Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Human nutritional requirements
The uses of energy and protein requirement estimates
Report a Workshop
A workshop to discuss uses (and mix-uses) of human energy and protein requirement estimates was held at the Massachusetts Institute of Technology (MIT), 27-29 May 1980, under the sponsorship of the World Hunger Programme of the United Nations University. The objective of the workshop was to identify and comment on issues arising from the application of requirement estimates such as those prepared by the 1973 FAO/WHO committee. (It was hoped that identification of issues might serve as background for the planning of proposed 1981 FAO/WHO/UNU committee to be charged with review and revision of current requirement estimates and suggestions for their uses. The mandate of the workshop did not extend to the proposal of specific revisions.)
Participants, drawn from the fields of nutrition, economics, demography, and anthropology, were selected to provide experience ranging from the generation of requirement estimates through national economic and food and nutrition planning to the examination of demand and need at the community level. The co-chairmen were Professors George Beaton of the University of Toronto and Lance Taylor of MIT.
This report is presented as a series of interrelated themes that developed during the meeting, each leading to the recognition of issues or alternate approaches in the use of nutrient requirement estimates. On several occasions members noted that many of the issues arising in the present workshop had been identified and discussed in the 1973 FAO/WHO report. A major recurring problem in this field is the failure of users of that report to read or comprehend its comments and recommendations. An overall recommendation of the present workshop would be that the FAO/WHO/UNU committee be as explicit as possible about its meaning, and that all users be encouraged, by every means possible, to read the text of the report as well as the summary tables. However, in the present workshop there were irreconciled divisions of opinion about the appropriate and inappropriate use of nutrient requirement estimates, and about the methods by which such estimates should be related to population data. Clearly, this remains a matter to be addressed further by the FAO/WHO/UNU committee.
THE PLANNING FRAMEWORK
National planners typically think in terms of improving "shortfalls" from desirable levels of food supplies. Traditionally, these were measured at the national or per capita level through comparisons of per capita supply (food disappearance data) and per capita need (demo graphically weighted requirement estimates). More recently, and particularly during the past decade, attention has been given to disparities or shortfalls in the distribution of supply relative to need, the identity and magnitude of sub-populations having intakes less than required, and the determinants of both supply and demand across and within these sub-populations.
In such distributional analyses, data have seldom been available describing the actual distribution of usual food and/or nutrient intakes among individuals. Rather, information about actual or probable distribution of household income, together with estimates of income elasticity of demand, have been used in generating a proxy for the food purchased for human use. The amounts actually ingested have remained unknown.
The requirement estimates published by the FAO/WHO or national committees have often been used without critical consideration of their appropriate interpretation and application. Conversely, among those who have given detailed consideration to such issues as the variability of individual requirements, there are major disagreements as to the degree of correlation to be expected between intake and requirement, and the appropriate statistical approaches. As will be discussed later, there has been a growing realization of the significance of individual and community adaptations to various levels of energy intake and of a need to consider requirements for expected levels of performance.
These uncertainties, ambiguities, and controversies have clouded, and to a degree discredited, the use of requirement estimates as a part of national planning. The members of the workshop emphasized the need for clarification and understanding, since estimates of human nutritional requirements can and should be a part of the input in national development planning.
ALTERNATIVE POLICY INDICATORS
The present workshop focuses on the use of nutrient requirement estimates in national planning. The members emphasized that requirement estimates are not the only, or necessarily the most important, nutritional input in the planning exercises (see box, below). More generally, other "nutrition impact statements" in both macro and micro-planning exercises (see box, below): More generally, other nutrition-related performance indicators should be taken into consideration. For example, food intake is a link between economic and social factors affecting the individual and his or her capabilities, activities, and general welfare. Therefore, it is appropriate to consider how information might be gathered about those aspects of human welfare that are most closely connected with food, and how this information might be used in planning. The question might be approached by asking what problems emerge when food intake is insufficient. The answer is most often expressed in terms of poor health, lack of social activities, loss of alertness, and so on. Accordingly, some useful indicators may be suggested, although which of these provides reliable information is a matter for further research and policy judgement. Health-related measures would include, for example, mortality and morbidity rates, and measurement of growth and body composition. Relevant aspects of the way of life might be leisure activity, the pattern of play of children, and the behaviour of individuals. In addition, the composition of the diet might prove to be a useful indicator - the kind of foods people eat.
