Settled Science

I mostly ignore, and tend to be skeptical of, most pronouncements on foods that supposedly kill us and foods that are supposedly superfoods.  I have a solid love of meat and have never let the fear of saturated fat stop me from enjoying a good steak from time to time.

I had heard that a lot of the "settled science" on saturated fat was iffy but I had no idea it was this bad.

Our distrust of saturated fat can be traced back to the 1950s, to a man named Ancel Benjamin Keys, a scientist at the University of Minnesota. Dr. Keys was formidably persuasive and, through sheer force of will, rose to the top of the nutrition world...

As the director of the largest nutrition study to date, Dr. Keys was in an excellent position to promote his idea. The "Seven Countries" study that he conducted on nearly 13,000 men in the U.S., Japan and Europe ostensibly demonstrated that heart disease wasn't the inevitable result of aging but could be linked to poor nutrition.

Critics have pointed out that Dr. Keys violated several basic scientific norms in his study. For one, he didn't choose countries randomly but instead selected only those likely to prove his beliefs, including Yugoslavia, Finland and Italy. Excluded were France, land of the famously healthy omelet eater, as well as other countries where people consumed a lot of fat yet didn't suffer from high rates of heart disease, such as Switzerland, Sweden and West Germany. The study's star subjects—upon whom much of our current understanding of the Mediterranean diet is based—were peasants from Crete, islanders who tilled their fields well into old age and who appeared to eat very little meat or cheese.

As it turns out, Dr. Keys visited Crete during an unrepresentative period of extreme hardship after World War II. Furthermore, he made the mistake of measuring the islanders' diet partly during Lent, when they were forgoing meat and cheese. Dr. Keys therefore undercounted their consumption of saturated fat. Also, due to problems with the surveys, he ended up relying on data from just a few dozen men—far from the representative sample of 655 that he had initially selected. These flaws weren't revealed until much later, in a 2002 paper by scientists investigating the work on Crete—but by then, the misimpression left by his erroneous data had become international dogma.

In 1961, Dr. Keys sealed saturated fat's fate by landing a position on the nutrition committee of the American Heart Association, whose dietary guidelines are considered the gold standard. Although the committee had originally been skeptical of his hypothesis, it issued, in that year, the country's first-ever guidelines targeting saturated fats. The U.S. Department of Agriculture followed in 1980.

Don't these guys know this is settled science?  These saturated fat skeptics must be in the pay of big cattle.

The cherry-picking and small sample sizes are unfortunately a staple of science, but I particularly laughed at the practice of assessing meat consumption during Lent.

  • Daublin

    The commonality is that it's a field where nobody really knows. The uncomfortable truth is that for many things we'd like to know more, the most informed people in the world simply don't know. Not even after doing science to it.

    Nutrition is the poster-child example I use for these. Nutrition science is strong on deficiencies and on calory balance. For subtler issues, such as supplements, or different kinds of fats, there is just too little evidence to make strong conclusions. The best studies and best informaiton are therefore weak, but they get promoted because they are the best of a sorry lot.

    Sometimes the *best* is not *good*.

  • stanbrown

    Nutritional science is no worse than cancer, biotech, climate or any other science. It is probably better than economics, psychology, sociology and the rest of the social sciences. The problem with all academic research is that no one replicates anyone else's work. That minimal bit of quality control should be required before any such 'research; is ever used for making policy.

  • rxc

    If the push (or was it a nudge?) away from animal/saturated fats into vegetable oils with their trans-fat components, had been sponsored by an evil money-grubbing industry, you can be sure that legions of public interest lawyers and activists would be calling for their heads and suing them into bankruptcy. However, since it was a pet project of the progressive movement, which saw lots of possibilities for this experiment to help other aspects of their agenda (animal rights, lower population, less intensive agriculture, overall smaller human footprint, and, of course, better control over the populace), we hear not a word, not a peep of remorse or apologies.

  • MingoV

    Here's the scoop on saturated fats. Epidemiology studies by the CDC and others showed a correlation between saturated fat ingestion and atherosclerosis. Then it was determined that the correlation was because saturated fat increased cholesterol which increased atherosclerosis. But, here's the kicker, in 95% of American adults, there was almost no correlation between saturated fat or cholesterol and atherosclerosis. Almost all the correlation came from 5% of the population. That 5% had a partial or full-blown case of one of numerous inherited diseases of lipid and cholesterol metabolism. The CDC and other government health groups knew this, but they recommended low saturated fats and cholesterol to everyone figuring that it wouldn't hurt. They were wrong, of course. The recommended reduced fat (and therefore increased carbohydrates) diet resulted in a big increase in overweight or obese adults.

    They did the same thing a decade earlier with dietary salt. They knew that for the vast majority of people, salt intake was not a risk factor for high blood pressure. Low salt diets also harmed some people, particularly those whose work or play resulted in heavy sweating. It also hurt people taking lithium, making them more susceptible to muscle cramps and cardiac arrhythmias.

