Ingredients

The Bombbor publishing house represents the book of a chemist and popularizer of science George Zeidan “Ingredients. The strange chemical properties of what we eat, drink and apply to the skin ”(translation by O. A. Lyashenko).

When a hand reaches for a pack of chips, you should refrain and refuse yourself pleasure or can you eat a couple more pieces? Gathering for the beach, how many sunscreen should be applied to the skin, and how does SPF work? Is it still useful or harmful? The book “Ingredients” discusses the whole variety of chemicals that are in different ways in the human body. Is it true that processed foods are a real poison and how do they provoke obesity and other diseases? Why are we burning in the sun and what are the consequences of abuse of solar baths? What happens when chemicals from products and the environment come into contact with chemicals from our body? You will find out why there are generally processed foods, and then, together with the author, study the chemicals, the effects of which are exposed daily.

We offer to read a fragment of one of the sections of the book.

In 2005, the Greek epidemiologist John Ioannidis wrote an article with a completely harmless and unopulsing name “Why are the majority of published research results.” If at that time you followed the news of science, then you read it. Regardless of whether you agree with its name or not, it caused a resonance in the scientific community. The article launched several projects for reproducibility of results in psychology and fundamental studies of cancer (in them scientists repeated the experiments described in the articles to check if they will come to the same results). Joannidis moved to Stanford and switched his attention to the epidemiology of nutrition. He told the Canadian reporter that this science “should go to the garbage tank”, and Vox journalist, that “this is an area that has aged and died.” “At some point, we will have to bury the body and move on,” he added.

This is a statement!

Epidemiologists of food spoke in response, but they did not have enough drama of Johndis. Walter Willett, an epidemiologist of food from Harvard, replied: "You seriously distort what the epidemiology of food does." This is about the same thing as to answer: “We just have different opinions on this issue” - to the statement “Everything you are doing is a bunch of rotting shit.” Disagreements between Joannidis and Willett, which I will now call a war, erupted due to epidemiology of nutrition, are the largest and most deep rabbit hole of all that I read about, collecting material for this book. If you are funny terribly confused statistics, the struggle between Harvard and Stanford [1] , as well as doubts about the clarity of reason, then you will like this rabbit hole. Other things seem funny to me, therefore, instead of falling deep into you, I will vote you along its periphery. We will bend, look into the darkness and ask questions to the inhabitants of Nora, but will not go beyond the horizon of events.

One of the arguments of Joannidis against the epidemiology of nutrition was that we already discussed in the seventh chapter: the more hypotheses you check, the higher the likelihood that at least one of them will accidentally turn out to be statistically significant. However, he claims that this is especially true for food and diseases, because the number of hypotheses associated with them is endless.

Each of these lines is a potential experiment. If you took 300 products and 800 diseases, then you would receive 240 thousand potential tests. Even if only 5 % of them randomly revealed statistically significant connections, then you would still receive 12 thousand experiments that demonstrate correlations between, for example, kuumvats and anal abscesses [2] .

Joannidis also says that the studies that showed the connection between the food product and the disease are easy to publish (I agree with it in this). Such experiments more often attract the attention of the media, so you will more likely see the heading “Consumption of the European Salt [3] with an increase in the risk of egg cancer by 23 %,” and not the heading “bay leaf was not connected with anything concrete” [4] .

The counterargument of Willett and the company sounds like this: we do not just blindly check every possible hypothesis as robots, but use our knowledge of chemistry, animal experiments and metabolic experiments to narrow the list of hypotheses to the most reliable. In addition, we recently switched from testing individual products to test diets (for example, the Mediterranean diet), which reduces the total number of hypotheses and more accurately reflects how people eat in real life.

Joannidis gives another argument: the epidemiology of nutrition is based on observational, not randomized controlled studies. Let me remind you that during observation research you are not trying to change the behavior of people. In the best of them, scientists are gaining a group of volunteers, watching them for several years until any of the participants will develop cancer, cardiovascular diseases or any other diseases that are interesting to them, and then compare patients of the study with healthy ones. Did they smoke more? Did you play sports less? Have you eaten less mole?

We will talk about Joannidis's claims to observation research, but let's give them their due first: Willett claims (and I agree with him) that such experiments are successful. The best and most famous example is smoking. Early evidence of the harm of this habit was obtained mainly due to observation. Let me remind you that the number of randomized controlled studies mentioned in the report of officials of the public health system from 1964 was ... zero. Of course, this is not quite an epidemiology of nutrition, because we do not eat cigarettes, but this science also conducted successful observatory research.

The Willett team gave several examples, the most fresh of which was a observational study of trans fats. Scientists also noted that they rely not only on them, but also on randomized controlled studies, however, as they rightly noticed, the former are cheaper, and the latter are sometimes unethical or impractical.

So, what does Johannidis do not like in observational research? Firstly, the fact that they are observational and they can be compared more with the nuisan of public pools than with the addition of urine to Menzurka. Even if they could identify genuine connections between specific food and the disease, they would not show whether the product causes a disease. The second problem is memory. The conversation about this will be long, so get ready.

[1] In 2010, Harvard made Johndis Adjunct Professor of Epidemiology, thereby taking a clear position ... on both sides of the barricades. Typical Harvard. - approx. aut.

[2] Everything is a little more complicated. Some variables are not independent, so the real numbers are lower. Nevertheless, the idea remains the same. - approx. aut.

[3] European salt, or European sea, is a type of fish from the Cambal -shaped detachment.

[4] It should be noted that I invented both headings. - approx. aut.