NUTRITION IMPACT STATEMENTS
Members of the United Nations University workshop on the Uses of Energy and Protein Requirement Estimates felt that it would be extremely useful to devise approaches that, with minimal information about the nature of the nutritional situation and about the nature of distribution (effective demand) of food within the population, could still be the basis for introducing nutritional considerations into national planning. Beyond that point there is a relationship between the type of policy decision and the extent and precision of distributional information that is needed, i.e., intake versus demand and requirement versus need, as well as factors affecting each.
There are at least two approaches to inclusion of nutritional considerations in planning, whether this be concern for food, health, or other aspects of development planning. The approach that has been promoted by the nutrition community as most desirable is that in which nutritional objectives are defined, and planning is undertaken to meet these objectives. A variant that is potentially more consistent with existing planning processes, and hence more acceptable to those involved in planning, might be the consideration of the potential or probable nutritional impact, positive or negative, of a proposed policy decision. That is, nutritional considerations might follow rather than precede the primary policy development process, but precede implementation. The extent and depth of analysis in the preparation of such a "nutrition impact statement" might vary with the data base and infrastructure available to generate the statement. This does not preclude consideration of nutritional implications; it merely sets limits to the possible extent and precision of the consideration.
The actual form and content of a nutritional impact statement would, of course, vary from case to case. Nonetheless, in implementation a number of questions repeatedly arise. They include:
- What variables, whether or not controlled by policy makers, actually affect distribution of nutrient intakes and requirements of a population? What are the functional relationships involved?
- Within the limits of existing knowledge, which proposed or presumed indicators of "nutritional status" of a population can be accepted as having some meaning, and which not? Should these indicators focus on physical health or extend to areas of social and functional performance?
- What are the major gaps in knowledge of these issues that
further research might help to clarify?
A realistic impact statement would, at a minimum, deal with these
issues.
For an impact statement to have any real effect, a strong political commitment is required to make nutritional considerations a part of food planning and of national development planning. In turn, a prerequisite to the generation of this commitment may be an awareness of the probable extent of nutritional problems in the population such as might arise from a general assessment of the situation, and of the probable implications of this situation for future development of the population and country.
Even though approaches to such assessments remain imperfect and even controversial, it is the responsibility of the concerned scientific community to provide guidance as to the most suitable approaches to be used in the interim while better approaches are being developed. It also behoves that same community to make clear the limitations to interpretation of present approaches. This needs to be done in a constructive rather than destructive manner. Where confusion and disagreement abound, the probable reaction is no action.
Patterns of consumption do change with income. For example, it has been observed across and within populations that energy derived from fats as a percentage of total energy rises with income, and that, simultaneously, the percentage of energy from carbohydrate declines. Total protein remains relatively constant, but the proportion of animal protein increases. Moreover, there may be an increase in demand for "empty" calories, particularly fats and sugars, as income rises. These typical changes in diet may provide an aggregate background against which impacts of food policy may be judged.
REQUIREMENT ESTIMATES
1. Energy Requirements
The 1973 FAO/WHO report provided estimates of average energy requirements for individuals of various ages and, for adults, of various categories of physical activity. Because of difficulty in defining physical activity in populations, the "reference man" and "reference woman" requirement estimates, adjusted to observed body weights for the population in question, have tended to be used as absolute values applicable to ail populations. Since 1973 there has been a growing consensus among nutritionists that the use of such absolutes is inappropriate and has led to misunderstanding. The basic reason is that individuals and populations can adapt to broad ranges of energy intakes. If body weight and composition are maintained, the adaptations include mainly the level and efficiency of physical activities, but also changes in basal metabolic rate (BMR), efficiency of energy utilization, and patterns of activity. These mechanisms play roles whose magnitude in energy conservation and expenditure must be considered in relation to the habitual physical activity of the populations and their intake level at energy equilibrium.
