dapto wrote: ↑Wed Aug 05, 2020 7:05 am
Without knowing why something is unhealthy, you are describing correlation and not causation.
This is a misconception, that's not at all how science works. You can determine that there exists a causal relationship without knowing its mechanisms.
E.g. even without any knowledge of gravity we can determine that lifting and then dropping an apple causes it to fall.
Having the mechanism of action is even better including things like the equations that describe gravity and whether there's an exchange particle (the Higgs boson) at work there, but causation can easily be established experimentally beyond any reasonable doubt without knowing the mechanism -- indeed, in the case of gravity it still remains poorly understood (arguably more poorly understood than the effects of meat on human health).
Would you assert that because we have not found the Higgs boson that we can only claim correlation between the act of dropping an apple and that apple beginning to fall? If not, then you need to rethink your understanding of correlation vs. causation.
dapto wrote: ↑Wed Aug 05, 2020 7:05 amA good example of how this can lead to flawed thinking is the debate about a carbohydrate diet verses a diet high in lipids or fats or a diet high in both which is what many from the US eat. Does saturated fat cause heart disease or is it correlated with heart disease because people who eat a diet high in saturated fat often consume something else along with it?
You're confusing cofactors with a nullification of causation.
In general, saturated fat consumption raises LDL cholesterol (it depends on the type of saturated fat too, e.g. lauric acid from coconut behaves differently from most animal fats but that's unnecessary nuance here if we're just talking about meat), high LDL causally increases risk of arterial plaque formation -- plaques which are literally made of cholesterol.
We know this is causal for many reasons, but some of the best evidence comes from statins which act to lower blood cholesterol and treat the underlying cause -- with extraordinary evidence of efficacy.
There are things that may make the plaques worse or help to clear them, but you can't form cholesterol plaques without blood cholesterol.
A good example of a correlation that muddies the waters is dietary cholesterol, because dietary cholesterol correlates with saturated fat, it's harder to determine if it's the endogenous production of cholesterol or the dietary cholesterol that's the principal cause of the plaques. Both play a role, but right now it seems to be that endogenous production is much higher than consumption except for in hyper-responders -- that means if the two sources of cholesterol are equally harmful, then it's the saturated fat that should be the main focus of public health messaging. Food like shrimp that are high in cholesterol bot low in saturated fat should not be a big issue for most people.
There is another theory that dietary cholesterol might be absorbed much less, but that it may be oxidized and damaged in a way that endogenous cholesterol is not, thus making it more prone to forming plaques (which are composed largely of oxidized cholesterol). Endogenous cholesterol can be oxidized in the body, but that saved step might make dietary cholesterol worse than it appears based on the numbers alone.
And yeah, it's well known that eating a lot of vegetables (antioxidants) can help prevent plaque formation, yet another variable that needs to be considered.
None of that negates the well established causal link between dietary saturated fat and heart disease. There being other variables at play doesn't cancel out causation.
To extend the cyanide analogy, would you say large doses of cyanide just aren't poisonous and aren't causally associated with death because there's another variable (B-12) that can blunt its effect?
Anything that can kill you, from saturated fat (blunted by vegetables) to a car crash (blunted by a seat-belt) has mitigating factors that can be protective. It doesn't mean that thing doesn't cause death and that it's only a "correlation".