Meaningless Associations

      23 Comments on Meaningless Associations

A couple of studies caught my eye recently. When I opened our local newspaper this morning, I saw this headline over a short article:

Exercise cuts womb cancer risk

Women who regularly work up a sweat exercising have a 30 percent lower risk of developing endometrial cancer, a new study says.

Researchers at the United States’ National Cancer Institute analyzed 14 previous studies and found physical activity cuts the risk of endometrial cancer by 20 to 40 percent when compared to sedentary women. The study was published online Wednesday in the British Journal of Cancer. It was paid for by the National Cancer Institute.

“We already knew that maintaining a healthy body weight is an important way to reduce the risk of womb cancer, but our study showed that physical activity has a protective effects of its own,” said study author Steven Moore of the National Cancer Institute.

I exercise regularly and believe exercise is good for your overall health, but with all due respect to Dr. Moore, his study does NOT show that physical activity protects against cancer.  All it shows is that women who exercise are also less likely to develop endometrial cancer.  That’s an important distinction, and one that far too many researchers and health reporters don’t understand.

I looked up the full text of the study and, as I suspected, the previous studies Moore’s team analyzed were all observational studies.  The trouble with observational studies is that while the correlations researchers find in them may sometimes point to cause and effect, very often they’re simply the result of comparing different kinds of people.

Perhaps you recall the estrogen fiasco.  Examining data from the Harvard Nurses Study, researchers found that post-menopausal nurses who took estrogen had a lower rate of heart disease.  Based on that finding, doctors were ready to start prescribing estrogen to every middle-aged women in the country.  Just one little problem:  when estrogen was tested in a controlled clinical study, the women who took it ended up with a higher rate of heart disease, not lower.  Same thing happened in a clinical study with men.

As it turned out, the nurses who took estrogen were simply more health-conscious than most other nurses and were taking estrogen because they believed it was good for them — not because it actually was (at least for heart health). Health-conscious people are different. They exercise more, eat less sugar and other junk food, make sure they get enough sleep, take their vitamins, are less likely to smoke or drink alcohol to excess, and more likely to see a doctor if something doesn’t seem quite right with their health.  Estrogen wasn’t making nurses healthy … healthy nurses were taking estrogen.

We could be seeing the same phenomenon here.  Exercise may, for all we know, help to prevent endometrial cancer.  Or it could simply be that health-conscious women — who engage in host of healthier behaviors that help prevent cancer — are also more likely to exercise.  The point is, this study doesn’t tell us diddly about cause and effect.

In the full text, the authors explain that they examined the previous studies to determine rates of sedentary behavior — which, in a rare case of clarity for an academic paper, they describe as “too much sitting.”  They found a positive correlation between too much sitting and developing cancer.  They also found a negative correlation between moderate-to-vigorous exercise and cancer … in other words, more exercise, less cancer.  They speculate that since obese people have a higher cancer rate, exercise may prevent cancer in part by preventing obesity.

I sincerely doubt that.  First off, as research has shown over and over, exercise has a minimal effect on controlling weight.  Secondly, I think they’ve got the equation backwards.  We don’t become fat because of “too much sitting” — we sit around too much because our bodies are accumulating fat, thus storing the fuel that would otherwise be burned, compelling us to move around more.

In a recent interview with Jimmy Moore, Dr. Robert Lustig recounted how kids given certain cancer treatments gained weight and became sedentary afterwards.  Lustig suspected the treatments had somehow screwed up their insulin levels.  When he gave them an insulin-suppressing drug, they not only lost weight, their parents reported they became much more active — without being encouraged to do so.  With their insulin levels down, their bodies were no longer storing fuel and wanted to burn it.

Sugar and fructose tell our bodies to store fat.  Going into fat-storage mode makes us feel lethargic and lazy.  Sugar and fructose are also cancer’s favorite foods.

So it doesn’t surprise me that people who engage in “too much sitting” are more likely to develop cancer.  But I doubt sitting around is the cause, or that exercise is the cure.  I think it’s more likely sugar is the cause, sitting around is a related symptom, and cutting sugar is the cure … or at least the best prevention.

This study came to me in an email:

Antagonistic people have increased cardiovascular risk

Antagonistic people, particularly those who are competitive and aggressive, could be increasing their risk of MI or stroke, new research indicates.

Studying more than 5000 people in Sardinia, Italy, US scientists found that those who scored high for antagonistic traits on a standard personality test had greater thickening of the carotid arteries on ultrasound compared with people who were more agreeable. Intima-media thickness of the carotid artery is an independent risk factor for cardiovascular events, say Dr Angelina R. Sutin and colleagues in their paper published online August 16, 2010 in Hypertension.

“We found that although men tended to have thicker arterial walls than women, antagonistic women had arterial walls similar to that of antagonistic men,” Sutin told heartwire. “So the association between antagonism and arterial thickness was much stronger for women.” And although arterial thickening is a sign of aging, young people with antagonistic traits already had such thickening, even after controlling for confounding factors such as smoking, she said.

