Cover Image
close this book Measuring drought and drought impacts in Red Sea Province
close this folder 5. Nutritional status of children in Red Sea Province, November 1985 to November 1987. Mary Cole and Roy Cole
View the document Summary
View the document Introduction
View the document Methods
View the document Results
View the document Conclusions
View the document Discussion
View the document Future directions.
View the document Appendix 5.1. Data collection form, nutritional surveillance teams, Oxfam
View the document Appendix 5.2. Claasifications of coded variables.
View the document Appendix 5.3. Ecozones in Red Sea Province (from Watson, 1976).
View the document Appendix 5.4. Seasons by month and ecozone, Red Sea Province.
View the document Appendix 5.5. Classification of fled Sea Province into food security zones, 1987.
View the document Appendix 5.6. Locations of sampled sites, nutritional surveillance cycle 1.
View the document Appendix 5.7. Names of sampled sites, nutritional surveillance cycle 1.
View the document Appendix 5.8. Locations of sampled sites, nutritional surveillance cycle 2.
View the document Appendix 5.9. Names of sampled sites, nutritional surveillance cycle 2.
View the document Appendix 5.10. Locations of sampled sites. nutritional surveillance cycle 3.
View the document Appendix 5.11. Names of sampled sites, nutritional surveillance cycle 3.
View the document Appendix 5.12 Locations of sampled sites, nutritional surveillance cycle 4.
View the document Appendix 5.13. Names of sampled sites, nutritional surveillance cycle 4.
View the document Appendix 5.14 Locations of sampled sites, nutritional surveillance cycle 5.
View the document Appendix 5.15. Names of sampled sites, nutritional surveillance cycle 5.
View the document Appendix 5.16. Locations of sampled sites, nutritional surveillance cycle 6.
View the document Appendix 5.17. Names of sampled sites, nutritional surveillance cycle 6.


1. Gender differences in nutritional status.

From the figures for mean percentage reference median weight for height it would appear that females had a better nutritional status than males in Red Sea Province, although the difference was so small as to be of little practical significance. Equal proportions of males and females were malnourished, but in 1986 significantly more females than males were severely malnourished. In large samples relatively small but significant differences in mean percent reference median weight for height may be a reflection of differences in distributions above 80% weight for height. The mean percentage weight for height in two samples can be significantly different while the percentage of classified as malnourished is not, because of varying proportions in the different categories of well nourished children. Changes in mean percent weight for height can be used as an indicator of a worsening nutritional status in the community before increased malnutrition is seen. Mean percent weight for height also helps to describe the total distribution of percent weight for height in the sample. It is not safe to assume, however, that because one group has a higher mean percent weight for height than another (in this case females versus males), that the higher group has lower rates of malnutrition. From the results of this study, it would appear that females are at higher risk of malnutrition, particularly severe malnutrition, than males.

Significantly higher proportions of males than females were found in the samples from both cycles 1-3 and 4-6. We lack enough information to say whether these significant differences in proportions of males and females are a true reflection of a difference in the population, or are a result of a bias in the sample. If these figures are representative of the population, and assuming that the proportions of males and females are equal at birth, higher proportions of males than females in children 75.1-115 cm suggests a higher mortality rate in females than in males. This would be consistent with higher rates of severe malnutrition in females than in males, but there are no mortality data available to support or refute this hypothesis. Mortality data are important in interpreting change in nutritional status. An apparent improvement in nutritional status may be due to mortality in severely malnourished children. In Red Sea Province it is difficult to collect reliable mortality data as death, particularly in children, is a taboo subject. The health care infrastructure in the province is weak and the health information system extremely incomplete.

An alternative explanation of the unequal proportions of males and females is that the sample was not representative; that it was biased towards sampling males. There are no data on refusal rates by sex to arrive at this conclusion, but there remains the possibility that cultural factors may prevent women from bringing forward female children for measurement. One observation from the raw data: in some settlements the raw data is grouped into batches of males and females. This suggests that there may have been some division by gender during the process of assembling the children to be weighed. If this was the case, it would have added a complicating factor to the process of obtaining a truly random sample. If future surveys are undertaken, accurate records should be kept of refusals.

