It's a really interesting subject, and a subject where marketing and politics and even religion tend to obscure the actual data. I'm a data guy myself, so when the actual data is available, I trust the data over the papers or books.

A huge problem with nutritional science and the links it tend to make with disease is that RCTs tend to be too small and too riddled with regression to the mean effects to be of much value, while practicality forces these studies to use questionable surogate endpoints, and epidemiological data tends to have associations so small and so clustered and confounded to make things hardly any better. This appart from the facts that questionaire based epidemiology yields highly inacurate data, and that multi variate analysis often wildly adjusts the data without a solid causal model, or at least without a causal diagram published in the paper that could be independently validated or critisized.

Next to being a data guy, I write fiction, and reading books and even papers on nutrition, Ive been seeing a trend where the books and even papers are really strong on narative but really poor on the actual math. I often joke that nutritional science papers and books are my favourite subgenre of speculative fiction.

Let me give a huge example. The book The China Study is a book that is based on a large and very broud questionair based epidemiological study. The book makes claims about animal protein and animal foods in general being really bad for human health.

The data used to be publicly available, but was deleted, however the data is still available in internet archives, so a few years ago I looked at the actual data.

While the data was of pretty poor quality, there were many usefull fields en enough to do a simple cluster analysis of food stuff versus longevity, and so I did just that. And guess what? The foods most positively correlated with longevity were mostly animal sourced foods, with saturated fat being at the absolute top of the list and animal protein not being much below.

And realize the book was written by an actual scientist with decades of work in the field. A scientist writing a book that claims the absolute oposite of what tha actual data suggests, and I'm saying sugest, because as a data guy I can confidently say that the actual data is nowhere near good enough to make any actual causal claims.

And it doesn't end with individuals. Things like this are often orginized by doctors and scientists with strong ties to either extreme animal rights groups (like the PCRM that is basicly a branch of PETA), have been exposed as actualy countering real science from being published (the True Helath Initiative), or are linked to (people with known financial ties with) the sugar and/or plant based oil industry like the Eat Lancet mess. The more you look into the science, data, and the people involved in the studies and publications implicating animal sourced foods, the same mostly American faces pop up.

If however instead of looking at RCTs and epidemiological studies, you look at chemistry and bio chemistry, then you get a glimpse of a tiny bit of real science hidden under the narative. While unheated fats might not actually ever be bad for you, the same can't be said for heated cooking oils. Want to worry about fat consumption, don't worry about saturated fat, worry about aldehydes formed when deep frying in unsaturated fats. Want to worry about protein, don't worry about animal vs plant protein, worry about HCAs and AGEs formed in the browning process of your meat. And, yes, worry about nutrient deficiencies.

In short, the data does NOT justify any causal claims of meat, saturated fat or animal protein perse causing cancer or heart disease. The bio-chemistry and chemistry DO justify being concerned about curring and preparation of meat (and other foods). So it's not fish over chicken over red meat, its soup over grilled over deep fried, actually.