The extent to which socially or economically productive activity will be constrained by restriction of food intake is difficult, if not impossible, to predict from a knowledge of food intake alone. Various societies demand different levels of activity and different energy expenditures. For example, industrialized societies, through mechanization, may greatly decrease the per capita needs of their population relative to non-industrialized societies, where a higher level of physical activity may be required. Conversely, social structures of societies may become adapted to chronically low levels of dietary energy intake such that a rapid increase in food availability and consumption might be associated with positive energy balance and a tendency towards obesity rather than an increase in social or economic activity. Similarly, the partitioning of activity between "work" and "non-work" is a function of societal incentives, opportunities, and pressures. Within societies, the roles of different sub-groups may be adapted to the differing situations of the particular sub-groups. As a result, the variation in energy needs either to maintain the status quo or to move to a new level of equilibrium may be very large among these sub-groups.
In the case of children, adaptations to constrained energy intake also occur. It has been shown that growth may be maintained, perhaps at modestly reduced rates, by an adaptive reduction in physical activity. There is concern that this may result in reduced stimulation to their psychomotor development from interaction with their environment, and have lasting effects on learning behaviour.
For the above reasons, a UNU Working Group on Protein-Energy Requirements under Conditions Prevailing in Developing Countries, meeting at the Massachusetts Institute of Technology from 19 to 23 May 198û, whose report is in press, cautioned against attempts to define a single set of energy requirements applicable to all societies. The present workshop adds that current anthropometric indices of nutritional status consider energy balance only in terms of physiologic deficit and do not measure or take into account the social or economic significance of physical activity. There Is need for conceptual and practical clarification of both requirement estimates and definition of "nutritional status." Without a clearer understanding of the broader meaning of these terms, it will be very difficult to reach meaningful decisions about the numerical requirement estimates, and it will be even more difficult for planners and others to make use of the work of expert committees offering estimates of human energy requirements.
2. Protein Requirements
For protein requirements, the situation differs from that of energy. There is no recognized mechanism of adapting to intake below conventional estimates of requirements that does not involve some degree of change in body tissue composition, and hence may be presumed to mark the beginning of a deficiency state. Such changes are generally assumed to have adverse biological effects or to increase the potential for such adverse effects. The social adaptations that influence energy requirement have no effect, upward or downward, on protein requirement. In children, however, adaptations that involve differences in growth rates would be expected to affect protein as well as energy requirements.
Again unlike the case with energy, there is little or no risk to the individual associated with protein intake in excess of actual need. It has been customary to identify and recommend intakes of protein that are sufficiently high to cover the needs of nearly all individuals in the population category. Operatively, this has been defined as the estimated average requirement plus two standard deviations. Thus, the way in which protein requirements are expressed is in direct contrast to the approach taken for energy, for which average requirements have been used. The assumption in the case of energy is a strong correlation between intake and requirement among individuals having unlimited access to food, presumably mediated by a physiological regulation of intake as well as the previously discussed adaptations of requirement to match intake. Since there appears to be little or no correlation between intake and requirement of protein among individuals, there is no other way that national planners could assume that population needs are being met.
It was noted with concern that a fundamental misunder-standing of the differences in meaning between published estimates of energy and protein requirements persists. The workshop members pointed to the lack of consensus among nutritionists regarding the appropriate functional specification of the protein requirement distribution across individuals in a population. Clarification of these matters at the 1981 FAO/WHO/UNU meeting will be essential.
One topic that was singled out as requiring consideration for both energy and protein was specification of nutritional requirements associated with post-illness weight recovery, and more generally the concurrent change in requirements associated with disease states. Although knowledge about food requirements imposed by these conditions may be important at the individual or household level it is probably of limited importance in assessing the adequacy of total energy or protein supply for a population. The higher requirements imposed by periods of post-illness weight recovery are usually preceded by periods of reduced food intake in the same individual such that the overall need for that individual over the total time course is moderated.
Conversely, at the level of the particular individual, the availability of additional food during the post-illness weight recovery period is essential. The relatively greater require" meets for protein over energy for growth as a percentage of total requirement indicates that the minimal ratio of protein to energy also increases in this phase. The energy density of the foods may be important also. Therefore, both quality and quantity of the food consumed may take on particular importance in this context. The importance of energy density of diets has been identified by many groups. The FAO/WHO committees on Fats and Oils in Human Nutrition and on Carbohydrates in Human Nutrition have discussed this matter. These are matters that warrant special attention by the 1981 FAO/WHO/UNU committee.