She cautions, however, that this was a population-based sample and more research needs to be done in clinical settings.

Kudos to Dr. Sutin for exercising caution.  Associations don’t prove anything.  Their purpose is to suggest hypotheses for clinical research, not to serve as evidence for conclusions.

I have to admit, as a survivor of the Los Angeles freeway battle zones, I kind of like believing that the tail-gaiting, bird-flipping, I’m-more-important-than-you drivers I dealt with were thickening their arteries every time they made another dare-devil lane change.  (If you’ve never driven in Los Angeles, a significant portion of the drivers believe if they conduct a dozen death-defying maneuvers every minute or so, they’ll eventually discover the secret fast lane.)

However, I doubt aggressive behavior or even having an antagonistic personality causes heart disease in and of itself.  I suspect those people are just wired a little differently.  Perhaps they pump out a lot more stress hormones, which causes aggressive behavior and damages their arteries at the same time.

One of the most antagonist people I ever met was a young manager at a comedy club.  His favorite motivational technique was yelling at the top of his lungs.  As you might imagine, he wasn’t beloved by the staff.  I found out later he died of a heart attack at age 26.  I didn’t like the guy, but I was sorry to hear about his demise.  The punishment seemed to vastly outweigh the crime.

To close on a lighter note, a reader sent me a link to this picture.  Apparently the Modest Proposal I posted earlier in the week (let’s consume vegetarians for food and save the planet) is already being put into action.

I never considered storing them in the pantry … although CAN VEGETARIANS could be taken more than one way.


23 thoughts on “Meaningless Associations

  1. Nathaniel

    Hey Tom, I was studying for the law school admission test today and I came across this question that instantly made me think of nutritional science. I think you’ll see why. In this part of the law school admission test, you are presented with short arguments and then asked questions about them. (Full credit to the LSAC).

    Science writer: All scientists have beliefs and values that might slant their interpretations of the data from which they draw their conclusions. However, serious scientific papers are carefully reviewed by many other scientists before publication. These reviewers are likely to notice and object to biases that they do not share. Thus, any slanted interpretations of scientific data will generally have been removed before publication.

    Which one of the following is an assumption required by the science writer’s argument?

    I’m sure you can see the problem here so I’m not going to list all of the choices. The correct answer is,

    B) In general, biases that slant interpretations of data in serious scientific papers being reviewed for publication are not shared among all scientists.

    Some question-writer at the LSAC is a pretty shrewd commentator on the state of nutritional science, if you ask me.

    Indeed. And scientists have learned to seek out peer-review from colleagues who share their opinions.

    Best of luck on the exam.

  2. mezzo

    Stress-related sounds reasonable to me. Dr. Malcom Kendrick has similar suspicions.

    I think he’s onto something. Stress, high blood sugar, smoking … they can inflame your arteries.

  3. LeonRover

    Is it the abstract writer, the journalist or the deliberate failure to distinguish causation from ass-ociation?

    It would be a more realistic view of the study to hypothesize that healthy women exercise more, that healthy women show a reduced risk for development of cancers and thus the specific cancer mentioned in the study. The association measured is an association induced by the factor health on 1) the behaviour “exercise” and 2) on an outcome “development of cancer”.

    Just creased by this cartoon –

    It’s both the researcher, who imputed cause from a mere association, and the journalist, who failed to see the problem with imputing cause from assocation.

    Outstanding cartoon.

  4. Scott

    As many people who have known me for a long time can attest, when I changed my diet, my personality changed pretty drastically. I went from being explosively in your face to fairly easy going.

    But along with the change in diet was a natural lowering of my blood pressure. Additionally, I’m no longer in continuous pain from arthritis. I can easily believe that the rampant rudeness of people in general would decrease dramatically if we all started eating primally.

    Good point. I tended to feel tired and cranky far more often when I ate too many carbs.

  5. AllenS

    There is a correlation between exercise and health, but just as they have the exercise/obesity causation backwards, so does the medical community have the health/exercise causation backwards. IMHO, exercise doesn’t make you healthy, rather healthy people exercise because they feel good. Even the controlled exercise studies fail because they always let the exercise and non-exercise groups self-select. If you don’t want to exercise they’ll let you join the non-exercise group and visa-versa. They claim to do this so that the groups have stricter adherence to their assigned level of activity. But it clearly pollutes the data.

    BTW, the sign should read “Canned Fruit” and “Canned Vegetables.”

    Yes, whoever put up the grocery sign could use some remedial education.

  6. Nathaniel

    You just have to love correlations. There is literally a peer-reviewed study out there which found that gray hair, baldness, and wrinkles are associated with increased risk of heart disease.

    I’m surprised the media hasn’t been telling us to buy hair dye, toupees, and botox in order to protect ourselves.

    Yikes. I use the example of gray hair being associated with heart disease to show what’s wrong with association studies.