  • NL7

    Also, dietary cholesterol does not produce blood cholesterol. Simply eating cholesterol from a cow doesn't transmute into human cholesterol. In fact, the worst cholesterol types, such as triglycerides, tend to come from carbohydrate consumption. So a low fat diet, which is usually by necessity a high carb diet, is also going to be a high-triglyceride-producing diet. So a low fat diet is worse for the thing it supposedly addresses.

  • me

    Charismatic personalities and a subject matter of interest to everyone... with people being people. What could possibly go wrong? For some fun with fats, look up reddit.com/r/ketoscience.

  • marque2

    One interesting fact. Stearic acid - the prime fat in animal meats is actually a strong reducer of cholesterol. Look it up on the wiki or even the NIH.

    There is a good movie which covers all the fraud in detail - while debunking super size me at the same time. It is called "fat head" and is available free on Hulu.com.

  • marque2

    The real problem is that in order to get government grands you need to show something profound and hopefully useful to a government agenda. Much of this junk science is from Universities chasing government money and adjusting facts to make it look like there is an urgent tragedy about to occur so they can get more money to study it, while government agencies get one more aspect of control or a new revenue stream.

    We should probably cut a good 90% of government sponsored research.

  • Joe_Da

    There is some indication that Salt in the diet acts as a trigger to let you know when you have eaten enough.

    This brings a to mind an study done in the late 60's early 1970's with first graders whereby, the kids at a summer camp were allowed to serve themselves from a cafeteria line with good nutritious food and junk food, candy ice cream etc. On day one virtually all the kids ate the junk food, by day 2 or 3 virtually all the kids began eating the vegetables, and other food because that was what their bodies told them the needed.

  • kidmugsy

    I don't trust Science that isn't Science. To be trusted it must, preferably, be studied by controlled experiments. If those aren't available, then the next best thing is an observational study that tests a well-defined mathematical model of some particularly simple system: the Solar System filled that role for Copernicus/Brahe/Kepler/Newton. After those two, after a huge gap, come observational studies that (i) turn up a very large effect, (ii) satisfy the Bradford-Hill criteria, and (iii) lead to lab science that explains and supports the results: the deadliness of smoking cigarettes falls into that category. Much of front-page science is best assumed to be rubbish - sometimes it's mendacious rubbish.

  • FelineCannonball

    It is unfortunate that conventional wisdom (based on more than one study) was wrong for so long, but the truth on saturated fats has been out there for 20 years or so.

    Never trust dietary studies that use fewer than 30,000 subjects and study them for less than 20 years. Also make sure they don't over-interpret, misunderstand their own survey, or screw up the statistics.

    There are a handful of studies that can be trusted, starting with the Harvard nurses study. They indicate that calories from saturated fat raise risks for heart disease no more than calories from carbohydrates. Transfats, on the other hand, raise the risk considersbly. As does obesity. Polyunsaturated fats like fish oil do actually lower risks.

  • MingoV

    Nutrition science itself is fine. Misuse of data and selection bias (excluding reports you don't like) by governments and others is the problem.

  • Matt
  • HenryBowman419

    Gary Taubes book Good calories, Bad Calories discusses this topic in quite a bit of detail. Unlike Teicholz, he doesn't come to definite conclusions, other than the conclusion that many studies, notable those of Keys, were horribly botched. But, his book is chock full of detail and does not have a thesis other than "such studies should be done scientifically if possible".

  • AnInquirer

    In my limited education on nutrition in college, we were not referred to the Keys Study. Rather we were instructed to compare the U.S. and Japan. With a virtually non-existent red meat diet, Japan had very low incidence of heart disease, but as they became more Westernized and increased meat consumption, their incidence of heart disease has been rising.

  • Don

    Nutritional science bears more than a passing resemblance to economics and other sciences of complex systems for a reason. Not only do the variables you're trying to measure change, but the yard stick you're using to do that measurement also changes. What I've noticed is that this specific combination (large number of uncontrollable, and even unknowable) variables, combined with a lack of a "standard candle" for measurement purposes, leads to these sorts of issues.

    I think what is really needed is a new set of testing criteria for complex, highly variable systems, and a consistent method of its application, similar to what natural philosophers did when they applied deductive reasoning and observation to get "the scientific method".

    The problems seem to be so common, and so widespread across disciplines that they would appear to be systemic in nature.

    At least, that what it looks like from the cheap seats (there's not a lot of controversy in Comp Sci, just math ;^).

  • Don

    That's because they (the Japanese) ALSO increased their non-rice grain intake and intake of grain-based vegetable oils (corn, canola, etc.).

    While you were look a the one hand (red-meat), the other hand (vegetable oil) was palming the coin.