In that rightfully famous routine by Abbott and Costello, What, if you recall, was on second (base). Who was on first. I Don't Know was on third. And so on, into the mayhem of historically comedic genius.
Of course, comedy was the only real point of the routine. But it may be worth noting that Bud Abbott, the perennial straight man, knew exactly what he was talking about. The names of all players in the field were hopelessly confusing, unless they weren't. Unless you knew exactly what you meant when you said them -- and he did. Lou Costello, ever the butt of all jokes, was utterly lost in the same terminology.
Two recently published, opposing papers about nutritional epidemiology suggest we have much this same problem when it comes to diet and health. Unlike baseball, however, in the case of nutrition, What, rather than Who, is on first. How replaces Who, and is on second. And what we know, and how it all plays out -- actually matters.
The "what" of nutritional epidemiology is what we measure. We measure, for instance, weight, height, body fat, and fat distribution with the BMI, waist circumference, and body composition analysis by various means. We measure total food consumption on metabolic wards, or with self-reported intake via recalls, diaries, or questionnaires, translating totals into calories. We measure the energy expenditure of physical activity, again on metabolic wards or with sophisticated VO2 max machines; with apps and trackers; or by asking people to tell us what they do. And we measure the pattern of diet, in controlled settings, or with those same food diaries, simple recall instruments, food frequency questionnaires, or with new and developing methods such as photographs of food and interactive, online systems.
All these same names for all these same players can yield clear understanding or utter confusion. So argue those two opposing papers in the recent literature.
The first, published in the International Journal of Obesity, and garnering considerable media attention, makes the case that self-reported intake of energy expended in exercise, and consumed in food, is so egregiously inaccurate as to be useless. The authors propose we consign such measures to the dustbin of history, and move on -- although to what is rather less clear. The team suggests the measures we require still need to be invented, or at the very least, refined.
The second paper, however, in Advances in Nutrition, makes rather the opposite case. The authors acknowledge the many limitations of how we measure dietary intake and energy expenditure. But whether or not prevailing measures are of use is much dependent on exactly what that use is. What, after all, is on first.
We don't, for instance, base the conclusion that energy imbalance prevails in modern America because we have reverential faith in self-reported data regarding calories and exercise. Rather, we base that conclusion on the observable fact that so many of us have gotten fat. Energy imbalance is the cause of a self-evident, robustly documented fact.
In this context, measures of energy intake and expenditure are put to various uses. They are used to characterize trends over time, in individuals and in populations. They are used to inform judgments about attribution. Rarely, though, is either of these measures the direct or primary basis for any policy or recommendation related to diet or exercise for control of weight or promotion of health.
The BMI, or body mass index, is a notoriously poor measure if being used to gauge the fatness of an individual. Muscle elevates BMI even more than fat does. And the BMI is blind to body fat distribution, which matters quite a bit to the health effects of obesity. A case could be made on this basis, and indeed has been, for abandoning the BMI altogether. But while the BMI serves poorly as a high-fidelity measure of body composition or health status in an individual, it serves adequately to track trends at the population level. If we had concomitant epidemics of obesity and totally awesome muscles, that might not be the case. But until that day dawns, it suffices. To date, a rise in average BMI reliably means we are getting fatter, not fitter, alas.
I am prompted to juxtapose the academic debate with a personal anecdote involving me and my elliptical machine. I acquired my old-model, LifeFitness elliptical from a warehouse store that resells retired gym and hotel exercise equipment roughly 15 years ago. I have pounded away on it almost every day for 40 minutes throughout those 15 years, and it just keeps getting the job done. I love it.
Among its now quaintly antiquated functions, my machine does purport to measure my energy expenditure in calories: calories burned per minute, and the total at the end. I can think of three ways I might use this information:
1) I might use the reported energy expenditure to determine the calories I should eat every day. I might then track both my calories in and my calories out, and ignore whatever happens to my weight. If my clothes stop fitting, my clothes must be shrinking.
2) I might, alternatively, start out using my machine's report to guide my daily intake of calories, but then adjust over time based on my actual weight.
3) Or, I might use the information my machine gives me, but for something totally unrelated to my "energy balance."
If I were to practice choice 1, and perhaps choice 2, the precautionary tale in the International Journal of Obesity would pertain. But in fact, I make choice 3.
I have long noted the high probability of profound inaccuracy in my machine's measure. I travel often, and use exercise equipment in gyms all over the world. They are ridiculously discrepant from mine, and from each other. So if my machine is accurate, most others are not -- and vice versa.
But I don't really care. I simply use the measure on my machine to gauge if I am working about as hard today as I did yesterday, and as I do on an average day. Once I establish the level of energy expenditure that corresponds to a good, hard workout, I can use that as the floor I should clear every day. If I clear it, I am putting in a good effort. If not, I am probably slacking off, however unintentionally.
This is exactly how I use the information my machine delivers. And based on how I feel each day, and on records I've kept over portions of the past 15 years, I do trust my machine to be consistent with itself, even if not accurate. However wrong it may be, it is wrong in much the same way every day. When I push harder, it tells me more calories are burned. When I'm not sure how hard I am working, it reliably tells me.
As for calories in, I don't count those at all. I certainly think, unlike some, that they count -- but I find that controlling energy intake follows far more comfortably from making calories count than from counting them. I just eat wholesome foods in sensible combinations, and the result is filling up on the right number of calories. The proof is that I weigh just about what I did when I graduated high school.
So, the only measures of energy balance I apply to myself are: Do my clothes fit? Can I do the same number of chin-ups and pull-ups as yesterday? I don't care how many calories I ate yesterday, and it doesn't matter if my elliptical erred high, low, or was spot on.
The paper in the International Journal of Obesity tells us that some of our measures in nutritional epidemiology are very imperfect. This is true.
The paper in Advances in Nutrition tells us that perfect measures don't exist, but that the right conclusions for policy and practice can be reached just the same when how we measure is appropriately contingent on what it will be used for.
Better measures would be better. But we have learned enough, despite the flawed measures we have, to establish across a stunning expanse of years, methods, and populations, the fundamental associations between dietary and activity patterns and health outcomes. We know enough to prevent roughly 80 percent of all chronic disease and related premature death. I humbly suggest that as we squander this luminous opportunity to use what we do know, the flaws in measures we use to learn what we don't are not our biggest problem.
So yes, there are limitations to How on second, that could, in theory, land you on third, with -- "I don't know." On the other hand, the real action is on first. From my point of view, despite the imperfect measures we have -- we really do know what's What.
David L. Katz, MD, MPH, FACPM, FACP played in the outfield, and frankly, not all that well.
Director, Yale University Prevention Research Center; Griffin Hospital
President, American College of Lifestyle Medicine
Editor-in-Chief, Childhood Obesity
Author: Disease Proof