There is little information on infant and child feeding practices in Red Sea Province, and how they vary with the sex of the child. Colin Alfred undertook 10 weeks of field work in April - June 1986 in which he used open-ended questionnaires and interviews to collect information designed to increase understanding of the effect of food aid on communities. In discussion groups with women, a consensus emerged that "...boys ate more than girls, but not when they were young". The question of whether women felt that males needed more food than females was not addressed. Also, expressed opinions may not reflect practice, especially in times of stress.

2. Co-operation and "survey fatigue".

Although it is not within the scope of this report to comment extensively on survey design and practical difficulties in the field, there are some points that should be highlighted here. The percentage of places receiving repeat visits does reflect the possibility of different areas of a settlement being surveyed at each visit. In the case of larger settlements these areas could have been some distance from each other. Conversely, it is possible that in some areas one community leader would have to be consulted before a number of settlements could be visited. The number of repeat visits to individual community leaders could therefore be considerably higher than the number of repeat visits to individual settlements. The total number of visits by all Oxfam and non-Oxfam teams should be considered in assessing survey fatigue. Oxfam Food Monitors, World Food Programme Monitors, SERISS and Ministry of Health teams all toured Red Sea Province with varying frequency. The British Red Cross, Norwegian Red Cross, Euro-Action Acord, International League of the Red Cross and the Sudanese Red Crescent work in rural areas of Red Sea Province also.

The difficulties associated with survey co-operation increased not only with increasing numbers of visits, but with decreasing relief rations and decreased drought related stress. The unpopular and invasive nutritional surveillance was seen as a necessary evil associated with relief distribution. The nutritional surveillance teams themselves often emphasized the links between nutritional surveillance and the relief ration in an attempt to ensure co-operation. As discussed previously, refusal rates must be accurately recorded in the future. Refusal rates for an invasive and unpopular procedure are in themselves an indicator of stress in the community.

3. Changes in height distributions between cycles 1-3 and 4-6.

The increase in proportions of children less than 75 cm (approximately one year old) between cycles 1-3 and 4-6 was highly significant. Possible explanations for this change include:

a. Sampling bias.

b. An increased birth rate in cycles 4-6.

c. A decreased infant mortality rate in cycles 4-6.

Better maternal nutritional status would contribute to both an increased birth rate and decreased infant mortality. Again, the lack of vital statistics for the province and a lack of detailed data on refusal rates make it impossible to evaluate the change in the height distributions of the sample.

4. Nutritional status of children, total province.

Province-wide figures for mean percentage reference median weight for height and percentage of malnourished children provide a good starting point for describing nutritional status. The advantages of province-wide figures include a large sample size, which increases the probability of finding statistically significant results, and results that are not influenced by population shifts within the province (although Red Sea Province is far from a closed system, and population movement does occur across provincial boundaries). The disadvantage of province-wide results is that they mask variation within the province.

In Red Sea province as a whole mean percentage reference median weight for height increased, while the percentage of malnourished children decreased between cycles 1-3 and 46. The increase in mean percentage reference median weight for height was very small, and despite being statistically significant, is of little practical significance. The decrease in the percentage of malnourished children is encouraging, but we lack the historical data necessary to evaluate the degree of malnutrition in cycles 1-3 and 4-6 or the change between them in the context of what is usual variation in Red Sea Province.

There were no formal relief programme objectives in terms of nutritional status set at the beginning of the programme against which to assess the decrease in malnutrition.

