Cover Image
close this bookNutrition Guidelines (MSF, 1995, 191 p.)
close this folderPart I: Nutrition Strategies in Emergency Situations
View the document1. Food crises
View the document2. Assessment of the nutritional situation
View the document3. Interventions: ensuring adequate general food availability and accessibility
View the document4. Interventions: selective feeding programmes
View the document5. Evaluation

2. Assessment of the nutritional situation

What do we need to know?

Information collected in order to take a decision, and to implement, alter, or stop programmes must be as clear and precise as possible.

· General information:

- Identify the origin of the problem (harvest failure, increased prices, population movements).
- Identify the population that is effected (number, ethnic groups, displaced, villages, camps).
- Identify other factors that may alter needs for intervention (other organizations, timing of harvests, national strategies).
- Realise the logistic constraints.

· Basic health information to be gathered must include (see Ref. 12, Annex 1):

- Mortality rates (crude and under five mortality rates);
- Major infectious diseases (measles, diarrhoea);
- Nutritional status of the people;
- Water availability (number of litres/person/day);
- Number of persons per latrine;
- Shelter;
- Amount of food available (Kcal/person/day).

Where do we find the information? (see Ref. 10 and 13)

· Capital Level Visit

- To collect all available information on relevant health, population and food/agriculture issues.
- To contact other organizations and find out their knowledge and plans to work in the affected area.
- To establish a realistic understanding with the authorities.

· Visits to the Field

- Talk to representatives of the population, or those with special knowledge (chiefs, health workers, agricultural extension agents, spiritual leaders, etc.).

- Talk to the affected people, in order to assess their ability to cope with the situation and their prioritization of needs.

- Observation (geographical area, water resources, shelter, adequacy of the food system).

The initial evaluation is extremely important and needs to be global, brief, concise and fast in order to allow an appropriate intervention as quickly as possible (see Annex 1). However, if a high level of precision is required for decision making, it is necessary to use a structured survey methodology.

The feasibility and usefulness of the information will depend on if the situation is <<simple>> or <<complex>>. A <<simple>> situation (like a well established camp) is where the population is:

- almost totally dependent on external food aid,
- population figures are well known,
- the population is easily accessible.

A more "complex" situation (like an open area) has many different food sources and the population is very spread out and inaccessible.

Measuring Malnutrition

There are 3 major clinical forms of severe protein energy malnutrition - marasmus, kwashiorkor and marasmic kwashiorkor. There are various clinical signs useful for diagnosis, but most obviously a marasmic child is extremely emaciated and a child with kwashiorkor has bilateral oedema (see Ref. 1 and 3). However, clinical assessment is not practical for managing nutritional programmes and monitoring and comparing large scale food crises.

Most standardized indicators of malnutrition in children are based on measurements of the body to see if growth has been adequate (anthropometry; see Ref. 14, 15, 18, 20).

· Height for age (H/A), is an indicator of chronic malnutrition. A child exposed to inadequate nutrition for a long period of time will have a reduced growth - and therefore a lower height- compared to other children of the same age (stunting).

· Weight for age (W/A), is a composite indicator of both long-term malnutrition (deficit in height/"stunting") and current malnutrition (deficit in weight/ "wasting").

· Weight for height (W/H), is an indicator of acute malnutrition that tells us if a child is too thin for a given height (wasting).

For all 3 indicators (W/H, W/A, H/A), we compare individual measurements to international reference values for a healthy population (NCHS/WHO/CDC reference values).(Annex 13).

In emergencies, W/H is the best indicator as:

- it reflects the present situation;
- it is sensitive to rapid changes (problems and recovery);
- it is a good predictor of immediate mortality risk;
- it can be used to monitor the evolution of the nutritional status of the population.

· Bilateral oedema is an indicator of Kwashiorkor. All children with oedema are regarded as being severely acutely malnourished, irrespective of their W/H.

Therefore, it is essential to assess W/H and the presence of bilateral oedema to define acute malnutrition.

· Middle upper arm circumference (MUAC), is another anthropometric indicator. MUAC is simple, fast and is a good predictor of immediate risk of death, and can be used to measure acute malnutrition from 6 months to 59 months (although it overestimates rates in the 6-12 month age groups).

However, the risk of measurement error is very high, therefore MUAC is only used for quick screening and rapid assessments of the nutritional situation of the population to determine the need for a proper W/H random survey (see Annex 2).

(For the assessment of adult and adolescent malnutrition, see Annex 3.)

Measuring the Nutritional Status of a Population

Anthropometric surveys allow us to quantify the severity of the nutritional situation at one point in time, which is essential to help plan and initiate an appropriate response.

