I recently took my boys for their annual checkups and each time the pediatrician asked the same question: “How much milk do they drink?”
“Three times a day,” I answered. She smiled and said “good.”
But according to this new study, my doctor is giving me bad medical advice.
A good look at the study reveals it is yet another case of correlation, not causation, that’s feeding these inflated and hysterical headlines.
First, the correlation. Student researchers looking to get their name out there love using this trick and they’ve done just that with this latest example of junk science. This is how it’s done: The researchers collect data on how people are eating (in this case, they took data from the National Health and Nutrition Examination Survey, called NHANES) and find a habit of the American public (in this case, drinking milk and eating meat) and then they line that data up against the number of people who are getting cancer (in this case by looking at data from the National Death Index).
And what did they find when comparing these two data sets? That the people who consumed animal protein died of cancer more often.
But it’s important to realize a few things about the NHANES data. Now, I don’t want to totally dismiss it because it is a valuable source of information but it does have limits. First, NHANES relies on face-to-face interviews with people where they are asked to recall one day of eating. Well, guess what happens when you are asked very personal questions, in person. You lie! That’s right. You lie? Why? Because it can sometimes be embarrassing to admit your weaknesses. Secondly, people have bad memories, which only adds to the uncertainty of the data collected.
Also, NHANES data is only a tiny piece of a very large pie. NHANES doesn’t have complete information on all Americans. Rather, the survey examines a nationally representative sample of about 5,000 persons each year. Do those 5,000 people accurately represent the entire U.S. population? Maybe. Maybe not. Does that mean we chuck all NHANES information? No. But perhaps we shouldn’t’ use it to draw conclusions like “milk as bad as smoking.”
Regarding those inflated and hysterical headlines: Well, no surprise there, right? Headline writers are good at getting clicks and they have to in today’s competitive media market. But, people should remember that this latest “milk kills” study is only one study that says there’s a possibility that you’ll increase your chances (got that, chances) of getting cancer or that there might be some random, vague correlation between gorging at 10 cent wing night and cancer. Yet there is a ton of rigorous scientific studies that actually show smoking causes (CAUSES) cancer. Spot the difference?
When you read most of the stories this study has generated, you get the sense that the researchers are backpedaling; fumbling around saying “ummm…we didn’t actually mean people should, you know, drop the glass of milk and ummmm….take up smoking…no, no, no, that’s not what we meant!” Okay, I’m paraphrasing a bit.
But consider this statement from the lead researcher:
The size of the effect we're finding was similar, however I think the association with smoking mortality is way more clear-cut," says Canon. "There's been a lot more research, and there's a lot less potential for confounding factors. For nutrition, it's really hard to unravel what someone's diet is and really quantify that—whereas it's quite easy to say, 'Do you smoke a pack of cigarettes a day?'"
Canon emphasizes that this definitely isn't a reason to continue (or start) smoking.
Well, at least Canon managed to get that good “Don’t smoke!” message in there. Meanwhile, plenty of Americans are going to forgo healthy, nutritious milk and protein because of a weak correlation to cancer and a study that should never have generated a headline.