Junk Food: from Confusion, to Clear and Simple Truth

I was somewhat startled, and quite concerned, to learn directly from the source while sipping coffee together, that my wife was confused by a study on junk food, generating headlines such as: “Junk Food Not to Blame for America’s Obesity Epidemic.” I was startled because my wife, an expert cook and mother of 5, has become very knowledgeable about nutrition over the years in her own right; because she is highly intelligent; and because she is embarrassingly well educated, with a PhD in neuroscience from Princeton. In other words, if she’s confused, everybody’s confused. That’s why I was concerned.

There is no legitimate cause for confusion in this study. Rather, we have painted ourselves into a corner with a plethora of misguided “either, or” and “action, opposite reaction” approaches to nutrition that have generated a whole lot of heat while dousing the common light. This is fixable, of course, and the timing is good- because two noteworthy fixes are imminently in the offing. We’ll get back to those after addressing the study.

Let’s start here: in the contrived context of hyped up and dumbed down “either, or” approaches to diet, weight (and health), junk food causes obesity, or it doesn’t.

This study did not carry out any kind of intervention to answer that question. Rather, using a nationally representative sample of 5,000 adults from a federal survey, researchers looked at intake of candy, soda, and fast food and found little difference between the lean and overweight members of the group. They did, however, find particularly high levels of fast food intake, and the lowest reported intake of vegetables and fruits, in both tails of the bell curve: among the heaviest of the heavy, and among the skinniest. These two traits - underweight, and severe obesity - have long been strongly associated with adverse health outcomes. In those bad lands, junk food prevails.

But again, across the broad middle of the bell curve, lean and obese were both consuming candy, soda, and fast food at rather high levels, something we have certainly heard before. Does this then prove that fast food is unrelated to obesity, as the clamoring headlines seem to suggest?

Of course not, and a thought experiment will clarify quite readily.

Let’s go back in time to the 1950s and 60s when nearly half the U.S. population smoked. Now, imagine a nationally representative sample of 5,000 adults at that time, including a large number with, and a certain number without, smoking-related lung disease, such as emphysema. Smoking would be quite prevalent in both groups, because it was so prevalent in the entire population. Those results, if interpreted like the study about junk food, would result in this headline: “tobacco not to blame for America’s emphysema epidemic.” Given what we now know, and to put it in appropriately blunt terms: that fails to pass the sniff test. It smells bad, like tobacco smoke itself, because it is bad; it’s bunk.

Where is the problem with that study? Everywhere. 

Unless a “toxin” of any kind is so potent that it produces obvious and nearly immediate harms in absolutely everyone exposed, there will be variation in the effects observed. Some people will be more vulnerable, some less so. Some will succumb quickly, some slowly. Some will show one kind of adverse effect, some will show another.

This is all true for tobacco, and in no way attenuates our certainly about its implication in emphysema, lung cancer, and a long list of other awful things. If everyone in the population smoked, a snapshot of the population at any given time- and that’s exactly what the NHANES data used in the junk food study provide- would show both sick and healthy smokers. The interpretation that tobacco is thus exonerated is not just wrong, but flagrantly and self-evidently so. Rather, it means that: (a) not everyone gets sick from tobacco at the same time; (b) not everyone gets sick from tobacco in the same way; and (c) a comparison of individual health within a population can result in missing the larger truth about the health of the whole population. In a classic illustration of missing the forest for the trees, a compare-and-contrast between apparently sick and apparently healthy smokers in a given population would fail to reveal that the entire population had much more disease than a comparable population of non-smokers.

This is exactly the story for junk food. It doesn’t make everyone fat, but we - as a nation - are a whole lot fatter than we would be without it. It doesn’t make everyone sick in the same way. In fact, an emphasis on the adverse effects of “thin obesity” is directly germane. Some people with the most dangerous variety of “obesity” do not have an elevated BMI, the measure examined in the study. Rather, they have a normal BMI, but excess fat around the middle where it does the most mischief. What dietary pattern is associated with that ominous body composition? You guessed it: a junk food laden diet, otherwise known, to our shame, as the typical (or standard) American diet.

