Epidemiology for the Uninitiated

How do you plan and conduct an epidemiological survey? What is the difference between relative risk and odds ratio? How do you adjust for confounders by stratification or standardization. What are the limitations for ecological studies that compare exposures and diseases in different geographical locations? What is the difference between a cross-sectional study and a longitudinal study?

What does epidemiological research say about the potential benefit of earlier diagnosis? How do you define a disease case? How does incidence differ from prevalence and how do you calculate these two metrics? How do you analyze validity of epidemiological surveys? When should you aim for specificity and when should you aim for sensitivity? How come measurements of the same subject can differ? How can you figure out what conclusions can be draw from an epidemiology study, and which conclusions are not justified by the data?

With the current massive surge of disinformation about the coronavirus, it is now more important than ever to defend yourself against the onslaught of pseudoscientific quackery. Some people are pushing ineffective and harmful treatments. Other people want to pretend like the new coronavirus is somehow nothing to worry about even though there are millions of infected people and hundreds of thousands of dead. How can you as an individual to protect yourself and loved ones from such misinformation? One way is to grow your knowledge about scientific fields related to infectious diseases, medicine and epidemiology.

Luckily, there are a number of free epidemiology textbooks freely available online, such as Basic Epidemiology or Epidemiology and Prevention of Vaccine-Preventable Diseases. But some of those textbooks are very long and span hundred of pages. Is there some other epidemiology resource that is both brief and of high scientific quality? One that do not require that much background knowledge?

Epidemiology for the Uninitiated is the fourth edition of an introductory textbook in epidemiology written by Coggon, Rose and Barker published on the website for the British Medical Journal. It does not require any background in medicine, mathematics or science and its 13 chapters surveys the basic foundations of the area. In particular, it focuses on how to calculate and compare disease rates, how to estimate error and bias as well as the basic research designs that epidemiologists typically use. It also looks at screening, disease outbreaks and offers a primer on how to read epidemiological literature. If you are only doing to read one epidemiology textbook during this new coronavirus pandemic, Epidemiology for the Uninitiated is the one to read.

Follow Debunking Denialism on Facebook or Twitter for new updates.

The first chapter goes into detail about the definition and background of epidemiology. The second and third chapter discusses how to measure diseases in a population and comparing diseases rates between populations. The fourth chapter talks about error and bias, while the 5 chapter describes how to plan and carry out a study. The chapters six through ten go over different study designs, while the eleventh chapter explores disease outbreaks. The twelve chapter provides crucial insight into how to critically evaluate published epidemiological studies. The last chapter offers a few suggestions on what to read next.

What makes this textbook stand out? The primary reason is that it delivers the core mathematical tools for epidemiology without the bloat, offers detailed discussions of important epidemiological research designs (such as ecological, longitudinal, cross-sectional, and experimental) which includes strength and weaknesses of each approach. The entire textbook can be read within a day, as it is equivalent to only approximately 60 or so pages of a conventional book.

Perhaps the best section is the final chapter that offers tips and tricks for digesting epidemiological literature, highlighting the importance of identifying potential sources of bias, why it is strongly discouraged to put a too strong focus on statistical significance in epidemiology, why magnitude and confidence intervals do better and the crucial difference between causality and confounding.

A lot of scientific skeptics are surely familiar with discussing a certain scientific or medical topic on social media, only to be hit with a torrent of links to scientific papers in the form of PubMed abstracts that the commentator claims support their case. Learning more about how to critically read and evaluate published studies is a key skill that all scientific skeptics should strive to master. Anyone can link to studies that they think support their position, but it takes critical thinking and scientific skill to be able to read and evaluate papers to find out what conclusions can be drawn and not drawn from the results.

There is no PDF version of Epidemiology for the Uninitiated, so it can only be read online. A cache of the online version can be found here.