3. Indices of Diet Quality
In the present workshop, emphasis has been placed upon the use of requirement estimates in the assessment of the probable adequacy of food intakes and food supplies. This is an appropriate first consideration. However, there are major dangers in restricting the focus of interest to energy. Food must also provide the required nutrients - protein amino acids, fat and fatty acids, vitamins, and minerals. Further, the food available to the individual must be of a type that is both acceptable within the setting of that culture and socio-economic group, and feasible of ingestion in the sense that the bulk is not too large as determined by the energy and nutrient density. From these considerations, two physiological criteria and one sociological criterion of food quality emerge:
a. minimally suitable concentration of nutrients per 100 kcal in the mixed diet as consumed;
b. minimally suitable density of the mixed diet as consumed (energy content per volume occupied in the digestive tract);
c. acceptability of individual foods (both for food preference and cost) for inclusion in the diet.
Both energy and nutrient density will have to be considered in connection with different levels of energy requirements as encompassed in the concept of recommended energy allowance (REA), discussed below.
In the third aspect of food quality - the attributes of foods that influence their acceptability and use - knowledge is fragmentary. Yet clearly, this is fundamental to any approach to the inclusion of nutritional considerations in national planning. In practice, such measures as apparent price and income elasticities of demand for existing commodities are relied upon, and may succeed in influencing consumption toward nutritionally desirable goals by affecting both income and food prices. There is reason to believe that such responses are culture-specific. In the case of "new" foods, there is dependence on pilot studies to test acceptability and later to assess demand. This is not a fully satisfactory approach, since there are many examples of failures in the attempted introduction of foods that are new to a particular culture. There is need for systematic investigation of the determinants of food behaviour. Such research might lead to prediction models and hence a greater flexibility and reliability in planning food availability at the community and family level.
It is worth noting that another type of research that should be considered would address the question: "What are the real constraints to food intake of young children in the community?" With this type of understanding, we might be in a much better position to predict whether quality factors, behavioural factors, supply factors, anorexia, or even satiation of actual need, limit observed intake, and then to consider implications for action. Such research has been proposed by WHO.
A PROPOSED APPROACH TO THE DESCRIPTION OF ENERGY REQUIREMENTS
The workshop discussions focused on concerns particularly relevant to planning rather than to the problems of defining absolute protein and energy requirements. The latter are to be subject to detailed review and possible revision by the 1981 FAO/WHO/UNU committee.
In considering energy "requirements," it was recognized that at least two different meanings of the word appear in common usage. For an individual, one meaning refers to the amount of energy needed for the maintenance of health at a sedentary level of activity (plus the additional needs for growth in a child); it does not take into account the energy expenditures for activity above this relatively sedentary level. The second meaning is normative - the amount of energy required to maintain zero energy balance (intake = expenditure) at some specified level of activity and body weight in a given social setting. Both meanings have relevance. Lack of clarity about which meaning is intended causes confusion, as in the assessment of the situation in a population of individuals. The workshop suggested that a new differential terminology be developed. A tentative suggestion, adopted for purposes of the present report, was that the latter meaning be associated with the term recommended energy allowance (REA) to denote a desirable level of intake and of related activity, while the former meaning be associated with a terminology such as subsistence energy need (SEN) to denote the base-line needs for the maintenance of physical health. It is to be noted that the workshop was more concerned with the differentiation of concepts than with the particular terminology.
In a given social context, one could conceive that a series of REAs (or a range about some central value) might be established by national or regional authorities, reflecting various possible social roles and their patterns of physical activities. These might then be seen as a part of development planning goals. The REAs might be used as reference points in assessing the nutritional impacts of policy changes or of changes in the pattern of life.
In contrast to these "targets," the estimates of SEN might be used to describe requirements in the sense of minimal need or, as discussed earlier, the lower physiological limit to adaptive changes. Together, and applied to population groups, SEN and REA might serve to describe the range from population survival to a desired level of population function, i.e., social and economic productivity as influenced by energy intake. It was recognized that in the case of populations failing between these two levels of intake, it would be very difficult to make statements about the probable distribution of either health problems due to a failure to ingest sufficient energy to meet existing energy requirements, and hence to maintain body composition and growth, or of functionality, linked to a failure to ingest sufficient energy to permit desired levels of social or economic productivity. As discussed earlier, other indices of these distributions might be needed. This realization should not detract from the use of the SEN and REA levels in planning and assessing progress towards a goal, or even the description of gaps to be filled in achieving this goal. It does imply the need for care in describing and interpreting the results of such examinations.
