usage—and until recently the bulk of research on malnutrition—has been directed to understanding inadequate intakes of macro- and micronutrients. The measures of central concern are observed through analysis of biological tissues (e.g., serum), observation of well-established physical (e.g., anthropometric) and clinically observable consequences (e.g., blindness), and by inference from data on intake. For example, anthropometric status is commonly used to assess malnutrition of children under age 5 (de Onis, Blösner, Borghi, Frongillo, and Morris, 2004).

As malnutrition acquired a central role in scientific conceptualization, it was often mentioned jointly with the idea of hunger, to the point at which the two often became virtually synonymous. Nutritional scientists as well as social advocates therefore sought to describe the inequalities of access to adequate food and its consumption. One approach was to compare intakes of a nutrient for a given gender and life stage group with an established reference value, such as the Recommended Dietary Allowances (RDAs).

Some problems with using the RDA approach stem, in part, from its conceptual underpinnings. To cover the needs of nearly all of a group, the reference values were set at very high levels. Consequently, a proportion of the population may consume less than the RDAs but still have adequate nutrient intakes. Another problem is purely technical. It is difficult to use a single interview to assess usual nutrient intake in a biologically meaningful fashion. For instance, vitamin A intake varies considerably over time, and only the mean intake over a period of weeks is meaningful nutritionally, because vitamin A is stored and body reserves buffer the variability of intake. Further technical problems relate to the accuracy of reported intake and of the information used to translate food intake into nutrients. As a consequence of these problems, assessment of nutritional adequacy through interviews and analysis of the record in relation to the RDA is no longer considered appropriate (Institute of Medicine, 2000).

The United Nations Food and Agricultural Organization (FAO) took a different biologically based approach to define undernourishment as not ingesting enough food to meet energy needs. Operationally the FAO indicator is calculated from national food energy balance sheets. These balance sheets estimate the total energy available for human consumption nationally by adding total energy produced plus energy imported plus the change in stocks minus energy exported, energy wasted, and energy used for other than human consumption. FAO then creates a synthetic distribution of energy consumption for each country in which the mean is total energy available (from the balance sheets) and the variance is taken from another source, typically an estimate from a nationally representative household expenditure survey that accounts for energy exported and energy used for other than human consumption (Naiken, 2003). The resulting estimated distribution of undernourishment (i.e., food energy consumed) across countries is