Blog | Friday, September 28, 2018

Nutrition and the art of disparaging science

It is back to school season, so let's start this with a pop quiz:
1. Which are more nutritious, garbanzo beans or jelly beans?
2. Which are more nutritious, walnuts or donuts?
3. Which are more nutritious, lentils or lollipops?
4. Which do you think is the more wholesome, nourishing breakfast of champions: steel cut oats and mixed berries, or pepperoni-stuffed-marshmallows?
5. Do you need a PhD degree, MD degree, or the ability to cite a specific randomized controlled trial to answer questions 1-4?

I trust you had no difficulty with questions 1-4, and answered “no” to question 5. Unless, that is, you have a MD or PhD, and are among the high-profile contrarians claiming we know nothing at all about nutrition because the scientific methods of nutritional epidemiology are suspect. I do at times wonder if members of this guild follow their own logic where it leads, and let their kids have jelly beans and donuts for breakfast every day, lollipops for lunch and dinner.

I rather doubt that (with one possible exception), in which case, and with all due respect to their nihilistic passions, their practice belies their preaching. Intentionally, or otherwise, they are dabbling in hypocrisy. We either do have some basis to know something about the basic care and feeding of Homo sapiens, or we don't. If we don't, then really, absolutely anything goes. Maybe a Coca-Cola/hot dog smoothie is just as good for us as, say, raw almonds.

In order to choose garbanzo beans over jelly beans, walnuts over donuts, oatmeal over pepperoni for improving diet and health requires acknowledging that we actually know the basics of better and worse nutrition, know them on the basis of common knowledge and sense. Science is part of that mix, but certainly not all of it, any more than we know to use water rather than gasoline for putting out fires on the basis of randomized trials.

That is the first, dangerous fallacy in the nihilistic nutrition narrative, that what we don't know from the right kind of science, we can't know at all. That's not true about water, fire and gasoline; breathing versus not; the likely fate of an apple we toss up into the air. It's not true about nutrition, either.

In other words, science itself is being misrepresented. Science is a method of our own devising to extend our knowledge and reach beyond the limits of our native perceptions. We need science for microscopes; we need science for telescopes. We don't need science to see the writing on the wall. (OK, we needed science to devise the alphabet, manufacture the chalk, and build the wall; but you know what I mean.)

We rely on science to know the molecular composition of sandpaper; to know that it is rough, we just need to touch it. Science can reveal what is not already obvious; the limits of science do not preclude the truths that are obvious without it. Apples tossed up fall down, they don't float away. We knew that before ever Newton bothered to ask: why? Answering why, well, that calls for science.

Nutrition is the same. We are well advised to consider that every wild species on the planet knows what to eat absent recourse to randomized trials, and our species was no exception before we invented science, food science included, and simultaneously fed ourselves into epidemic obesity and chronic disease, while talking ourselves into a state of perpetual, gullible pseudo-confusion.

There is, simply, and from my perspective sadly, a booming cottage industry in the art of disparaging all aspects of nutrition science. I write to label it a dark art, a cheap art (like paint-by-numbers), an easy art (like being a food critic without knowing how to boil an egg) and to oppose it, because what we know reliably about nutrition and lifestyle is enough to eliminate 80% of premature death and chronic disease. This is entirely about dietary patterns, not at all about the isolated (and frankly, unknowable) contribution to health of some number of hazelnuts (see footnote 1), or cups of coffee.

The costs of excessive parsing, prevarication, procrastination, and profiteering are, truly, measured in years lost needlessly from lives, life lost needlessly from years. I reflect on that almost daily, and have for the many years of my career in public health. At times it makes me cry. At times it makes me angry.

The prevailing narrative goes something like this: Nutrition science advised us to cut fat, or add oat bran, or steer clear of gluten. It told us that if we did that, all would be well. We listened, but here we are, fatter and sicker than ever. So, actually, nutrition science got it all wrong, and is entirely unreliable.

But every element in this narrative is manufactured nonsense. Nutrition science showed, for instance, that diets high in saturated fat from meat and dairy produced less good health outcomes in general than diets based principally on plant foods, and thus, lower in saturated fat, and often, total fat. That was, and is, entirely true, and robust across all forms of research, from nutritional epidemiology at the scale of whole populations, to basic studies of mechanistic pathways.