The workshop proposed tentative definitions of the concepts associated with the two terms:
Subsistence energy need (SEN) is the energy intake that is considered adequate to meet the minimum needs compatible with health for persons of the specified age, sex, and body size category. The SEN represents an average value for a group or population of such individuals. These minimum needs depend upon three major variables interrelated in a complex way: (a) age and sex, (b) body size and composition, and (c) climate and other micro-ecological factors. The SEN also includes the energy needed to maintain sedentary activity and should allow for an adequate rate of growth based on currently accepted standards, and desirable levels of mental and biological functions, such as learning, immune mechanisms, pregnancy, and lactation. It would be expected that groups with mean intakes below the SEN would exhibit signs of reduced body weight, altered body composition, reduced growth rate, or impairment of physiological functions.
Recommended energy allowance (REA) is the sum of the SEN plus the extra energy considered necessary for (a) activity patterns greater than sedentary, or (b) replenishing an individual or population deficit (a transient increase in needs until rehabilitation has been achieved). The REA is seen as ensuring an intake adequate to meet the energy demands imposed by the individual's expected way of life as may be perceived in that society. The REA should allow for desirable functioning in that society. To avoid favouring obesity, it should not be set at a level above the desirable and feasible functioning in that society. Unlike SEN, observed levels of intake below the REA do not necessarily imply a detrimental effect on health; they would imply a limitation to achievement of expected levels of function and, depending upon the nature of the distribution of intakes, might include a range of associated states, including actual ill-health in some and adequate performance in others.
These concepts are not greatly dissimilar from the expression of estimates of energy requirement associated with different levels of physical activity in the 1973 FAO/WHO report. However, the present emphasis is toward an explicit statement of a concept and meaning that has relevance to planning. The distinction is made between an approach that might predict energy requirements to maintain the status quo (which would usually be the existing energy intake), and one that might predict the energy needs to achieve and maintain a desired level of social and economic productivity, given the relief of constraints to such productivity that may exist within the structure of that society.
The REA levels might be expected to be influenced strongly by socio-economic and cultural conditions in the country or region concerned. For example, as well as body size, sex, age, etc. as determinants affecting the SEN, the REA should take into account the energy requirements for normative patterns of work and leisure-time activity. These will differ from one segment of the population to another, for example as between relatively sedentary city dwellers and a rural agrarian population. As the societal norms change such that energy expenditures would be expected to change, the REAs would change in parallel. It is to be recognized that risk is associated with either direction of departure from actual energy needs - risk of inadequacy or of excess. Therefore, it is recognized that in individual societies, or in particular segments of societies, the normative REAs might be above or below the observed present intakes. It follows also that there is need to examine societies and populations in terms of their component parts rather than as an aggregate whole.
With due recognition of inter-individual variability, and of probable relationships between intake and requirement, judgements can be made about the extent to which existing intakes approximate desirable intakes (REA levels). in this send, the REAs might be used as a natural ban line or reference point for assessing nutritional impact.
SOME DISTRIBUTIONAL CONSIDERATIONS AND PROBLEMS OF ANALYSIS
1. Distribution of Energy Intake
In an assessment of existing situations, or in the prediction of the effects of policy changes, the planning model should take into account the joint distribution of requirements (appropriately defined - see above) and of intakes. BY examining the probable or possible effects of planning variables on them distributions, the planner may obtain vital data concerning constraints to achievement of goals or a potential benefit comparison of policy alternatives. (See later discussion of the problems with such analyses).
One methodological problem in generating relevant intake data for this type of planning model lies in the concept of the consumption unit. Conventionally, requirements are described in terms of categories of individuals of particular physiological characteristics (age, #X, body size, etc.), while intake data are almost always collected at the level of the co-mensal unit - the household. Proxy distributions such as income distribution are also conventionally related to household units.
It is possible, of course, to calculate a household per capita requirement as a weighted mean of the requirements of individuals in the household, or to describe the household in terms of equivalent consumer units weighted in accord with the relative energy needs of the categories of individuals within the household. However, inequitable intra-household (intra-familial) distribution of food can lead to much greater deficits among some classes of individuals than would be expected from the household per capita figure. It is costly and difficult to obtain data about intra-family food intake distribution. Such evidence as exists from developing countries with limited or low food intakes does not support the general assumption that within households or families food is distributed in proportion to published estimates of energy requirements. It would be expected, though it has not been clearly established, that as total household intake rises, the intra-household distribution of intakes would more closely approximate the distribution of energy requirements. Thus, this problem may be more serious in attempts to predict the distribution of inadequacies within populations than in assessing progress towards defined goals.