Percentage reference median weight for height is a proxy indicator of nutritional status. The cut off of 80% weight for height for classifying malnutrition is largely arbitrary. We have no means of assessing what shifts in a proxy indicator around this arbitrary cut off really mean in terms of the health and welfare of children in Red Sea Province. More information is needed on the Beja perception of child health, acceptable and unacceptable levels of thinness and shortness in children of varying ages, and the expected variation in these both seasonally and in times of stress. By developing a locally appropriate cut off for malnutrition which could be used alongside international standards, future programme objectives could be both more realistic and aimed at a degree of change which has practical significance and visible impacts for the beneficiaries. Unless we have more detailed local information we are in danger of either devoting time and money to programmes which do not address a real need, or, more alarmingly, withdrawing from programmes when objectives which do not reflect needs have been met "successfully".

5. Variation in nutritional status by district.

With the exception of North Tokar, the nutritional status of children in the southern half of the province was worse than that of children in the northern half of the province. Although there is low rainfall in the north of the province, the absolute variability in rainfall is not great (see the accompanying paper by Roy Cole "Drought, food stress and the flood and rainfall record for Red Sea Province"). The north is therefore adapted to conditions of permanent drought, and did not experience a dramatic worsening of conditions in 1984/85. The worst nutritional status was seen in the south of the province, which has the highest variability in rainfall and flooding and experienced the biggest drought impacts in 1983/84. Nutritional vulnerability is not linked solely to absolute conditions but to variability in conditions and the degree and frequency with which exceptional conditions occur (see the accompanying paper by Roy Cole "Measuring drought and food insecurity in Red Sea Province in 1987 and 1988: a technique for the rapid assessment of large areas.").

The higher nutritional status of children in North Tokar district than the rest of southern Red Sea Province could be a reflection of the relatively rich resources of that district. These include agriculture in the Tokar delta and major khors, the economic opportunities and facilities associated with the town, and rich natural resources for local rural industry. South Tokar, which had the worst nutritional status in Red Sea Province in 1986, is exceptional also in that it has a high population of Beni 'Amer "refugees" who have become destitute through the combined effects of drought and restricted access to traditional grazing and agricultural areas in Eritrea caused by war. In addition, there is a high population of Eritrean refugees both in camps and absorbed into the areas around larger towns such as Garora. Traditional land rights restrict the movement of the Beni 'Amer into the North, the war in Eritrea has cut them off from their own traditional areas in the South, and the situation has been exacerbated by the pressure of incoming refugees. South Tokar district can be cut off from the rest of Red Sea Province for several months during late summer because of flooding in Khor Baraka.

Those districts with above average nutritional status in 1986 (North Tokar and Rural Port Sudan) did not improve in terms of percentage of malnourished children in 1987, despite a general improvement in nutritional status province-wide, regular relief food deliveries and good rainfall. Possible reasons for the lack of improvement include:

a. The level of malnutrition seen in these districts in 1986 represents the best possible level achievable in the prevailing conditions. That is, malnutrition is primarily a result of variables such as poor water supply, infectious disease, poverty and a lack of entitlement rather than a result of inadequate food supplies. This being the case, relief rations would not have a significant impact on malnutrition rates, and an alternative strategy focusing on primary health and a general strengthening of the ability to cope with environmental and economic variability would be more appropriate.

b. The level of malnutrition seen in these districts in 1986 is a level that is accepted either by individual families or the community. That is, the child's health and development are considered to be normal and there is no obvious functional disability, despite being below internationally recognised cut offs for malnutrition. This will be referred to as an "acceptable" level of malnutrition. Once this level is reached, resources may be thought to be more appropriately channelled into strategies to increase future security rather than into immediate household consumption. Given the extremely marginal and variable nature of the environment it is likely that long term strategies for increasing security would have a high priority once children were perceived as being out of danger. The comments made previously about the need to investigate local perceptions of malnutrition and their relation to child health and welfare apply here also. If a threshold of acceptable malnutrition does exist, and it is at a level which compromises child health and development, vulnerable children should be targeted with commodities specifically for their consumption rather than assuming that the benefits of a more general ration will trickle down. Examples such as corn soy mix (CSM) or a more varied basket of local commodities suitable for weaning foods are discussed above. Children are unlikely to benefit from supplementary foods when the food supply for the whole family is inadequate. A twopronged approach that ensures sufficient food for the family together with targeted foods for weaning age children stands the best chance of success.