  7. Ms. X

    I’ve never driven in LA, but I did spend a week in Boston once. I was white knuckled driving around the first day, then I figured out the way to survive, even flourish, was to simply throw away everything I knew about good driving and drive as badly as I possibly could. It worked. It was unstressful, badness became a force in and of itself.

    My approach in L.A. was to become the nicest driver on the road. Want to cut me off because you’re in a huge hurry? Be my guest. I’ll wave when I pull up behind you at the next red light.

  8. Crusader

    Ever since I blew up in my car at a couple of left-wing talk radio hosts, I told myself NEVER again. I was only hurting myself and rewarding my enemies. Let them harden their arteries by feeling the RAGE all the time. I won’t.

    That’s why I prefer the humorous approach. Laughing is good for you.

  9. Jan

    Can a vegetarian? Picking the meager flesh of their skinny carcasses, then you’d have to drag out the pressure canner…seems like too much trouble to me. I wonder what a good vegetarian-based stock would taste like.

    I suppose there’s enough meat for soup.

  10. gallier2

    Going low-carb was the best thing that happened to my driving, strange but true. I commute on the A31 between Thionville and Luxembourg and this section is one of the heaviest traffic axis in Europe (the petrol in Luxembourg is 20/30 cent a litre cheaper than anywhere else in Europe, so all the lorries try to refuel there). In old days BLC (before low-carb) it happened often that my concentration went away, or if I encountered a dangerous situation, that my pulse went high and BP was enormous. After low-carb, nothing could get me out of my cool. A dangerous situation, direct reaction, braking, turning whatever, but no acceleration of pulse, no physical reaction, nothing. That was a change. And the tendency that I had to get sleepy after 1 hour driving was blown away, never happens again. Even if I’m not a rigorous paleo/low-carber (my weight creeped back to were I started because of to mych exceptions) this was one of the changes that remained even after re-eating too much crap (the other things being the afternoon sleepiness that has wanished and the gastric reflux which is only a shadow of its former nastiness).

    Sounds familiar. Avoid the blood-sugar roller coaster, get your mood-enhancing fatty acids and amino acids from your diet, and it’s amazing how you can feel calmer and more energetic at the same time. We should make Tony the Tiger the poster boy for road rage.

  11. Lori

    I drove all over LA last year and didn’t think it was any worse than driving here in Denver. I keep my printed street finder in the car, figure out where I’m going before I start, put on some good music, merge properly, yield to everyone who wants to get in front of me, go with the flow of the traffic, and don’t expect to get across town quickly. I don’t race to the next red light and don’t get excited about anything. It makes driving a generally stress-free experience. Since strokes run in my family, it just might help keep me from having one.

    You did it the right way; become the nicest driver on the road and don’t expect to get anywhere quickly.

  12. Gwen

    Darn it! LeonRover beat me to it! I had that comic tagged in my Delicious list and was going to tell you about it when I read this blog post, but he already did it! There’s already a new push by law-lickers to get the FDA to regulate food colorings because they cause ADHD (according to some specious observational evidence).

    Good grief … must be the food coloring, all right, because we wouldn’t want to blame the sugar and flour in the foods we decorate. (If any of you out there are using food coloring to decorate your meats and vegetables, you should probably stop just in case.)

  13. Jan

    The potato guy has already gone off the diet. He says on the home page that he’ll eat ’em pretty plain: “No toppings, no chili, no sour cream, no cheese, no gravy, just potatoes and maybe some seasonings or herbs and a little oil for some of the cooking.” In the comments section of his “blog” he stated on Day 1 he went to dinner with his family for his wife’s birthday and had mashed potatoes and steak fries. Mashed potatoes have butter and milk in them – or in the case of a restaurant, margarine and evaporated milk.

    I’d love to see his before and after numbers re: weight, blood sugar and lipids, but I bet we won’t hear anything but “Oh, I feel wonderful and still love potatoes!”

    Since he’s doing this as a P.R. stunt to promote potatoes, I have to wonder if he’ll confess to, say, screamingly high blood sugar or triglycerides.

  14. emily deans

    (The artificial food colorings are especially bad and seem to promote a histamine response that does lead to increased ADHD symptoms separate from sugary carbs)

  15. lovely lentilla

    I like the Breast Cancer & Brassiere Wearing association myself… to get the point across that association is not causation.

    That’ll work.

  16. Levi Wilson

    If living in LA has taught me anything its that the lane you’re in will eventually move and the lane that’s moving will eventually stop. And you don’t know when any of it will happen.

    And the sooner you learn that, the less stressful the drive.

  17. Carol Bardelli

    (If you’ve never driven in Los Angeles, a significant portion of the drivers believe if they conduct a dozen death-defying maneuvers every minute or so, they’ll eventually discover the secret fast lane.)

    I really did laugh out loud on that. I used to live in LA. That’s why I live in rural Nevada. We drive like demons here, too but there’s only one lane in any direction. I drive a full size V8 van, I win by size and speed alone. Pass me. Go ahead.


Leave a Reply

Your email address will not be published.