Modern Epidemiology, 4th edition. TL Lash, TJ VanderWeele, S Haneuse, KJ Rothman. Wolters Kluwer, 2021.
Modern Epidemiology, 4th edition. TL Lash, TJ VanderWeele, S Haneuse, KJ Rothman. Wolters Kluwer, 2021.
复制标题
DOI：
10.1007/s10654-021-00778-w
复制
发表时间：
2021-08
影响因子：
13.6
通讯作者：
Ahlbom A
中科院分区：
文献类型：
--
作者：
Ahlbom A
流行病学是由它的主题而不是它的方法来定义的，它在书中说，就像其他科学分支一样。然而，流行病学的教科书通常是关于可能引起混乱的方法和理论。这本书的第一句话指出，流行病学是研究人类疾病发生和健康状况的科学。描述性流行病学测量疾病频率和其他健康指标如何随时间、地点和人而变化;麦克马洪在他的开创性教科书中使用了类似的流行病学定义，并根据这些参数组织了部分内容。病因学或分析性流行病学评估暴露（包括原因）对疾病频率的影响。虽然绝大多数学术流行病学似乎属于第二类，但现代流行病学强调描述性研究的必要性。但它也指出，描述性和分析性流行病学之间往往没有明确的界限，它们在书中没有分开对待，只是有一章是关于监测的。书中说，现代流行病学直到世纪中期才开始发展，尽管在此之前已经出现了一些复杂的流行病学例子。现代流行病学的发展是因为新的研究问题需要新的工具而开始的，这是一个合理的猜测。这是流行病学开始调查在开始接触后很长时间才出现的疾病的时候，结果是罕见的，并且有多种原因。流行病学逐渐成为一门独立的科学学科。需要对基本概念进行定义和命名，以便在报告和讨论中使用，这是将流行病学转变为一门独特学科的必要条件。鉴于流行病学是如何定义的，最基本的概念是那些用来衡量疾病频率的概念。令人有点失望的是，在这方面仍然没有一个统一的术语，作者们感到有必要为核心概念提供替代术语。令人欣慰的是，尴尬的累积发病率已经消失，取而代之的是合理的发病率比例，但患病率仍然是患病率，而不是类似的患病率比例。值得注意的是，这本书花了25页的措施的发病率和患病率，一个主题，似乎相当微不足道，也许已经如此，如果措施只发生在生命表一样的模型。但现实使事情变得复杂，因为实际计算这些措施的人群和情况差异很大，通常对流行病学不利。为了准确地计算出这些数字，我们必须认识到其中的复杂性，但在书本和教学中，这些复杂性往往被淡化了，在介绍流行病学定义的章节之后，在其他章节之前，有两章是关于因果推理的，一章是关于在规范框架中体现的推理，正如作者所说的那样，另一章是关于因果关系的数学模型，包括有向图和经典的饼图模型。这些主题的突出位置是合乎逻辑的，但也证明了因果关系在流行病学中的核心和明确作用。这些章节都很简洁，但包含了大量关于因果关系及其评估的想法。然而，有趣的是，作者们没有定义因果关系或因果变量。事实上，这本书说，什么构成一个原因仍然是一个争论和讨论的问题。其中一个主题是一个变量是否必须是可操作的，以...
Epidemiology is defined by its subject matter rather than by its methods, it says in the book, just like other branches of science. Yet, a textbook in epidemiology is typically about methods and theory which may have spurred confusion. The first sentence in the book states that epidemiology is the science that studies disease occurrence and health states in human populations. Descriptive epidemiology measures how disease frequency and other health indicators vary with time, place, and person; MacMahon, who used a similar definition of epidemiology in his groundbreaking textbook, organized parts of it according to these parameters. Etiologic, or analytic, epidemiology assesses the effect of exposures, including causes, on disease frequency. Although the vast majority of academic epidemiology seems to belong to the second category, Modern Epidemiology emphasizes the need for descriptive studies. However, it also states that there is often no clear border between descriptive and analytic epidemiology, and they are not treated separately in the book, except that there is a chapter about surveillance. Modern epidemiology started to develop only in the middle of the twentieth century, the book says, although some examples of sophisticated epidemiology had appeared earlier. It is a reasonable guess that the development of modern epidemiology started because new research questions required new tools. This was when epidemiology started to investigate diseases that manifested long time after start of exposure, where the outcome was rare, and multiple causes were abundant. Epidemiology gradually turned into a scientific discipline in its own rights. Fundamental concepts needed to be defined and also named such that they could be used in reports and discussions, and it was a sine qua non for the turn of epidemiology into a distinctive discipline. Given how epidemiology is defined the most fundamental concepts are those that are used to measure disease frequency. It is a little disappointing to note that there still is not a unified terminology in this respect and that the authors sense a need to offer alternative terms for core concepts. It is a relief to see that the awkward cumulative incidence is gone and replaced with the logical incidence proportion, but prevalence is still prevalence and not, the analog, prevalence proportion. It is noteworthy that the book spends 25 pages on measures of incidence and prevalence, a topic that may seem rather trivial, and perhaps had been so if the measures had only occurred in life table like models. But reality makes things complicated because the populations and situations where these measures will actually be calculated vary widely and are often unfriendly to epidemiology. The involved complexities must be appreciated to get the numbers right but are often downplayed in books and teaching.Right after the introductory chapter that defines epidemiology and before everything else are two chapters about causal inference, one about reasoning embodied in canonical frameworks, as the authors phrase it, and one about mathematical models of causation, including directed graphs and also the classical pie models. The prominent placement of these topics is logical but also a testament to the central and explicit role that causality has been given in epidemiology. These chapters are succinct but contain a wealth of ideas about causation and its assessment. Interestingly enough, though, the authors hold back from defining causation or causal variable. Indeed, the book says that what constitutes a cause remains an issue of debate and discussion. One of the topics is whether a variable must be manipulable to be …