Assessing and defining food needs (Special Supplement 1)

Assessing food needs

There are three broad approaches in use for assessing the food needs of a population. These approaches are often used simultaneously in an assessment.

'Economic', i.e. judging need against people's current or predicted ability to access enough food. These assessments measure the process by which households become food insecure and malnutrition occurs. The techniques now most widely used are based on estimating household 'entitlement' (see, for example, Seaman et al, 2000) or by defining livelihood groupings (see, for example, Young et al, 2001). The causes of their food insecurity may be multiple, for example, environmental (e.g. drought, flood affected) and socio-political (e.g. displaced, refugee, conflict affected).

Economic assessments, which are based on the Household Economy Approach (HEA), are typically used to assess the ability of defined wealth groups, (e.g. 'very poor', 'poor', 'middle' and 'better off' groups) within a defined population, to acquire food and non-food goods, although they can also be applied to individual households. This method can be used to predict the probable food needs of a population and when those needs are likely to arise, or to estimate current food needs. Livelihood methods typically identify livelihood groupings and determine their respective food security needs, which may include food aid or other food security interventions.

'Outcome' measures, i.e. measures of current physical distress. Anthropometric assessment, e.g. weight-for-height of a population sample of children, is the most widely used, although adults are sometimes measured. Anthropometric assessments give an estimate of the current prevalence of malnutrition in the measured group, typically children aged 6-59 months. There are no techniques in use for estimating the nutritional status of adolescents. The interpretation of anthropometric assessment of adults and the elderly presents difficulties.

Anthropometry can be used with an economic assessment to confirm an economic prediction after the predicted event has occurred and only if no intervention has been carried out, i.e. if it is anticipated on 'economic' grounds that a population will be unable to access sufficient food, anthropometry can be used to confirm this. Figure 2 shows predicted food aid needs for the forthcoming agricultural year of poor households in different food economy zones in North Darfur in October 2001, and their relationship to anthropometric outcome. In the absence of a response to the October predictions, the food economy zone with the greatest deficit ('Goz') was also the zone that saw the highest increase in rates of malnutrition measured in January and, subsequently, in April 2002.

Anthropometry can also be used to estimate the prevalence of current distress. However:

• In locations with a high prevalence of malnutrition in noncrisis years, there can be difficulties in separating the effects of the crisis and the underlying rate of malnutrition, unless baseline data is available for the same season. Absence of baseline data can also be problematic where there is a low prevalence of malnutrition in a crisis year. Box 1 shows rates of malnutrition gathered by cross sectional surveys in two districts in Malawi during an acute food shortage period. The rates of malnutrition are higher than national baseline rates (though baseline rates for these districts and season were not available) but are not exceptionally high compared to other emergency situations. Interpreting the severity of the situation based on these data would be impossible without an understanding of the food security and health situation.
• Anthropometric assessment is often limited to children aged 6-59 months, and therefore excludes adults, adolescents and other groups who may be in need. Sometimes recommendations made in anthropometric survey reports are restricted to interventions targeted at young children. Yet, the survey is intended to inform an understanding of the situation of the whole population, as it is assumed that young children are the first to show signs of malnutrition in a population which is facing a nutritional crisis. In some situations, however, this is untrue. For example, in Southern Sudan the highest malnutrition rates were observed in adolescent boys, resulting in feeding programmes targeting the wrong age group (Salama and Collins, 1999).
• A high prevalence of malnutrition cannot be reliably used to infer that that there is significant household food shortage, e.g. malnutrition may be caused by disease.
• Acute malnutrition is usually a late indicator and, therefore, when the objective is to prevent malnutrition, economic or livelihood assessments may be more appropriate than anthropometric surveys.

'Physiologically vulnerable groups', i.e. groups who have specific physiological food needs including older people, pregnant and lactating women, and the sick. The consistent poor quality of food aid rations (see Section 6) and widespread gender inequalities that exist in many societies mean that even if household rations are deemed adequate, there may be population subgroups who are at risk of malnutrition. The actual level of food insecurity faced by these groups depends on the context, e.g. if a household has sufficient food there may be no reason to suppose that an older person has special additional food needs.

The term 'vulnerable' should not be used to mean the same thing as need, nor should it be regarded as an absolute status (i.e. female headed households are always vulnerable). Rather, it should be used to denote risk of malnutrition. The degree of risk should be assessed in each emergency, otherwise groups which are in particular need may be overlooked and others, whose needs are less profound, may be included (Darcy and Hoffman, 2003). Those who are vulnerable to malnutrition will vary from place to place and should be specifically assessed rather than assumed. In practice, food aid is, in many cases, targeted at small numbers of relatively arbitrarily defined 'vulnerable groups' on the basis that they are easily identifiable and can be reached relatively simply.

Defining food needs

A Dinka woman waits for the rain to stop so she can make a fire and prepare dinner from her food ration at Ajip Bahr-el ghazal

Donors, government and NGOs may define needs differently, even given the same information (Darcy and Hoffman, 2003). For example, in a place during non-crisis years, most poor rural households may obtain their food partly from their own production and partly by purchase. Crop failure, or some other shock, may lead to the situation where a household could acquire enough food, without external assistance, but only at the cost of selling livestock or other assets and/or giving up nonfood expenditure, on health care, education, clothing and other basic needs.

In such a case, an economic assessment might lead to two different definitions of need. For example:

1. The need to ensure that the household did not have to sell productive assets, and would still have the means of survival after the crisis had passed.
2. The need to meet minimum survival needs, assuming that the household would exchange all other assets.

People affected by crisis also have a view of need and of the best use of food assistance, if, in practice, this is less clearly articulated or not heard. In many situations, a large proportion of a population face no risk of starvation and are chiefly preoccupied with preserving their assets and securing their means of subsistence in the long term, e.g. minimising current consumption to avoid selling livestock or other productive assets. Some communities may perceive those in greatest need to be those who are socially vulnerable, e.g. the old, widows, orphans, etc, and may not agree with criteria based only on economic and physiological vulnerability (Sharp, 2001). Some communities may be opposed to targeting food aid. In parts of southern Sudan, obligations to redistribute food within the community determine that although hunger may be concealed for a time, it then becomes evident in a large proportion of the population. Generosity in sharing food is religiously sanctioned (Harragin and Chol, 1998) so that targeted households or individuals are subject to considerable social stigma if they do not share the food. Furthermore, "A welfare system is no good when no-one within the structure has any resources. They will thus claim that they are all vulnerable and, if offered a food distribution, will feel that it is fairest spread equally among all of them.. Distributing food liberally and to all takes away the stigma attached to being a recipient of charity" (Harragin and Chol, 1998).