Ingredients: the Strange Chemistry of What We Put in Us

author: George Zaidan

related books: Statistical Rethinking, Observation & Experiment

Main take-away

  1. Foods contain a lot of different chemicals. The reason is because (i) the plants that compose our food produce metabolites to protect them from pests and herbivores, and (ii) processing food (e.g., to preserve it longer) introduces lots of other compounds, either by adding it or from reactions of the process.

  2. To understand how a substance can cause harm, it is crucial to unravel the mechanism. For example, UV and cigarettes mess with your DNA, causing mutations.

  3. In many cases, the effect of health benefit or detriment is relatively small with a (partially) unclear mechanism. The gold standard to measure this effect is using a randomized controlled trial. These are often unfeasible, leaving observational studies as the only option. Unfortunately, such studies are complicated because of confounding factors, p-value hacking etc. It is hard to decisively establish that some food is good or bad for you.

Are Cheetos bad for your health, and how do you know? Zaidan's quest to answer this question starts with chemistry and ends with statistics. Chemistry, as generally accepted by non-chemists scientists, is relatively straightforward. You have a bunch of reagents, there react and form one or more products. Food science is just applied chemistry, so it should be easy, right? One of the messages of this book is that chemistry outside of the beaker is messy. Our food is made from plants and animals, who make thousands of compounds for various reasons. Some interactions between our health and our food intake are clear-cut. For example, the deficiency of a particular vitamin, say vitamin C, can hamper several vital processes such as good collagen production, leading to nasty diseases, such as scurvy. Supplementing your diet with the missing vitamin usually leads to a miraculous recovery. Luckily in many countries, nutritional deficiencies are becoming rare. Compare with the health benefit of occasionally eating an avocado. This fruit likely has several compounds that might have some beneficial effect on several processes. Chemistry is quite helpless to disentangle these. So, if uncovering the mechanism between food and health is nearly impossible, one has to resort to science. Zaidan explains why randomized trials, while theoretically the best way to answer such questions, are unfeasible to perform in practice. So we are left with observation studies, susceptible to random effects, confounders, p-hacking and generally suffer from all causes and effects being entangled. The bottom line is that it is tough to what food does to our health. I enjoyed reading this book! It is an excellent introduction to food science and observational studies alike.

Who is this for?

The book explains the very basics of food chemistry. Its main topic, however, is epidemiology studies and why they are so hard. I found this a more accessible and gentler introduction to randomized trials and observational studies than many layman books specifically about statistics. The book is also quite funny.