Food for Thought
by Harold J. Morowitz - 1980
The sciences of nutrition and toxicology require deep methodological reexamination. (Source: ERIC)
Reprinted with permission from Harold J. Morowitz, "Food for Thought," Hospital Practice, vol. 11, no. 11, November 1976.
Nutrition, on first inspection, appears to be such a down-to-earth discipline that one has a moment of doubt about viewing it in relation to abstract subjects like philosophy of knowledge and theory of probability. But the issue is forced upon us. Someone out there is very concerned about what we eat and shouts at us over television, puts pictures on the cover of consumer magazines, clutters our mail, and sets a rash of titles in front of us at the bookstore. We are advised to avoid sucrose and saturated fatty acids. We are instructed both ways on vitamin C, and we are placed in unholy terror of the dread cholesterol.
I must confess that this "hate sterol" campaign is beginning to get to me. After all, we are all about one half of one percent cholesterol, and this self-negation seems a bit pathological. Cholesterol provides me with cell membrane material and keeps my organelle envelopes in an adequately fluid state. Besides, I dislike getting angry at molecules that occupy such an important position on the metabolic chart of animals; it seems like an assault on nature. Actually, when I begin to think about cholesterol as the precursor of both male and female sex hormones, I can get downright romantic about this substance. All of which serves to introduce my theme. How do I know that any or all of those people out there advising me on eating know what they are talking about? Once that awesome question is raised, we move into the philosophical domain of the theory of knowledge, epistemology.
Let us start by trying to state the basic task of nutritional science, which is to formulate a diet over the lifetime of an individual that will optimize health, well-being, and longevity. This calls for providing the necessary chemical components in the right proportion and avoiding or minimizing toxic substances. Such a protocol assumes: first, that we know what is essential; second, that the quantitative requirements are uniform enough to be meaningful for a population; third, that deleterious substances have been identified; and fourth, that the foods and toxins can be separated. This separation is often complicated by the observation that a nutrient at one level of intake can be a poison at a higher level, as in hypervitaminosis. The first two assumptions bring into focus the kinds of experiments that establish requirements or nutritional values of foods. These are carried out on rats or on humans and suffer from a number of methodological difficulties. The mapping from rodents to Homo sapiens is not precise. This is often presumed to be answered by picking strains of rats that resemble humans with respect to known requirements, a procedure that guarantees nothing with respect to unknown requirements. The difference in life-style between rats and humans simply cannot be neglected. The experiments themselves are long-term, costly, and often difficult to interpret. Discovery of subtle effects requires, for statistical validity, a far larger animal population than is usually practical.
A flagrant example of the problems of rat experimentation is found in the May 1976 Consumer Reports, where breads are "nutritionally compared." Rats in groups of six were tested for 16 weeks on a sole diet of one brand of bread per group. Without any discussion of primary data or statistical significance, the breads were rated for quality. Such ill conceived and inadequately reported experiments are supposed to guide the consumer.
Human experiments are invariably short-term with respect to life span and also suffer from the problem of too small a population. It is relatively easy to spot well-defined relations such as vitamin C and scurvy, but it is exceedingly difficult to evaluate less dramatic correlations. All of the problems met with in nutrient requirement experiments are encountered in toxicity studies. Acute toxic effects can be established, but the long-term nexus between various food additives and incidence of disease is extremely hard to come by.
A further methodological difficulty must be considered. Various nutrients and toxins have combinations of antagonistic and synergistic effects that render the response to single factors of limited usefulness. In the language of the systems theorist, we are dealing with a multiple input system, in which the signals are integrated in some manner that, at best, is incompletely understood. Having to consider interactions, even at the elementary level of two at a time, intensifies the problems, since for every N possibility there are N(N-l)/2 binary combinations to deal with. These difficulties are a result of the complexity of the system and cannot be avoided by simple devices of analysis.
Some additional information may be available from evolutionary considerations. If we study the foods that humans have eaten during the development of the species as determined by paleontological investigations, we have a long-term, if imprecise, set of data that might yield some useful clues. In addition, the diets of our primate relatives may contain further information. In some ways we know a good deal more about the nutritional status of domestic animals than about humans. The information is of immediate economic value in agriculture, and the organisms are more cooperative with the experimenters.
All of the preceding adds up to the fact that most of what we are told about nutrition is neither true nor false; it is indeterminate. The experiments that have been carried out are often inadequate to develop conclusions within the ground rules of probability and statistics and the accepted notions of scientific verification. The information available to diet planners consists of: a small body of universally accepted results such as the pathways of intermediate metabolism, a set of direct minimum requirements to avoid dietary deficiencies, data on toxic substances and levels of acute toxicity, and a very large body of resultsmany of which do not measure up to the minimum standards of statistical acceptability.
Given the Babel of information and misinformation, the question emerges: Is there a rational method to decide what to eat? To seek an answer, we must go back to a two-hundred-year-old idea that is usually referred to as "Laplace's Principle of Insufficient Reason" and states that events are to be assigned equal probabilities if there is no reason to think otherwise. E. T. Jaynes has recently extended the principle to state that "in making inferences on the basis of partial information, we must use that probability distribution that has the maximum (informational) entropy . . . maximally noncommittal with regard to missing information subject to what is known." Without going into a detailed computer program, we can infer that the most rational aliment will be obtained by fulfilling known nutrient requirements, minimizing intake of proven toxic substances, and randomizing everything else. In other words, the most variable diet that fulfills known constraints is the diet of choice. The clue is variation: use different foods, change brands frequently, and never use the same recipe twice. Bring on the escargot, candied grasshoppers, roast guinea pig, alfalfa sprouts, macadamia nuts, falafel, turnip greens, squid, mare's milk, and seaweed salad. Minimize the intake of packaged foods, since they contain additives of unknown
toxicity, and where they are unavoidable, keep varying the suppliers according to additives.
Beyond our personal diets, it is clear that the sciences of nutrition and toxicology require deep methodological reexamination. It has been my impression that these sciences have been largely under-emphasized in medical school curricula so that the individuals whose advice is most sought about foods, the practitioners, are inadequately trained and often unfamiliar with the full range of theoretical issues involved. The first step to remedy these problems consists of a probing examination of the problems of obtaining valid results within these fields of study. The second step would seem to be an upgrading of diet-related disciplines by medical faculties. Whatever else man may be, he is a mechanism that responds to a lifetime of inputs with a series of outputs. Those inputs that arrive through the gastrointestinal tract describe a large part of the human experience. At present they are of two categories: normal biochemicals that have been part of the human milieu for millenia and new organic compounds that have only recently been synthesized. Before we ingest any large quantity of these organics, it would be nice if we knew what we were doing. Nutrition could be a major field of medicineand a most positive onewhich aims at promoting good health rather than responding to system failures.