Friday, July 9, 2010

"Epidemiology is Bogus"

The paleoblogosphere is humming with excitement the past few days over the glorious work done by raw food blogger Denise Minger in her personal examination of the China Study, a large data set of epidemiology studies used by researcher T. Colin Campbell to formulate his book The China Study: The Most Comprehensive Study of Nutrition Ever Conducted and the Startling Implications for Diet, Weight Loss and Long-term Health. In it, Campbell comes to the conclusion that avoiding animal protein is the best way to avoid all sorts of diseases of civilization.

Denise's post uses the same China data set to implicate wheat as a major factor associated with heart disease and cancer, not animal protein. Kurt Harris of PaNu makes a point in his analysis of Denise's work that is the closest to my own struggles combing the literature - association studies are interesting, but troublesome. Conclusions from such studies should be viewed with a furrowed brow. "Hmmm, that's intriguing. I wonder why red-tailed baboons who eat less algae live longer than the red-tailed baboons who eat more algae? Why don't we do a prospective randomized controlled trial of red-tailed baboons and algae eating to sort that one out?" Because associations always come with confounding factors - turns out red-tailed baboons who eat algae love race car driving, and have you ever seen a baboon wearing a seat belt? Sophisticated statisticians will try to account for all these "confounding factors," but such a task can be simply impossible when examining a complex system such as society, or the biochemistry of the human body.

The psychiatric literature is loaded with brief, often useless, short-term randomized controlled studies of lesser and better quality and association studies. It seems the nutrition literature is even worse. Large, long prospective trials of good quality are horrendously expensive, and may take decades to do properly.

This is why I feel the healthiest and most sensible way to eat is based on an evolutionary paradigm, and that evolutionary-based lifestyle measures (though I haven't blogged too specifically about them yet, as the nutrition aspect is my primary personal interest, these measures would include regular exercise, meditation (as a proxy for the intense in-the-now concentration we used to use for hunting and gathering), proper sleep, working with the hands, and various other social/fun stress-reducing activities) are likely to be most effective for our modern, unsettled minds. It is not that there is a huge amount of science data backing that up - it is that until we have exhaustively proven otherwise (which so far, in my mind, we haven't), at least scientists can concede that our exceedingly complex human design is based on adaptations for our ancestors' lives.

For public health and sanity, I believe in an evolutionary viewpoint. And animal protein :)


  1. Well put! The science is coming around more and more to support the Ancestral/Evolutionary framework, but it should still be the default starting point/theoretical framework from which to approach all aspects of human health (physical/mental/social/etc.). And Denise did a masterful job!

  2. I need to do some research as to how long paleolithic man lived, on average.

    Genes only "care" about one thing, and that is propogating themselves. If a gene doesn't express itself until well after the gene had propogated, the gene will not be selected out of the gene pool.

    Examples of genes that express themselves well after the reproductive years are over include those for Alzheimers disease and Huntingtons chorea.

    If paleolithic man only lived to thirty, it wouldn't matter if he at McDonalds and Taco Bell four times a day for 25 years.


  3. Steve - Paleolothic man had an average lifespan of 33 or so. There are a number of sources, but most will put you right around that number. However, once a hunter gatherer lives to 35, he or she is very likely to live to 70 - their average lifespan was so short because of high infant mortality, accidents, and things like homicide. The "evolution only cares if you live until your genes propagate" is the same argument used against the idea of a paradigm of a Paleolithic diet by T Colin Campbell himself in his "debate" with Loren Cordain. But consider - humans have grandmothers. We are the only primate species whose women are likely to live 20-30 years PAST menopause. In tropical areas, grandmothers gathered a lot. Women with long-lived mothers could have babies closer together, as grandmother would provide far more for the family than she would eat. Primates have 4-6 years between babies - humans had about 3. Therefore, at least in a womans case, there is no question in my mind that we have been selected for longevity. While there is some question as to when menopause occurred paleolithically, in modern hunter gatherers it is still around 50. Consider also that the first bronze age agricultural humans had a life expectancy of 18, their brains were 11% smaller (and still are), and the life expectancy did not surpass the grand old 33 again for 10,000 years, with the invention of antibiotics and vaccines.

  4. (in more "hunting-heavy" areas, such as the far north, grandmothers were sometimes euthanized because they could not provide for their own food - ancestral ways were not always pleasant, and I certainly have no romantic illusions of Paleoloithic man and what it was like. I know that modern hunter gatherers do not have diseases of civilization if they eat traditionally - no wheat, fructose, or vegetable oil. In general, nutrirent rich food, obviously choosing the best of what they could get their hands on. I can emulate that in modern life without getting into Paleolithic reencactment. )

  5. Emily & Steve,

    This was a recent Science Daily report that touched on exactly the comments Emily made regarding grandmothers:

  6. Thanks Jamie. Humans are a "K-selection" species (as opposed to r-selection). We role our genetic dice on a few offspring and spend a lot of time nurturing them. R-selection species like certain octopi reproduce once, have a gazillion babies, then die. As a k-selector, my genes are not only "interested" in whether I have offspring at all, but whether I'm around long enough and have the mental health and physical stability to raise those offspring so that they have healthy offspring of their own. This is, in part, why I hang my hat on evolution as an appropriate paradigm to base as a "default" for human health, though of course careful study of the scientific evidence to support or refute the hypothesis is what health science is all about. I hope.

  7. Denise's analysis of The China Study is heavily flawed and therefore invalid. Debunked by a cancer epidemiologist...

    Here it is...

    The proper testing procedure as stated by an expert on analysing stats...

    Campbells response to Denise..

  8. This is exactly why I think association studies should be taken with a zillion grains of salt, and not be the basis of our USDA nutrition recommendations. The critique and Denise's reply to the critique, in fact, points out the most astonishing thing (if it is true) about Campbell's results in the "China Study" - that they seem to match the raw data, not statistically corrected data. While Denise no doubt made some mistakes in her analysis (she was working alone, is not an epidemiologist, and burned the midnight oil), I can't imagine that Campbell at Cornell didn't have access to world-class statisticians. As far as I know, he has not addressed that particular criticism.

    Unfortunately, I took the time to read some that thread on the 30 Bananas a Day site, where various people vehemently questioned Denise's agenda (when I felt her criticism had a remarkably neutral tone), and derided her for her "failed veganism due to anorexia." Now I did not comb much further, but in the next several pages I didn't see anyone call out such posts for the unfortunate content. Of course everyone has a right to express their opinions, distasteful or not, but personally, I don't think that directing people to those posts will help support the vegan cause.

    In the end, I think we all have the same goal, herbivore, carnivore, and omnivore. We want good analysis of the scientific information we have in order to make informed decisions about what seems to be the best for our health and our morals. Debate either way is productive and meaningful in our search for health. The morality debate is also important - but perhaps not as meaningful for public health.

  9. Gday crew,nice blog.

    How come NONE of these pro meat bloggers have any REAL muscle/cardio with all that protein talk? :)

    Come and see if ANY of you guys can out bench press/dead lift us at

    Here are more websites for the doubters.

    Mike Arnstein ran a 2:28 marathon this year at Boston. He is the FASTEST runner in the raw food movement today. Long time vegan and now powered by sweet fruit. How come there is no competitive athletes eating this 'paleo fat diet?' Please shut me up and show me cos Im sick of seeing cardio and muscle deficient paleo crew trying to debunk the china study that us elite athletes are thriving on.

    Can you debunk me with a high fat eating paleo athlete that is a national level runner, cyclist, power lifter, UFC fighter like us vegans clearly have provided.

    Didnt think so.. :)

    Love, peace and banana grease.


  10. Actually, an ultimate fighting championship isn't the worst way to figure out this question of optimal human diet.

  11. Hi Dr. Deans,

    In addition to Denise's spectacular and awesome expose of the China Study, you might like my latest post on Campbell's animal research. Experimental science unlike epidemiology is not bogus :) but as it turns out Campbell's description of his animal research is not quite showing the whole picture. Here it is: