Book Reviews A Dictionary of Epidemiology

When glancing at the cover of the sixth edition of A Dictionary of Epidemiology, the first question that came to mind was whether a new edition of the dictionary was really necessary in an era when everything or almost everything can be found on the internet. It is a difficult question: instead of discussing whether a book is good or not, we debate whether it makes sense to have it in the first place. I will not hold back: a dictionary is absolutely necessary and it is a pleasure to hold the new edition and browse through it. Miquel Porta, editor of the fifth and sixth editions, discusses precisely this issue in the preface (very entertaining: recommended reading!) and suggests that ‘[the dictionary] can be more relevant and useful than ever before because nowadays we suffer from an unprecedented level of air pollution, noise and potential confusion’. We need an authoritative reference, even if we all know that definitions are never definitive and are often controversial. Indeed, being able to offer different views and modify definitions over time is a sign of authoritativeness. However, my view may just be outdated and isolated, so I asked 11 close colleagues, including senior and junior epidemiologists and biostatisticians, what source they would use for the definition of epidemiology when preparing teaching material. It was an open question, so they could cite more than one source. In total, seven colleagues would have used an influential textbook—I will not report titles and authors as this is a review of the Dictionary-; five would have used one of the previous versions of the Dictionary; and two would have used Wikipedia. This experiment would not be mentioned in the Dictionary as an example of good epidemiological practice but does suggest that the internet alone is not enough. A dictionary of epidemiology directly sponsored by the International Epidemiological Association is also a sort of building block of our discipline. Rather than drawing a line between epidemiology and other disciplines, it implies a shared language and core of concepts and knowledge. It does not mean that its readership is limited to epidemiologists; on the contrary it is a way to communicate with other disciplines and especially with the multitude of disciplines that typically interact with epidemiology in one way or another. Communication should always be two-way and the Dictionary, consistently, includes terms that do not really belong to epidemiology (see for example: epigenome or retrovirus, both mentioned in previous editions too). A multidisciplinary approach is definitely one of the hallmarks of epidemiology, as well demonstrated by the first sentence in the definition of the newly introduced term global health: ‘The international, transdisciplinary, and intersectoral research, knowledge, and policies for improving population health and health determinants on a planetary scale’. I am not the only one to believe that one of the main recent changes in our field is a formal methodological approach to causal inference. Epidemiologists have been discussing causal inference since the earliest days of the discipline, but we have now new tools and methods and we use a more formal approach to distinguish between association and causation. We are witnessing the flourishing of courses in causal inference at all levels—but not yet in medical schools, at least not in Italy—and thematic scientific journals. This big, or small, revolution was entered the fifth version of the Dictionary and has been further developed in this sixth edition. Just looking at the letter C as in cause, terms like confounding, confounder, collider, causal inference and counterfactual outcome have been profoundly revised or newly introduced. Given it is a dictionary, its appreciation is based on the reader’s scientific interests and field of work. I felt it natural to start from the definition of epidemiology, which has been slightly changed, mainly broadening its