Evaluation of the nutritional quality of food supplies: prediction of “desirable” or “safe” protein:calorie ratios

Abstract

A mathematical model, based upon the bivariate distribution, has been developed for the prediction of the risk of protein deficiency associated with particular ratios of protein:calories in human diets. Two sets of predictions are offered: a) the risk to the individual associated with the concentration of protein in his or her diet, and b) the population risk associated with specified average concentrations in diets. Empirically, the predictions are independent of age, sex, growth rate, and body size at least beyond the age of 2 years. Although in the derivation of the model an underlying condition was that energy requirements are met, it is suggested that the predictions are applicable under conditions of moderate, but not severe, underfeeding and in the presence of recurrent infections or other environmental stresses which exert proportional effects upon both energy and protein requirements. In the prediction of population risks, the important variables become the biological quality of the dietary protein and the variability of the protein:calorie ratio in self-selected diets. Data on these points are sparse. For diets having a protein utilization value of approximately 70 relative to that of egg or milk (as might be found in many of the developing countries), a 1% level of risk is associated with protein concentrations of 8 to 13% of calories, as the coefficient of variation of the concentration increases from 10 to 25%. Comparison of the predictions with average concentrations of protein observed in dietary surveys (coefficient of variation not known) suggests that a primary inadequacy of protein in regional diets is unlikely in most areas. However, the data suggest that the margin between observed concentrations and predicted desirable concentrations is not so large that the introduction of energy sources which fail to provide reasonable levels of protein can be viewed without concern.