Halaib district moved from having average rates of malnutrition in cycles 1-3 to having the lowest rates of malnutrition in Red Sea Province in cycles 4-6. Although the nutritional status of children in Halaib was better during cycles 1-3 than in several other districts, the changes seen in cycles 4-6 shows that there was still potential for improvement. Actual changes in nutritional status must be interpreted in terms of potential for change, which includes resource availability, prevailing conditions such as rainfall, and external inputs such as relief rations. The recent ERGO/Oxfam low level aerial survey of Red Sea Province has shown Halaib to be the least densely populated district, with an estimated population of approximately 17,000. This figure is substantially lower than the population estimate used for relief distribution. It is probable that the actual amount of relief grain distributed per family was considerably higher than that allocated. This, coupled with the fact that no severe drought effects were felt in Halaib district, may have contributed to the improvement in nutritional status.

Haya district was the only district with worse than average malnutrition rates that failed to improve between cycles 1-3 and 4-6. In cycles 4-6 Haya district had the worst malnutrition rate of any of the districts in Red Sea Province. The lack of improvement is particularly disturbing in view of the recent aerial survey findings that 28% of the total population of Red Sea Province lives in Haya district for all or part of the year. It was the district which suffered the worst drought impacts in 1983/84, and together with Derudeb the district in which camps for drought displaced people sprang up. There was no improvement in nutritional status between cycles 1-3 and 4-6 despite good rains and floods in Khor Arab, and regular relief distributions. Possible explanations for the lack of improvement must include those hypothesized for the lack of improvement in better than average areas; that the rate of malnutrition represents the best possible rate for the district, and/or the level was considered acceptable. The first explanation seems unlikely in view of the potential grazing, agricultural possibilities of Khor Arab and related areas, and the links and economic opportunities afforded by the railway and tarmac road.

Were relief rations allocated to Haya district inadequate? From the aerial survey figures it appears that the population of Haya district may have been underestimated, and the amount of relief actually available to each family less than that allocated Allocations and monitoring of relief in Haya district (with the exception of Tahamyam) was the responsibility of the Sudanese Red Crescent rather than Oxfam. The SRC has been and still is autonomous, and makes independent judgements on allocation levels. SRC seems to have been in favour of lower rations and more rapid ration reductions than was Oxfam policy for the rest of the province.

Another explanation for the continued poor nutritional status in Haya district could be that as families recovered from the effects of the drought they returned to their home areas, leaving behind a residual population of poor and destitute families. Alternatively, parts of families may have moved away leaving women, the elderly and sick children at the relief dumping points and camps. The populations being measured in cycles 1-3 and 4-6 would therefore not be comparable. With seasonal migration a strong feature of Beja life (and indeed an essential survival strategy) the problem of defining the population being surveyed at various times of the year applies to all districts and adds yet another complicating factor to survey design and the interpretation of results.

There is not enough information available to assess retrospectively the plausibility of any of the proposed explanations for the lack of improvement in nutritional status in Haya district. It is a pity that the emphasis of the nutritional surveillance was on a quantitative statement of the degree of malnutrition to the exclusion of a more holistic investigation of the causes of malnutrition. The quantitative approach was rooted in the original function of the nutritional surveillance teams for rapid assessment in an emergency situation, and failed to evolve with the changing emphasis of the relief programme and the changing conditions in the province. The opportunities to investigate and learn from situations such as the lack of change in Haya district were lost.

6. Malnutrition rates by height of the child.

The highest rates of malnutrition in Red Sea Province were found in children 65.1-75 cm in height. This corresponds to approximately ages 6-12 months. Disturbingly high rates of malnutrition were found in children 55.1-65 cm tall also. It would be expected that breastfeeding would play a major role in meeting nutritional requirements of children in these height categories, and therefore that rates of malnutrition would be relatively low. Aside from the possibility of a sampling bias towards babies who were sick, and the error introduced by the non-standard measurement technique used for children in these height categories, the high rate of malnutrition in under one year olds in Red Sea Province could be a reflection of inadequate breastfeeding and/or low birth weight babies who fail to catch up with expected growth patterns.