There is another important limitation to the study I have not seen addressed. The researchers defined the trifecta of soda, candy, and fast food as “junk.” Alas, the bounds of comestible American junk extend far beyond those narrow confines. What about breakfast cereals with some kind of sweetener as the first ingredient? What about multicolored marshmallows as part of a complete breakfast? What about cheese puffs and pork rinds; donuts and muffins; toaster pastries, Snackwells, Doritos, and deli meats? 

NIH data suggest that roughly half of the calories in the diet of the typical American child come from this broader swathe of foods reasonably catalogued as “junk.” When an exposure is this universal, it becomes virtually impossible to determine its overall impact on a group by comparing those succumbing, to those managing to resist. Another thought experiment should convince you.

This time, we compare survivors and those not so fortunate among the Titanic’s passengers. Since exposure to a ship wreck in the North Atlantic did not differ between those who died and those who survived when the ship went down- apparently North Atlantic ship wrecks are “not to blame” for drowning. Who knew?

There are, just the same, two genuinely useful messages in the study’s mix.  

The first is that, despite the strange cottage industry that has sprung up to argue otherwise, calories count. In a reprise of the notorious “Twinkie diet,” the study showed that some people with a very high intake of junk food were skinny, not fat. If total food and calorie intake are low relative to the body’s metabolic demands for energy, the result is weight loss, not gain, no matter how bad or good the fuel in question. Conversely, studies have shown that overweight results when even high-quality fuel is over-consumed.

Calories count. But of course, so does the quality of food. The fallacy propagated by a noisome minority is that there is any need to choose between the two. In general, the higher the quality of one’s food choices, the more readily one controls the quantity of calories it takes to feel full and satisfied. The more dubious one’s food choices, the more one plays into the designs of Big Food to maximize the calories it takes to feel full, as so brilliantly chronicled for us all by Michael Moss. Quality and quantity both matter, and the best way to manage the latter is to attend to the former.

The second important message in the study is that the BMI does not measure health. At the population level, there is a strong correlation between BMI and health outcomes, justifying the use of this rather crude measure in epidemiology. But heavy does not always mean unhealthy, any more than skinny means vital. The adverse health effects of junk food extend far beyond the BMI, and by not addressing that- the study reminds us of it.

So, only in a world where tobacco is not to blame for emphysema, and shipwrecks in frigid waters not to blame for drowning, is junk food unimplicated in obesity. Junk food, per se, does not make us fat; as the report suggests, overeating makes us fat. But junk food is engineered to make us overeat, and while that doesn’t always work- it clearly works often. And even when junk food isn’t making us fat, it can make us unhealthy in many other ways that are harder to see- until it’s too late.

The truth, then, is simple: we, the people, would be much healthier, and many of us leaner, if junk were not a food group.

This, in turn, is part of a larger, and comparably simple truth: the fundamentals of healthy eating, and living, are clear, evidence-based, and consensus based. But since this is obscured by a confederacy of fools and fanatics; by an analysis of parts that only serves to hide the whole; by seemingly erudite answers to misguided questions; by hyperbole and titillation; and by predatory profiteers- there is a need to shine a light on it. The truth here is simple, but alarmingly hard to see through a dense haze of business-as-usual.

As noted, two timely efforts are directing their high beams at these very shadows. The first is a remarkable conference, sponsored by Oldways, and taking place in Boston just over a week from now, devoted to mapping out the ground common to diverse experts in nutrition. I for one am highly confident there is room there for all of us, and I look forward to standing there with colleagues and singing a rousing chorus of Kumbaya.

The second, a campaign rather than a conference, is the just-launched True Health Initiative. This effort, predicated on a genuinely intimidating mountain of evidence, and an unprecedented coalition of influential voices spanning the globe, is all about turning what experts know about healthy living into what everyone knows, and actually uses. Check us out, and if persuaded, please join us. For whatever it’s worth, my wife already has.

So there you have it: eat well, don’t smoke, beware of haze, and look out for icebergs.