out appropriate and valid statistical analysis for these hallmark study designs. A modern-day Volume III is said to be in preparation, which I look forward to reviewing at a later time.
Cite: Goldstein ND. The Updated Modern Epidemiologist's Library. Apr 25, 2023. DOI: 10.17918/goldsteinepi.
