Cover Image

PAPERBACK
$34.95



View/Hide Left Panel

when the distribution of requirement is skewed, the EAR cut-point method results in a biased estimate (in this case, an underestimate) of the true prevalence of inadequacy.

For which nutrients are the requirement distribution not symmetrical?

One nutrient for which it is known that requirements are not symmetrical about the EAR is iron in menstruating women. Because requirement data are so scarce, it is often difficult to investigate the shape of the distribution of requirements for every nutrient in every life stage and gender group. Indeed, there is virtually no information on the actual characteristics of any requirement distributions except perhaps protein in adult men and iron in adult women (FAO/WHO, 1988; FAO/WHO/UNU, 1985).

In the absence of additional information about the shape of the requirement distribution of a nutrient, it is implicitly assumed in this report (and the DRI nutrient reports) that the unknown distribution is symmetrical around the median requirement (the EAR).

When it is known that the distribution of requirements is skewed, the full probability approach can be used by computing a risk curve that reflects the skewed requirements. The FAO/WHO (1988) adopted a log normal distribution to model iron requirements in women and applied the probability approach under the log normal assumption.

The effect of skewness on the bias of the EAR cut-point method is likely to be significant. Even moderate amounts of skewness in the distribution of requirements may result in noticeable biases in prevalence estimates with the cut-point method. Therefore, when the distribution of requirements is known to be asymmetrical, as for iron in menstruating women, the probability approach, not the EAR cut-point method, is recommended for assessing the prevalence of nutrient inadequacy.

What Happens if the Variance of Requirement Is Greater Than the Variance of Intake?

At least in North America, the situation where variation in individual requirements is greater than variation in individual usual intakes is most likely to arise for institutionalized subpopulations—for example, prison inmates or residents of a long-term care facility —who are all fed similar diets. Figure 4-11 illustrates this scenario:



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement