Once upon a time, the vaguely suspect legal group Center for Science in the Public Interest set their sights on alcohol, claiming it caused breast cancer.

Media was a smaller community then - three television networks - and for national stories local reporters often rewrote wire pieces from a centralized pool. If you groomed a few journalists as allies you were able to place a large chunk of public thinking in your grasp, and breast cancer had a lot more sympathy than something like cirrhosis.

For cirrhosis, the public blame is on too much drinking, but on breast cancer, which likely affects almost every American at some point, it was easy to blame the chemical. And that is what trial lawyers like CSPI want to do.

CSPI did what activists do, they promoted a study that used a few hundred raging binge drinkers in the 1970s and extrapolated that out to everyone, and then boosted their calls for lawsuit action by citing every pooled analysis of cohort studies (.e.g. JAMA. 1998;279:535-540) - the kind of epidemiological data dredging Harvard School of Public Health publishes every month - they could find.

They have long traded in an ''attitude of crisis'' mentality, and their tools are 'health sensationalism' and half-truths.  The founder, Michael Jacobson, was a thug when it came to scientists who opposed his agenda, and an overt sexist when it came to women in science on top of it. His goal was to find companies to sue, that provided the bulk of their revenue, and then they could claim victories in meaningless gestures like getting a pointless sulfite warning on wine. 

There is no known maximum safe level of alcohol, human biology is too variable to create one, but groups like CSPI weaponized the language of science against it and spun "no safe level" to mean no level of alcohol is safe - as in "there is no safe dose of alcohol." Which the science evidence shows is false. Even in pregnant women.  America is pretty Puritanical by heritage so saying pregnant women can have a glass of wine leads to indignant outrage, but Europe does not have American Puritanism about a glass of wine during pregnancy and they don't have any more birth defects than America has. Yet women are shamed and told they should be abstinence only because that pilot study of binge-drinking alcoholics in the 1970s became famous for its analysis showing alcoholics did give birth to more babies with birth defects.

Alcohol is a class 1 carcinogen, the International Agency for Research on Cancer made that ruling long before they were hijacked by environmental consultants such as Chris Portier, Martyn Smith, or anyone Linda Birnbaum at National Institute of Environmental Health Sciences (NIEHS) is willing to hire. But that is at high doses. Plenty of people who got throat cancer also drank alcohol. A few even smoked cigars. Those can be statistically associated as long as you make sure that dose does not matter and you use 5 orders of magnitude; one dose of throat and mouth irritants is the same as 10,000 doses. Which is what IARC does.

Their claims about breast cancer and alcohol fizzled out after a while. Their desired outcome, a class action lawsuit against Big Alcohol where the cy pres award just happens to go to CSPI, never happened. But prior to that plenty of epidemiologists saw media buzz about it and armed with mouse studies and statistical correlation started producing their own papers to jump on the bandwagon.  Modern lawyers who have convinced juries that baby powder or a weedkiller might be able to cause human cancer, can thank CSPI efforts to link everything in use to a disease. 

Still, they failed on alcohol. Instead, statistics and mouse studies could even show alcohol was beneficial, and a whole field of woo grew up around that.

Like a low-fat diet still being the rage among Baby Boomers, there are a number of people who still believe alcohol "causes" breast cancer - including the U.K. National Health System. But they are using a proxy, saying that alcohol can increase the level of estrogen and then that higher estrogen over the long term can increase risk of breast cancer. Nothing in that is actually a scientific claim that alcohol causes 1 in 13 cases of breast cancer, but they claim it plainly anyway.

Statistics are still not biology, any more than theoretical physicists can create time travel. Yet epidemiologists are the life sciences equivalent of kooky theoretical physicists who think if they can do it using numbers, it is real - and other scientists just have to prove it for them.

If that is science, then I am Doctor Who.