Reports from the field suggested that a large percentage of lactating mothers were unwell. Diseases such as scurvy and severe anaemia have been noted on field trips in 1988. Women do not have access to such a variety of foods as men because they are restricted to the tent. Men eat outside the tent, sharing food with neighbours. Men are served first, before women and children eat the remaining food. Men travel away from the community have the opportunity to eat in restaurants and coffee shops.

Maternal malnutrition affects both the quantity and quality of breast milk, particularly when the mothers' own body stores have been depleted by persistently poor nutritional intake and the stress of disease and pregnancy. Poor health could also affect infant feeding practices. Personal observations and anecdotal evidence from midwives suggest that some women who are in poor health will either stop breastfeeding or not establish breastfeeding even if their breast milk is adequate. Breast milk production is stimulated by feeding. If feeding stops or is infrequent a vicious cycle develops where breast milk production decreases and the mothers' perception of her inadequacy for breastfeeding is reinforced. More information is needed on women's perceptions of suitable and unsuitable health and environmental conditions for successful breastfeeding.

Maternal pre-pregnancy nutritional status and weight gain during pregnancy affect the birth weight of the infant. Poor nutritional status and low weight gain in pregnancy can result in low birth weight babies, who are effectively malnourished from birth and subsequently at high risk for diseases which would worsen their nutritional status further. A prerequisite for improving the nutritional status of children is to improve the nutritional status of women. All women of child bearing age should be considered for targeting as a vulnerable group. Particularly at risk are very young women who face the nutritional stresses of pregnancy and lactation before they have completed their own growth. In Red Sea Province it is common for women to be married and bearing children before they are out of their 'teens.

Targeting pregnant and lactating women for what the World Food Programme describes as "vulnerable group feeding" poses many problems besides the most obvious of identifying and reaching such women. Traditional practice is to restrict food intake in the third trimester of pregnancy to avoid large babies. Large babies are thought to cause difficult births, and to have been afflicted with the evil eye (belief in the evil eye is powerful and influences all aspects of rural Beja life). The period of food restriction coincides with the time that it is customary for a woman to leave her husband and return to her family for the birth; an extra mouth to feed in a household that may already be strained. In addition to a general restriction of food intake there are specific food taboos that apply in pregnancy. These seem to vary but examples of forbidden foods include eggs, dates and camel's milk. Any supplementary feeding programme must take such practices into account.

Policing household level food distribution is obviously impossible and undesirable. It is unlikely that commodities distributed for pregnant and lactating women will be consumed by them unless there is sufficient available for the whole family. One approach which has been used successfully in the Gambia is to provide pregnant and lactating women with high protein high energy biscuits which are less likely to be absorbed into the family food supply. This approach is impractical for Red Sea Province as there is no health care infrastructure through which to distribute the biscuits, large external inputs would be required, and the system would not be locally sustainable. Given adequate environmental conditions there is a sufficient quantity and variety of local foods to satisfy the nutritional requirements of pregnant and lactating women. What is needed is a greater understanding of women's support structures, their priorities for and control of their own income, their influence on decision making at various levels and how these change with factors such as wealth and drought related stress.

Of children over 75 cm tall, highest rates of malnutrition were found in the 75.185 cm height category. This corresponds approximately to children 12-24 months of age. In Red Sea Province weaning begins with the gradual introduction of supplementary foods between six and nine months of age, and is completed between 1824 months of age. Breast milk alone becomes inadequate to meet the nutritional needs of a child at around six months of age. Results from this study indicate that children of weaning age were at high risk of malnutrition. Children of weaning age have high energy and nutrient requirements for growth, but cannot take large quantities of food. Weaning foods must therefore have a high energy and nutrient density and be fed in small frequent amounts. In Red Sea Province weaning foods include ground sorghum porridge (asayda), goat milk, goat ghee, rice, and water in which dried dates have been soaked Sour milk and camel milk are not considered suitable for children. Breast feeding continues for up to two years, or until the woman becomes pregnant.

