Hacker News Reader

Comment by grandalf | Hacker News Reader

The evidence is not just from the epidemiological study, there have been lots of follow-up studies to examine the processes in detail.

I would argue that a broad, epidemiological study is the most persuasive sort of evidence to use when considering whether to make drastic changes. Otherwise you're dealing with information like "omega 3s prevent x, y, and z" so you end up taking an Omega 3 supplement along with the dozen other supplements while eating Wendy's for lunch every day.

Thinking that health can be understood on the basis of "taking supplement x is good" is the root of the current supplement gold rush (people sell tons of snake oil substances that contain "omega 3's" etc.) and a highly irrational way to behave.

View on HN

>follow-up studies to examine the processes in detail.

Examining the processes isn't sufficient to provide proof unless it's a completely thorough examination. Biological science isn't at that point yet, and so what this usually means is that group A performs an observational study, group B finds a process that would explain group A's results, assuming group A's results are true. That's different from showing that group A's results were true.

Human, and animal, physiology is very complicated, and it's just plain wrong to assert that we know that eating meat decreases longevity. We don't. There is some evidence to believe it, and some evidence to not believe it at this point (both epidemiological studies and at the molecular understanding of the processes involved).

Epidemiological may be persuasive, but it's misleading because you take a single variable out of hundreds that may not even be known. So you may get Texans who become vegetarians based on longevity studies in some region of China, but then live an even shorter life because their lifestyles and the particular vegetarian foods that they eat aren't as conducive to longevity. This history of epidemiological studies are full of cases like this.

I don't dispute your logical points, but I think there is still a rather strong case for eating a plant-based diet.

Clearly, humans can eat meat (during most of human evolution calories were scarce and eating meat conferred a selection advantage)...

But why would you reason that all foods are equally beneficial and harmful? To me, broad epidemiological studies can offer clues on classes of foods and some of the costs/benefits that they confer.

I'm not reasoning that all foods are equally beneficial and harmful, and I accept that it may very well be that a plant-based diet is superior to a meat-based diet. I reject the idea, however, that we know enough to say that this is true, and I absolutely reject the idea of significantly changing behavior based solely on epidemiological studies -- the epidemiological studies should be the starting point for much further study. To date, most large-scale controlled studies on diet have unfortunately been shelved. It should entirely be possible to separate multiple groups on multiple different diets and see what happens over a period of, say, ten years -- one diet as vegetarian, one diet as meat/fruit/nut but no grains, a fish+vegetarian diet, while holding other variables constant (such as calories). This is what we need to do to know, and this is what has consistently been proposed, but not done.

You are correct.

However the study that you call for has been done in laboratory animals -- TCC found that (controlling for calories, etc.) the animal proteins he studied led to the growth of cancer cells, while plant based foods did not cause cancer to develop, even in an environment of radiation, etc., that would typically be thought to be the cause of the cancer.

I fully agree that the study you mention would be hugely beneficial, but by your logic one ought to smoke cigarettes if one chooses, or at least ought to have done so in the 1980s before more conclusive evidence began to emerge... even though most doctors/scientists had held a strong belief that smoking was a bad idea for decades.

If one were to base all of his health decisions on studies that were conclusively done in humans (without regard for any animal studies, etc.) one would be limited to a 1950s understanding. Humans are not mice, and the studies don't all correlate perfectly, but many do. And they contain valuable (though not necessarily conclusive) information.

Reply on news.ycombinator.com