Abstract
Introduction
Fascination with the idea that some diet is “best” is, literally, perennial. 1 Popular interest in a parade of food and ingredients as contenders for silver bullet or scapegoat is seemingly as persistent. 2,3
These 2 fixations of popular culture pose a singular challenge to nutritional epidemiology. For there to be a rebuttal, predicated on evidence, to marketing hype and misdirected hope, the relevant evidence must be generated.
The charge has been levied, repeatedly and at volume 4 that nutritional epidemiology is flawed in its methods for pursuing reliable evidence and fundamental truths. The result of such assaults, aided and abetted by the hospitality peer-reviewed journals routinely accord them, 5 is a state of persistent pseudo-confusion about diet that obviates the conversion of what we have long known 6 into what we do to good public health effect. 7
The charge of deficiencies in nutritional epidemiology is true at least in part: there is much yet to learn. That, of course, is true in every field of human inquiry. The proposed remedies, however-notably, the contended exclusive need for ever more randomized controlled trials, and the unfailing advantage in ever larger cohorts—are not merely flawed, but actually corrosive to understanding.
Randomized controlled trials do not pertain to every important question in nutritional epidemiology, including the most fundamental of all: what way of eating produces the greatest combination of vitality and longevity? 8,9 To generate this answer directly, a RCT would need to span the lifetime of its participants. Of even more concern, RCTs can all too readily be designed to deliver the “winner” favored at the start, and then sanctify that conclusion under the halo of “robust methods.”
While critics rightly note inconsistent findings about the isolated health effects of given foods or nutrients, this is often because the dietary context, or population, have varied as well. What a food does or doesn’t do owes much to the company it keeps, and that is all too easily manipulated.
To feed the popular fascination in competing diets, the “what” of dietary patterns is at times conflated with the “how.” Various means of moderating daily energy (calorie) intake, or fostering satiety, are fully compatible with diverse diets of far-ranging composition. These “how to” elements of eating well and/or salutary energy balance are often pitted against “what to” dietary patterns, in pursuit of a winner. The answer in all such cases is a non sequitur, because the question is altogether vapid.
The French expression “je ne sais quoi” has been adopted into the English vernacular to mean: “a certain something.” The literal translation, however, of this slightly grammatically handicapped French expression is: “I don’t know what.”
When the effects of food or ingredient A on health outcomes are being tested, but the addition of that food causes reductions in the intake of food B-there is more than one variable in motion. Attribution of outcomes to just one of them becomes impossible because we, literally, “don’t know what” the causal factor is-even if our methods wear the imprimatur of a RCT. Similarly, when Diet A and Diet B are tested head-to-head, but they vary in both composition and objectively measured quality-we again have more than one variable in motion, and the attribution of observed effects is precluded.
The sine qua non of interpretable epidemiologic research is to compare 2 groups alike in “all ways but one.” 10 If we want to understand what a given food does to biomarkers, other aspects of diet must be matched. If we want to compare diets of differing composition, eating plans must be matched at least approximately for objective measures of quality.
The commentaries that follow are in the service of that transition, from je ne sais quoi, to sine qua non. There is not just one research method to advance human understanding-of nutrition, or any field-but there are universal elements of methodology that pertain to all good research.
The authors assembled here acknowledge that many valid questions about the best dietary pattern for specific goals in specific populations remain unresolved. We contend, however, that the basic theme of sustainably optimal eating for Homo sapiens in general is well established on the basis of evidence as diverse in its methods as consistent in its conclusions. 11 We maintain that the substitution of “junk” (i.e., ultra-processed food) for real food is a reliably dubious proposition. 12
For such knowledge to translate into collective action and serve the public good, it must be recognized and respected as established knowledge. In nutrition, as in all other fields of inquiry, there is space enough for both known and unknown.
We make the case here for dietary epidemiology done right, for methods worthy of such respect. As for diet itself, it is the single leading behavioral predictor of health outcomes in the modern world 13 the case for its respect we hold, accordingly, to be self-evident.