Successful weaning in rural areas is largely dependent on the availability of animal products such as milk, ghee and animal fat. The availability of these products varies both seasonally and from year to year. Replacements would be expensive and available only in towns. The vegetable oil and sugar provided in the relief ration are energy dense, but do not contain the wide range of nutrients found in milk and milk products. The vegetable oil (rape seed oil) is unpopular with some people because it has a fishy smell. It has been difficult to ensure equitable distributions of oil and sugar; they are valuable products which makes them particularly prone to "leakage". Oil and sugar may be sold by the recipients to provide cash. Sugar is highly prized and would be given preferentially to men and guests.

The most satisfactory way of permanently improving the health and welfare of weaning age children would be through long-term strategies to increase the resource base for pastoralism, investment in agricultural infrastructure and increased economic opportunities for the community as a whole. In the meantime, if a relief approach is adopted vulnerable weaning age children need to be targeted with more appropriate foods. Dried full fat vitamin A fortified milk is one, albeit controversial, possibility for the summer months when milk is in short supply. Corn soy mix was briefly distributed in Red Sea Province, with mixed reactions. Oxfam teams reported that the CSM was unpopular, and that women winnowed the mixture to extract the sugar. The Sudanese Red Crescent, however, reported enthusiastically that CSM had "the magical power to remove marasmus from our children" and asked for distribution to be extended. An alternative would be a relief ration that consisted of a more varied basket of nutritionally dense foods which would be suitable as weaning foods. Oxfam has tried this approach in some areas of Darfur, where families with malnourished children were targeted with a ration that included lentils and oil. An essential part of any such targeted programme would be an effective distribution and monitoring system. This should be done by women and through women, despite the difficulties that this poses in Red Sea Province.

Ideally, nutritional support to weaning age children would be part of a wider health care strategy that addresses issues such as immunisation and treatment of diarrhoea which directly affect nutritional status. Some common traditional medical treatments administered to children of weaning age result in either increased nutritional requirements because of physiological stress or decreased intake of nutrients. Examples include burning the head or other areas of the body to treat oedema (a common treatment for many other diseases also), and incising the gums and removing four teeth to treat tasniin. a disease associated with teething. Health and nutrition education for rural women should be increased. What primary health care facilities that are available concentrate on therapeutic rather than preventive and promotive health care. Community health workers are exclusively male.

7. Seasonallty In malnutrition rates.

Classification of seasons by month and ecozone failed to detect changes in malnutrition rate by season. The method used for classification was too rigid and insensitive to the great inter and intra-annual variation in rainfall and flooding seen in Red Sea Province. Seasonal changes in malnutrition rates are lagged behind the actual seasons. Changes are therefore masked, particularly as the rainy season is very short and the dry season long.

When re-examined by cycle, monthly changes in malnutrition rate became apparent. The worst rates of malnutrition were seen at the end of the dry season; the cumulative effect of the associated scarcity in cereals and grazing, reduced milk production and poor water supplies. Red Sea Province in the dry season is extremely harsh, with high temperatures and frequent dust storms. The incidence of respiratory and eye infections increases dramatically in the summer months, part of a vicious cycle of worsening nutritional status and increasing disease susceptibility. Lowest rates of malnutrition were seen in January and February, after the rainy season and harvest, and when improved grazing has resulted in increased milk production and livestock numbers.

When broken down by district the pattern of seasonal changes in malnutrition rates was not seen in the north of the province (Halaib and Rural Port Sudan districts). As discussed previously, the north of the province receives very little rainfall, and the distinction between the seasons is blurred. Seasonal changes in malnutrition rates in Rural Port Sudan district may have been modified by access to the supplies and facilities available in Port Sudan. Greatest seasonal variation in malnutrition rates occurred in the south of the province, which has the highest variability in rainfall and flooding.