Within the constraints of cost and feasibility, it may not be possible to obtain estimates of intake below the household level in the context of most countries. It may be possible to obtain other indices of probable food distribution as discussed previously. For example, if a survey is being conducted for other purposes, anthropometric or functional indicators of nutritional status may be used to infer information about intrafamilial distribution. Ultimately, the inference should be subjected to test.
In theory, it might be possible, through more systematic investigation in selected situations, to gain knowledge of the environmental variables that influence intra-household distribution and, indeed, about the influences on food intake in general. Care in interpretation and application would be needed, since it is recognized that factors affecting food selection operate in specific ecological contexts. The influence of particular factors may be very different in urban as against rural environments, in different climatic-geographic systems, and in different socio-political systems.
In addition to physiological regulators of total food intake, three broad classes of variables are generally recognized as affecting food selection and food intake. These include economic structure, social structure, and cultural factors. In the first category are such variables as the price of food, access to food production resources, cash income, and so on. The second set includes family composition (dependency ratio), and task allocation (domestic and non-domestic activities of household members), etc. The third set includes education cultural practices with respect to food, values affecting allocation of income and other resources between food and other wants, and so on.
The workshop members noted that, whether the planner wishes to develop an improved probability model of existing problems or to consider, by other approaches, the possible nutritional effects of policy alternatives, these determinants of food demand and the likelihood of different ecological contexts should be considered. More than maintaining the concurrent distribution of intake and requirement may be needed for effective planning.
Recognizing these ecological differences, it should be apparent that there is great advantage in examining sub-populations that are as internally homogeneous as feasible rather than addressing the population as a whole. Some suggested grouping criteria emerged from the discussions and include the following:
a. Rural and urban groupings.
b. Occupational categories of both men and women, including "while collar" groups, agricultural workers, and other specialized occupation groups
c. Regional ecological variants, such as highland versus lowland, hot zone versus temperate zone, and similar geographic variations. This may logically relate also to agricultural differences.
d. Different income or wealth levels within and among regions.
e. Household sizes and configurations that impose differential work loads, etc.
f. Ethnic, caste, and religious categories. In specific regions, special sub-groups often have important differences in dietary patterns, customary activity patterns, and other factors directly affecting nutrition levels (intake and requirement). In India, for example, caste membership imposes significant dietary patterns that operate independently from wealth and income. There are also religious prohibitions and proscriptions that often create stresses.
g. Various other "cleavage points" may be relevant in specific national and regional contexts, in relation to the geographic and ecologic variations in environmentally mediated behavioural requirements and in stress factors, especially parasitic and other infections.
h. Seasonal changes as well as other temporal variations should be considered in any assessment of nutritional patterns, deficits, and levels. Seasonal variations can have marked effects on family food availability as well as on levels of required physical activities. In addition, environmental stresses from ambient temperature and levels of infection have significant temporal patterns in many societies. Short- and long-term changes in family composition will also have effects on nutrition patterns in households.
2. Special Problems of Distributional Analyses
The foregoing discussion illustrates the very major problem of distributional analyses that are conducted for the population as a whole. The group strongly recommended that for such analyses, the population should be disaggregated as far as is feasible. That is, the analyses should be performed for sub-groups that are as homogeneous as possible, and the population should be described in terms of a collection of these separate analyses.
Such an approach does not circumvent the lack of information about intra-familial distribution of food. Although it can be recommended that family food consumption data be related to the composition of that family expressed as consumption units, or on some analogous basis, rather than assuming family composition to be equal to the average composition, this still does not take into account possible inequities in food distribution within families.
After substantial discussion, and with several members dissenting, it was accepted in principle that probability-based assessments of the fraction of a country's population of individuals having energy intakes less than their own REAs are valid. Nevertheless, there are formidable difficulties in making such calculations, even if the above precautions are observed. The members were divided in their judgement of the practical utility of such calculations. They were in-agreement on the need for considerable caution in the interpretations of the results of such assessments. The discussion focused upon the examination of concurrent distribution of intake and requirement for energy. The points made did not refer to the consideration of protein intake and requirement. A major difference exists between the two: in the case of energy, there is reason to assume some correlation between intake and requirement; in the case of protein, present evidence does not support such an assumption.