The message about food sources and their fat content was distorted into media sound bites, because that's business as usual. So we went from science that said “eat less beef, more beans,” to distortions that said: “cut fat.” Then industry got involved, and contrived an inventory of low-fat junk food that never existed before. So a message in the research literature about dietary patterns was adulterated into, “Just eat Snackwells, and all will be well.”

And then, we blamed our unfortunate results on “nutrition science.” But for one thing, the science never said anything like “eat low fat junk, and all will be well.” That message was generated by a toxic game of telephone with many elements of our culture complicit in, and profiting from it. And besides, we never took the advice anyway. Fat intake in the U.S. never went down; calorie intake went up around it.

But this is not a tale about dietary fat; the same thing happened when “just cut carbs” was the message du jour. The science never indicated that vegetables, legumes, whole grains and whole fruits were problematic, but these are “carbs,” all. The original message was aimed at reducing intake of refined grains and starches, and added sugar. But it, too, was adulterated by the usual suspects into: “just eat low carb brownies, and all will be well.” There was never any hint of science supporting any such nonsense, and the results, predictably, were: more of us fat, and more of us sick.

So here we are, in the aftermath of the latest high-profile rebuke of nutritional epidemiology, with many champions of nutrition, my friends and colleagues among them, celebrating the reality check. But what they seem to be overlooking is the calamitous eviction of baby with bathwater, and the conflation of cultural boondoggles with flaws configured into the science.

Science was right about the specific effects of oat bran on serum lipids; science was right about the implications of gluten ingestion for those actually sensitive to it. In the case of almost every historical folly or cultural fad, there was science that actually made sense, and remains right today, and then a toxic game of “telephone,” turning the message into profitable nonsense. Now, we have appended one more step at the end of this cascade of distortion: we have scientists maligning the science that got mangled by it.

The latest broadside is noteworthy not only for the pedigree of its author, but also for how it is glaringly wrong when right, and right when wrong. The commentary is wrong to imply that we need new or better studies for a basic, actionable understanding of the linkages between overall dietary patterns and the health outcomes that matter most, but entirely right that we need more and different research if we really are committed to chasing after the “active ingredients” in such assemblies. (More on “active ingredients” below.)

The commentary is entirely right that we know very little about the effects of any given food, independent of dietary pattern, on health outcomes. But it is entirely wrong to suggest that prior studies indicate an increase in life expectancy with every hazelnut consumed daily. No such claim is made in the study cited, and the translation of those data and their reasonable interpretation (e.g., an optimal intake of whole grains, vegetables, fruits, nuts, legumes, and fish, as well as reduced consumption of red and processed meats and SSBs, can lead to an important decrease—by ~80% in the relative risk of premature death when compared with intake“) into ridicule about life expectancy is itself a misrepresentation (and perhaps misunderstanding) of epidemiologic research methods. Yes, of course, nutritional epidemiology becomes flammably nonsensical if you turn it into a strawman before striking your match (see footnote 1).

Science, however potent, is always imperfect, and nutritional epidemiology subject to many more limitations than many of its other domains. But even so, science is not really the problem here. The conclusions of nutrition science have, for the most part, lined up behind a balanced assembly of mostly vegetables, fruits, whole grains, beans, legumes, nuts, seeds, and plain water across research methods, study populations, countries, continents, and decades. Adulterated though they may be by political motivations, even the Dietary Guidelines for Americans have been more right than wrong since their inception 40 years ago, and more alike than different across the iterations since. The apolitical advisory committee reports on which the guidelines are only partly based, even more so.

So, it's not really nutritional epidemiology, despite its genuine limitations, that's the problem. Rather, our culture has CHOSEN to be perennially confused about nutrition because: (a) that's a license for individuals to eat whatever they want; (b) it's a license for authors and publishers to keep pumping out quick-fix, fad diet magic; (c) it's a license for Big Food to keep putting lipstick on new pigs and pretending they think it's good for us; (d) it's a license for ever shifting, perky advice on morning shows and in health and beauty magazines; and (e) it's a license for Big Pharma to keep fixing what diet is unnecessarily breaking.

On the chance you are not convinced, I've thought of one fairly obvious and quite robust analogy: education. Let's imagine our culture chose to treat education the way we choose to treat nutrition.

As things are, we seem to have a shared, robust faith in the proposition that an education, from learning the alphabet to reading, from counting to addition, and everything that follows, is favorable to one's life trajectory as compared, on average, to no education at all, and resultant illiteracy and innumeracy. We don't tend to hear counter-arguments, telling us that “education science” is hopelessly impaired, and that we might well be better off with no education at all.