It is essential to take seasonality into account when comparing results of nutritional surveys conducted at different times of the year. Otherwise, an improvement in nutritional status could be unjustifiably attributed to an intervention such as provision of relief radons when it is a reflection of normal intra-annual variation.

The provision of a stable relief ration throughout the year did not mask seasonality in malnutrition rates. This could be for three reasons:

a. The ration did not provide the total requirements for energy and nutrients.

b. In terms of nutritional value, sorghum, oil and sugar are not a substitute for milk and milk products which are essential for children and are always in seasonal short supply unless the family has a large herd and is highly mobile.

c. Non-food factors which influence malnutrition rates and vary seasonally, such as disease incidence, are playing an important role in the aetiology of malnutrition in Red Sea Province.

8. Malnutrition rates by settlement type.

Camps for drought displaced families sprang up in early 1985, largely in Haya and Derudeb districts. They were usually associated with an existing town. The nutritional status of children living in settlements classified as camps was significantly worse than the nutritional status of children living in other settlement types in 1986, but not in 1987. The improvement in the nutritional status of children in camps in 1987 could have been due to either or both of the following:

a. Camps being disbanded and families returning to their home areas, leaving a residual urbanised population and blurring the distinction between camps and towns.

b. The nutritional status of the camp inhabitants improved.

There was no difference between the nutritional status of children living in railway settlements, towns and rural areas. It was expected that the nutritional status of children in railway towns would be better than the nutritional status of children in towns or rural areas. Railway workers were relatively well paid in 1986 and 1987 (LS 350 per month), they have access to food supplies delivered by train, and water supplies are delivered by tanker from Atbara or Port Sudan. This was not the case. Possibilities for the lack of difference between railway towns and other settlements include:

a. Within railway towns the children of railway workers were not differentiated from the children of non-railway workers.

b. The majority of people living close to the station are not employed by the railway.

c. Railway towns have large satellite communities who have settled to take advantage of the facilities, often because of stress in their home areas. Railway towns could therefore be made up of two populations; relatively advantaged railway workers, and disadvantaged rural people on the periphery. This would not have been picked up from the previous survey design.

d. There could have been no difference in the nutritional status of wealthy and poor families, either because malnutrition was caused by non-food factors or because the level of malnutrition was one which was acceptable to wealthier families. Oxfam nutritional surveillance teams in Darfur found no difference in rates of malnutrition in the children of wealthy or poor families.

The possibilities outlined above apply to the lack of difference between the nutritional status of children in towns and rural areas also. Towns have more facilities and economic opportunities than rural areas, but also have a peripheral population of very poor families. It is possible that similar populations are being measured in towns and rural areas as families move between them at different times of the year.

9. Rates of malnutrition by food security zone.

As might be expected, areas with low food security and high drought impacts had higher rates of malnutrition than areas with medium and high food security and low drought impacts. There was no difference, however, in malnutrition rates between areas with medium and high food security and low drought impacts. Possible explanations for this include:

a. An acceptable rate of malnutrition is reached in the medium food security zone, and represents a threshold below which malnutrition rates do not improve despite increased food security.

b. Malnutrition in areas with medium or high food security is caused largely by factors not related to food availability.

c. The food security zones are too large and mask variability within them. The zones were based on ecological rather than human criteria. There was a general east-west bias in the zoning, whereas differences in nutritional status seem to vary more in a north-south direction. A more detailed analysis of food security in Red Sea Province has since been carried out, but because of time pressures could not be included in this analysis.

d. Migration between the food security zones could have resulted in the same populations being measured in different food zones at different tames of the year.

If points a and b are valid, the usefulness of nutritional status as an indicator of food security is reduced. If point d is valid, it needs to be a major consideration in planning future sampling strategies.