Some of the reasons underlying these concerns and assertions for energy are:
a. A nutrient requirement is an attribute of an individual.
b. Requirements vary within individuals over time in relation to a series of conditions that are only Partially understood.
c. Requirements vary from one individual to another even after adjustment for major characteristics (e.g., activity, sex, body composition, age) in a manner that is poorly understood.
d. The requirements of any particular individual under any particular circumstance can be determined only by a complex and expensive process.
e. The adequacy or inadequacy of the observed intake of an individual can be judged only on a probability basis. For such judgement, information is needed about the distribution of requirements among such individuals (i.e., individuals of that category of age, sex, body weight, level of activity, etc.) and about the expected correlation between intake and requirement among such individuals. It follows that a particular person cannot be classified as "malnourished" on the basis of information about food intake alone. For a group of similar individuals, the most that can be expected is a prediction of that proportion of individuals who are likely to be inadequately nourished, without identification of which individuals.
f. To move to a population level, other than by examination of homogeneous sub-groups and aggregation of the results, presents formidable conceptual difficulties and assumptions about distribution characteristics and correlations that are at least heroic, and in the eyes of some, unacceptable.
g. Conversely, without information on the distribution of intake in relation to probable requirement, one cannot conclude that a particular population is undernourished as a result of a deficit in dietary intake; i.e., that an increase in dietary intake would be likely to effect an improvement - a necessary, basic planning judgement.
h. For purposes of planning, for example, in deriving estimates of the required supply of foods measured as energy for a politico-geographic planning unit, the best estimate should be built up by aggregating information from the maximum available sets of population subgroups differentiated by characteristics that are likely to affect the distribution of requirement or intake.
i. The judgmental or normative nature of the REA concept should be emphasized. In this context, the label "undernourished" should be reserved for persons whose energy intake falls below their SEN.
Bearing them considerations in mind, it is clear that the most complete knowledge a planner could possibly have would be characteristics of a population-wide, joint probability distribution of intakes and REAs conditional upon all the specified variables in the population. In formal statistical terms, estimation of the undernourished would then require integration of this joint distribution across individuals in the population.
Thus, the probability approach is conceptually useful. It is also clear that the existing base of information about the proper functional form and parameterization of the joint intake/REA distribution is very limited. The group agreed that there is need for substantial research in this area.
RESEARCH NEEDS
In the text of this report, a number of specific research issues have been identified. However, rather than reiterating these individual questions, the group emphasized one overriding message from the workshop. If we are to improve the planning process, nutritional knowledge must be put at the service of planners and communities. To a greatly increased degree, we must move research from the laboratory, the metabolic ward, and the office to the actual community wherein the problems we hope to address are actually located.
Second, as exemplified within the structure of this workshop, there is an absolute need for a multi-disciplinary involvement in such research. Repeatedly, the group was confronted with problems that lie in the interface of biological and social sciences, and that require considerable statistical expertise to disentangle. They require examination from the perspective not only of the nutritionist and social scientist, but also from that of the planner. The fundamental research need, then, may be for the promotion of a new approach to research, recognizing that because it is unorthodox it may receive limited peer support in comparison to more conventional approaches.
As an example of the type of question that might be approached in a new way, it was suggested that there is need for case studies of the situation and processes existing within the communities. This might be expressed in terms of the need to disaggregate data, including those on estimates of the prevalence of deficiencies, in terms of multiple host and environmental variables. It would be useful, in a systematic manner, to obtain information on food intake versus REA and SEN distributions in areas where locally relevant parameters can be concurrently examined. In such studies, the relationship between food intake change and functional change, including both social and biological function, could be examined.
The effect of social change on both intake patterns and activity levels (REAs) should also be examined. The role of infection, in terms of possible biological effects, of associated anorexia, and of the need for catch-up growth/rehabilitation can be studied within the context of the actual community and its social forces. In such cam studies, it would also be desirable to examine the actual constraints to food intake in young children and the factors affecting intra-familial distribution of food, recognizing that it is at least possible that in some circumstances there is no real constraint to food intake other than lack of demand or need for food on the part of the individual.