But if education were like nutrition, then doctoral-level experts in educational contrarianism would point out to us that we have no randomized trials to prove the merits of K-12, and absent such trials, in which we randomize thousands of comparable children to a full education or none at all, we might as well assume that NO EDUCATION is better.

Further, we might hear arguments from the specific to the general. In other words, we would hear that we cannot possibly know the exact benefits to one's life trajectory of, say, calculus, and thus, we must renounce all we think we know about “education” in general.

It would, indeed, be very difficult to isolate the benefits of studying calculus, just as it is very difficult to isolate the benefits of eating kale. Those who study calculus differ systematically from those who don't. Those who do well in calculus may have a better education in general, a more supportive family, be more affluent, or maybe even just be smarter. Until or unless we control for all of that and more, we would be hard pressed to attribute particular advantages across the life span to calculus. So, too, for hazelnuts.

I am not sure why we don't have a cottage industry in bashing the “science of education,” except for the obvious: probably no one has yet figured out how to profit from the sale of “alternatives to education.” If ever anyone does, perhaps those alternatives will proliferate just like fad diet books.

There is one final, critical dollop of hypocrisy atop the “nutrition science should start over from scratch” argument. In general, those most adamant that what most of us think we know (i.e., that we should eat food, not too much, mostly plants; let's call that “A”) is in doubt, are equally adamant that they KNOW the alternative truth (i.e., that we should eat more bacon, or put butter in our coffee; let's call that “B”). But this argument is to logic what…a bicycle is to a sea snake.

Uncertainty about the benefits of spinach does NOT mean that salami must be good for us, any more than lack of randomized trial data about water for putting out fire PROVES that gasoline must be better. Rather, there are three logical implications of emphasizing uncertainty in all dietary matters:

What we think we know, we know with considerable uncertainty at best; so focus on the big picture (i.e., dietary patterns) rather than banking too much on any details (i.e., nutrients, ingredients). By way of analogy, it's entirely obvious that climate change is advanced, and our carbon emissions are the reason for it, yet practically impossible to link the weather in any one place at any one time to any given drive or flight by any given individual.

Uncertainty simply means that whatever we choose, we could be wrong, so it always begs the question: which way would you rather be wrong? In my view, we have no meaningful uncertainty about the relative merits of plant-food-predominant versus animal-food-predominant diets for human health. But if we did, the even less contentious implications for planetary health would provide a very clear mandate for the “better guess.”

Most saliently: what's good for the herbivorous goose is good for the carnivorous gander (or something like that). In other words, if the state of nutrition science leaves us with substantial uncertainty about what we think we know, how can those arguing for alternatives be so certain, and adamant? It's not as if they have an entirely distinct approach to research. Rather, they seem to embrace all the same studies, with all the same flaws, provided they like the conclusions. Valid or not, it would at least be internally consistent to say: nutritional science is suspect so I'm not at all sure what's best to eat. It is not at all internally consistent to say: nutritional science is suspect, but I know the truth anyway, so eat more meat, or, don't worry about soda.

Rarely, there are critiques of nutrition science by those genuinely knowledgeable of the content, and experienced in the particular requirements for researching it. When such enlightened critics speak up, by all means listen to them. You will note that they consistently, explicitly differentiate between the many details we don't and in some cases can't know, and the general patterns we know with considerable confidence.

Valid criticism of science, in nutrition or any content area, requires such careful attention to the linkages between questions and answers. This is where active ingredients pertain. Consider that our well-founded confidence in the use of water for putting out fires is despite a lot of unstudied details we do not know: What pH of water is optimal? What mineral content is optimal? What water pressure, hose caliber, or parabolic angulation? What water temperature? We can, and do, know the general merits of dousing fire with water without knowing these pertinent details. But were there profit in it, we might read a paper telling us all the things we don't know about extinguishing fires, and inviting us to conclude we therefore know nothing. But here it's obvious that unanswered questions about details don't question the answers we have about the general pattern. Nutrition is the same.

Finally, there is one more tragic ingredient afloat in this toxic brew of nihilistic pseudo-confusion about diet: with each precautionary critique of nutrition science, we are learning to see clearly only backwards, not forwards.

