Half a decade ago, France's International Agency for Research on Cancer (IARC) tried to fight for its credibility in the face of a scientific onslaught against their latest epidemiology findings by actually lowering the "risk" of something.

Like everyone else, when it was announced they were 'studying' it - in IARC, that only means mouse models that support claims of cancer and surveys that can be linked to cancer - I assumed they would finally do what they had wanted to do since the early 2000s; declare coffee a carcinogen.

And get $15,000 an hour expert witness contracts from lawyers who could then sue, claiming someone who cut the lawn and drank a cup of coffee got cancer due to the coffee. 

Do you think this will give you cancer? Then you can be an IARC epidemiologist! Unfortunately, you are dumber than every illiterate 16th century peasant, who instead knew 'the dose makes the poison'.

IARC responded to the jeering they got from scientists over suspect claims about bacon, virtual diesel emissions and flame retardants by actually lowering the classification on coffee. They probably hoped that it would be a 'see, we don't just pick things members of our panels want banned' veneer but it actually showed how capricious and arbitrary their model is. Coffee is just as harmful as fire protection spray on grandma's old furniture. It is also just as harmless.

That is the problem with IARC epidemiology. It ignores dose. One shot of Scotch is considered the same as 10,000 to epidemiologists, despite everyone except epidemiologists and New York Times journalists knowing that makes no sense.

The US FDA pulled no punches in noting how wrong the methodology was on its recent aspartame claims; and that IARC used suspect studies and methodology over the objections of every scientific body, from FDA to the European Food Safety Authority. It is safe. It does not cause cancer. 

If IARC chooses a result and then picks mouse results to match it over the objections of chemists, biologists, and toxicologists, that does not mean anything is wrong with aspartame, it means something is wrong with IARC.