In fact, this excellent essay doesn t even mention bathmats but it may have to, when the next version comes out. We tried not to mention the entire headline really, we did but in the end, we caved, so here it is:
Why You Shouldn't Care That Yogurt, Mouthwash, Red Meat, Burnt Toast, and Bras Have Been Linked to Cancer.
This exposé by (Steven) Ross Pomeroy appeared in the Newton Blog of RealClearScience, and should be read by anyone who s ever had their wits scared out of them by some TV or news media headline asserting that one of their favorite (foods, activities, habits) increased their risk of getting cancer.
(The author is the assistant editor of RealClearScience, and a zoologist and conservation biologist by training).
His main targets are the researchers in epidemiology whose main goal is to accumulate vast quantities of data from people s recollections of their behaviors or diets over the past years and correlate that information with a variety of health outcomes. Using the common standard of statistical significance, i.e. that the association has at least a 95 percent chance of being real, they can torture and dredge the data by correlating hundreds of exposures with hundreds of outcomes, knowing full well that by statistical associations (chance), some statistically significant link will be detected. Does the ostensible cause-and-effect make any sense, or have any biological plausibility? Never mind, as long as the authors can get a paper published somewhere and, more importantly, make a news splash about their sensational discovery, that s all that matters.
Here s what Mr. Pomeroy says: I hate to break it to you, but almost everything in daily life has been linked to cancer: burnt toast, hot dogs, poor tooth brushing, you name it!
You now have two choices: panic or continue on with your day.
I recommend the latter.
Much of the health information you read online or hear on the morning news comes from observational studies -- scientists look at people who eat certain foods, or take certain drugs, or live certain lifestyles and see how their health compares with the health of people who don't do those things.
Those types of studies, observational, have indeed led to some major breakthroughs in science and medicine: e.g., the link between smoking and lung cancer was detected on a retrospective analysis, as were most of the data on the efficacy of various vaccines. Why? Two reasons: 1-no scientist would dare to consider doing a randomized, prospective study comparing a group of volunteers instructed to smoke vs. a similar (control) group mandated not to smoke. Never happened, never will. 2-when the results of an observational study are overwhelming (e.g., data analysis revealed that lung cancer patients were 30-fold more likely to have been smokers than not, retrospectively, sure, but still ¦), the results can be valid.
However, most such studies, especially the ones with blaring headlines, are more like women who used baby shampoo while pregnant had a 15 percent higher risk of their children becoming obese adolescents, or some such. Reporters (and medical journals!) love exaggerating such results, but to be considered even possibly a real signal of a cause-and-effect, an observational study needs to have a more than double variation between the study and control groups. (See the Bradford-Hill criteria for cause-and-effect for a sound background in this area).
ACSH s Dr. Gil Ross had this comment: Mr. Pomeroy makes this key point: many, if not most, of the epidemiological literature is tainted with this sort of data dredging. He cites an ACSH advisor s study (Stan Young along with co-author Alan Karr) from 2011 which showed that none of a group of such studies were able to be confirmed on attempted replication or statistical re-analysis. From that article s abstract: Any claim coming from an observational study is most likely to be wrong. (A similar theme can be read in today s NYTimes Science section, noting how often peer-reviewed publications in scientific journals are retracted).
So next time you see an article warning you about a new cancer-causing exposure, especially if it s about something you really, really like, pay it no mind. Chances are high that it s merely an attempt to scare you and get attention for the authors: so simply remain calm.