We use doubts about science to disparage what we thought we knew up until yesterday, but FAIL to apply it when headlines today tell us about a study saying just what we hoped (or perhaps hoped not) to hear. If nutrition science at the level of every detail is hard (it is), and if what we know about details is limited (it is), and if the isolation of effects to lone foods, nutrients, or ingredients is nearly impossible (it is), then it means we should NEVER get overly excited about any one diet study. It means we should base our judgments ever more cautiously on ever greater assemblies of evidence that stand up to ever longer tests of time. This, at least, is what it would mean if there were intelligent life down here.

Instead, we get hyperbolic headlines about the studies critiquing other studies, rush to flush baby and bathwater alike, and then…immediately overreact to the next cycle of hyperbolic, distorted nutrition nonsense as if the precautions pertain only up to the minute we hear about them. They are apparently abandoned with the next news cycle.

If you want to doubt nutrition science, have the intellectual honesty and integrity to doubt it all: yesterday's, and tomorrow's; the conclusions you like, and the conclusions you don't. The fad diet popular in any given moment is inevitably predicated on vastly LESS science than time-honored principles of generally healthful eating. Legitimate concerns about the robustness of evidence should defend against an endless cycle of silly fads. Instead, we seem inclined to apply our concerns selectively, raising doubts about what is almost certainly true, while failing to scrutinize the almost entirely unsubstantiated if we happen to find it…appetizing.

The repudiation of nutrition science has emerged as both cottage industry, and something of an art form. This is happening now perhaps because the importance of nutrition to all of medicine is simply no longer possible to deny, or avoid. Poor diet is the leading cause of premature death in the U.S., so all eyes, even formerly reluctant eyes, are directed this way.

Perhaps, like all movements, the movement into evidence-based medicine has taken on a veneer of religious fervor. After all, there is no decisive evidence that any particular approach to evidence-based medicine reliably improves outcomes, but there is the conviction that it does. Competing factions each seek to elevate a particular view of evidence above all others, and nutritional epidemiology has become grist in this mill.

Medicine has always been characterized as both art and science, so the subjective elements allow for a range of opinion. Views of expertise are shifting, and potentially, devolving. Thus, those outside of nutrition feel entirely comfortable disparaging a field in which they may have neither particular content knowledge, nor research experience. And of course, everyone has the megaphone of Internet access, and is eager to be heard above the din. Contrarianism is a highly effective method. Perhaps that, too, is fuel for this fire.

Whatever the reasons, on this bonfire of competing vanities and fractious voices, diverging agendas and divisive perspectives, we are burning down the forest we can't seem to see through the trees, burning down the long-languishing opportunity to apply what we know most reliably about food and health to add years to life and life to years, parsing the particulars we know much less well as we go. We just keep burning it all down. The carbon emissions alone are reason enough to stop.

Footnote 1 The commentary that prompted this column cites a meta-analysis to argue that adding hazelnuts to our daily diets directly increases life expectancy. However, the meta-analysis, which reports shifts in the relative risk of premature death with shifts in food intake, neither makes any such claim, nor do the results reported even support such a claim. Changes in the relative risk of premature death do NOT translate into years added to life expectancy. Consider, for instance, that I can expect to reduce my risk of premature death by as much as 80% by avoiding tobacco, exercising regularly, eating optimally, etc. That certainly does NOT mean that I can expect to live 80% longer than the mean life expectancy: [78 + (0.8×78)]=140. The math used in the commentary would lead to just such a preposterous conclusion. Further, reductions to the risk of premature death would not add ANY years to life after age 80, because death after age 80 is not premature; it is above the mean life expectancy in the U.S. (for men, at least). Finally, the association of routine intake of nut consumption with reduced risk of premature death is entirely consistent with a very considerable expanse of research, including not just observational studies, but also short term mechanistic studies, and intervention trials, including RCTs.

David L. Katz, MD, FACP, MPH, FACPM, is an internationally renowned authority on nutrition, weight management, and the prevention of chronic disease, and an internationally recognized leader in integrative medicine and patient-centered care. He is a board certified specialist in both Internal Medicine, and Preventive Medicine/Public Health, and Associate Professor (adjunct) in Public Health Practice at the Yale University School of Medicine. He is the Director and founder (1998) of Yale University's Prevention Research Center; Director and founder of the Integrative Medicine Center at Griffin Hospital (2000) in Derby, Conn.; founder and president of the non-profit Turn the Tide Foundation; and formerly the Director of Medical Studies in Public Health at the Yale School of Medicine for eight years. This post originally appeared on his Linked In page.