The prevalence of malnutrition in the 6 - 59 month age group is used as an indicator for nutritional status of the entire population, because:

- this sub-group is more sensitive to nutritional stress,
- interventions are usually targeted to this group.

In order to ensure that the estimate will be representative of the whole population, random, systematic or cluster sampling procedures must be used (see Part II).

During the survey, the nutritional status of individual children is assessed, prevalence of malnutrition is then expressed as the percentage of children moderately and severely acutely malnourished. It is very important to mention:

- the indicator (W/H, OEDEMA, MUAC),
- the method of statistical description (% of the Median, Z-Score),
- the cut-off points used.

Results should always be expressed as the percentage of children < -2 Z- Scores and < -3 Z Scores and/or oedema, to allow international comparisons as well as for statistical reasons.

However, it might also be necessary to express the results using a different classification system, if that is the method generally used in the area that you are working in.

The cut-off points most often used to define acute malnutrition for the different indicators during nutritional emergencies are:



Ideally, an anthropometric survey should be part of the initial assessment in every emergency situation. The malnutrition rate can be used:

- to establish the degree of emergency for the delivery of food aid;
- to plan complementary interventions;
- as baseline information to monitor the progression of the situation over time.

Survey information might be useful under certain circumstances. For example:

· Camp formation is usually sufficient evidence of problems with food supply, and an anthropometric survey is an essential part of the initial needs assessment.

· Health information systems or famine early warning systems indicate a deteriorating nutritional situation.

· Health organizations often have a good field presence and close contact with the population. When secondary information or field experience (from contact with health workers, local chiefs, extension workers, other NGO/government workers) indicates a major nutritional problem, an anthropometric survey should also be considered. Nevertheless, conducting a survey is expensive and consumes time and energy. You must consider the following aspects before actually starting a survey:

1) Will results of the survey be crucial to decision making?

If the needs are obvious, a survey may not be needed to mobilize full scale action and will only waste time.

Secondly, one should be prepared to act after identifying a problem, directly or indirectly. If this is not possible and the information will not affect anyone's decision, do not do a survey.

2) Is a survey feasible?

Risks from insecurity, and logistic and team capacity should be considered.

3) Is an anthropometric survey the only option?

In stable environments, it may be better to establish a system of data collection that helps track trends in the nutritional status of a population over time. Only when there are dramatic changes (i.e. deterioration of the nutrition situation, influx of people, natural disaster, epidemics) might there be a need for further anthropometric surveys to determine the absolute levels of acute malnutrition.

4) Are you able to get access to all of your population of interest?

It is essential to make a clear definition of the population of interest (political/ administrative boundaries, geographical regions, etc.), from which to draw a representative sample. If all sections of the population of interest are not reachable, * may not be worth doing a survey.


There are no hard and fast rules about the required frequency of anthropometric surveys.

In a simple situation:

In the emergency phase, surveys should be repeated quite often (i.e. every 3 months) as food supply systems are often weak, there may be considerable influxes of people, a greater risk of epidemics and elevated mortality rates.

Once the situation has stabilized and mortality rates have declined, the frequency of anthropometric surveys can be decreased.

In complex situations:

Organizing a survey in insecure regions or where the population is spread out is logistically more difficult and results are harder to interpret. A good anthropometric survey may provide critical information, but the frequency and regularity with which they are conducted should depend on priorities and the capacity of the team to implement the surveys.

A compromise may be to do an initial survey, then collect other data in order to monitor the situation (i.e. OPD data, hospital admissions, food availability and prices). Extreme caution is needed in interpreting this kind of non-representative data over time. If the groups measured are well known and the data shows consistent trends over time, then they may be a very useful source of information.

When trends indicate a rapidly worsening situation and there is a need for more precise estimation of the severity of the problem, then another survey should be conducted.

· An anthropometric survey should be part of the initial assessment of an emergency situation. Discuss beforehand how the results of the survey will influence decision making.
· When feasible and useful, anthropometric surveys should be repeated regularly.
· In more volatile situations, repetition of surveys may not be possible on a regular and frequent basis. An initial survey can be complemented with the collection of other data to monitor trends over time - a deterioration in the situation can prompt another survey.


There are crucial survey figures, that will be used for decision making and reporting:

· global acute malnutrition rate
· severe acute malnutrition rate

In addition to describing and quantifying the severity of the situation, one must also have information on:

- Factors that might bias the estimate of severity.
- The distribution of malnutrition in the population.
- Context factors that will influence interpretation.

Factors that might bias the estimation of rates of malnutrition

· Excess mortality of the most vulnerable might result in an underestimation of the true malnutrition problem (see Annex 9: <<gaps>> in the sample's age/height groups may show an under-representation of an age group or a sex group which might indicate elevated mortality).

· Timing or seasonality might make comparison of results from different periods hard to interpret.

· When malnutrition is mainly a problem in age groups other than the under 5 years (rare), survey results might underestimate the problem.

· Migration or absence of the worst effected families will tend to reduce the significance of malnutrition rates.

· Inadequate population data or access to certain segments of the population may mean that certain groups are left out of the estimation of malnutrition rates.

Distribution of Malnutrition within the population

The identification of population groups most affected can help target programmes more effectively. Sub-analyses of the anthropometric data may help suggest target groups:

· Malnutrition rates per age/sex group (Annex 9).
· Malnutrition rates by population group (displaced/residents, ethnic groups, etc.).
· Malnutrition rates per geographic area.

Contextual factors in the interpretation of anthropometric surveys

The interpretation of anthropometric surveys must take into account the context.

a) Interpretation in simple situations

Interpretation of anthropometric survey results should take into account 3 main aggravating factors:

· Mortality figures (see Annex 1 for <<acceptable>> mortality figures);
· General food rations (Chapter 3);
· Major epidemic outbreaks (measles, Shigella and other diarrhoeal diseases,...).

b) Interpretation in complex situations

In complex situations, both food availability and accessibility can be at the base of the nutritional problem.

Figure 3 outlines various factors that can influence food availability and/or accessibility and therefore may have an impact on nutritional status.

In complex situations, an anthropometric survey may not provide sufficient information with which to make informed decisions. In these conditions, other information (existing, information collected by rapid assessments or additional information collected during the anthropometric survey) will be needed to characterize the situation.


The survey data may therefore need to be complemented with the following additional information:

* Market prices of important cereals and livestock
* Information on access and availability of food
* Price and availability of seeds and tools
* Rainfall, pests and other agricultural information
* Major events with respect to mortality and morbidity
* Specific nutritional deficiencies (e.g. scurvy, pellagra,...)
* Food intoxication cases (rare)
*Information from health centres or clinics.

The interpretation of these data types will depend on an understanding of the normal circumstances, and what a change in the indicator actually means.

Often an important factor to consider is seasonality. The timing of harvests and hunger periods explains much of the yearly variation in the nutritional status of some populations. Some communities live constantly on the edge, due to poverty and seasonal food shortage.


· Context

In simple situations it is possible to use survey results to make straight forward decisions and implement standard programmes. In more complex situations it is important not to encourage camp formation through our nutritional interventions. In complex situations, it is essential to adapt standard strategies to support local coping strategies and reflect social and demographic patterns in the population.

· Survey timeliness

As stressed earlier, nutrition surveys are only useful if they inform decision making. Surveys should be conducted so that the results are ready in time for important decisions that need to be made. Survey results should also be shared with other organizations, to help them make more informed decisions.

· Causes

The causes of the nutritional problem may greatly effect the choices to be made in designing a nutritional programme. You may have an idea of the causes, as this knowledge may have prompted the survey (i.e. inadequate food rations for an extended period of time, crop failure). However, it may be necessary to investigate causes by asking community leaders about the problems or by asking the mothers of the malnourished.

· Logistics survey

Survey results may imply a certain intervention strategy, but programmes must be designed to reflect the logistical realities and the team capacities.

When conducting a nutrition survey, the various teams will travel extensively through the area of interest, they will have opportunities to talk to representatives of many communities and observe some of the constraints faced by the population.

The survey team should take advantage of the field experience to collect information useful for interventions:

· Teams can be asked to map out their paths to villages and report on broken bridges, impassable rivers, etc.

· Teams can report on the spread of the population in various areas.

· Teams can report on the availability of water and cooking fuel for the population.

- Surveys are expensive and time consuming so one should take full advantage of the opportunity to collect relevant information when conducting a survey.

- One should never ask too many questions.

- However, key information can help make the analysis of the nutritional situation and associated factors much stronger.

SUCH INFORMATION MUST BE IDENTIFIED BEFORE THE SURVEY - it is impossible to go back and ask the questions later.

· It is useful to consider more than just the traditional rates of malnutrition for decision making.

· Additional information collected and special sub-analyses of the data will allow speculation about causes and identification of most effected groups - thus enabling better design and targeting of the programme.

· The experience gained while conducting the survey can provide important additional information for practical considerations in programme planning.

The feasibility, reliability and usefulness of an anthropometric survey depends on the resources of the organization and the level of complexity of the situation.

In "simple" situations where food systems are controlled, the population is well monitored, anthropometric surveys are relatively easy to plan, conduct and interpret - interpretation must take into account the adequacy of food rations, mortality and epidemics.

· In more "complex" situations, anthropometric surveys are difficult to perform adequately and results are often insufficient to characterize a situation and make decisions - other critical information must also be collected to help interpretation.