10. Malnutrition rates and the relief ration.

The calculation of the average relief grain ration per family for the cycle during which the nutritional surveillance took place was not ideal. Nutritional status is a lagged variable; it does not change immediately but reflects conditions at some time in the past. To fully understand the relationship between the relief ration and malnutrition rates a historical perspective is necessary. Changes in malnutrition rate should be examined in relation to relief input over time rather than looking for relationships between absolute figures at one point in time. Unfortunately neither the Relief Section records nor the nutritional surveillance data were in a form which could be used in a longitudinal analysis without considerable work, well outside the time frame for this study. Coordination between the Relief Sections and nutritional surveillance teams to collect data in a form that would be amenable to continuous longitudinal analysis should be a priority in planning future relief efforts. Examples include better co-ordination between relief allocation zones and sampling units for nutritional surveillance, consideration of the timing of relief distributions in planning the frequency and timing of nutritional surveillance, more serious efforts to time relief deliveries on a regular basis, and making available cumulative records of relief inputs and running averages for allocations.

The calculation of the average relief grain ration per family per day at the time of the nutritional surveillance was considered worthwhile despite the limitations outlined above. This was for three reasons:

a. It was the best measure of average relief rations available.

b. As a rule, relief allocations did not vary greatly within one dumping point in 1986 and 1987.

c. The beneficiaries are thought to manage deliveries in such a way as to smooth fluctuations in the availability of grain through the cycle. Examples include borrowing grain against the next relief delivery if that delivery is delayed.

The lack of correlation between malnutrition and the relief ration is probably best explained by the limitation of the analysis outlined above. Other factors which complicate a direct relationship between the relief ration and malnutrition rates include:

a. Food availability may not be reflected directly in malnutrition rates because of the influence of non-food factors in the aetiology of malnutrition, and because of the possibility of an acceptable level of malnutrition (as discussed previously).

b. The nutritional status of children may not necessarily reflect the total needs of an area for relief input. Improving the nutritional status of children was not the only goal of this relief operation, and the need for relief could therefore remain after malnutrition rates have reached their lowest possible rates.

c. Nutritional surveillance data were not, or possibly never could be, appropriate in terms of time and space to make a meaningful contribution to the cycle by cycle management of relief allocations. Could such data be used more appropriately for long term planning and evaluation (for instance setting annual targets for total inputs), and/or more detailed case study type information to address the causes of malnutrition, which in turn would guide policy?

11. Variation in malnutrition rates by settlement.

In the 1986/87 Sudan Emergency and Recovery Surveillance System (SERISS) study of nutritional status in Northern Sudan (Ministry of Health/USAID), 28% of the total variation in weight for height z-scores was explained by the 22 variables examined in the survey. The remaining 72% of variation remained unexplained The Oxfam nutritional surveillance teams collected information on similar but fewer variables than the SERISS survey, and it is unlikely that regression of the Oxfam data set would result in a higher percentage of explained variance in nutritional status than that found in the SERISS study. If the SERISS figure of 28% explained variance is taken as a guide, the effect of settlement in explaining variation in percent weight for height in the Oxfam survey was considerable. It is probable that settlement was the single most important variable in explaining variation in nutritional status in the Oxfam data set, as it was in the SERISS data set. The practical implications of this finding are:

a. Sampling should include as many settlements as is practical in order to ensure an accurate assessment of the district and province In Red Sea Province family size is small and the number of children present in each settlement typically small. Sampling therefore inevitably involves visiting a large number of individual settlements.

b. Investigations of the causes of malnutrition should concentrate on looking at individual settlements and the differences between individual settlements which are influencing nutritional status. This was the approach taken by the nutritional surveillance teams in 1988 and 1989 (see the paper by Fatma Gebreil "An analysis of areas in Red Sea Province with persistently poor nutritional status", Oxfam Port Sudan 1989.).

c. General relief rations and/or supplements for vulnerable groups need to be targeted by settlement. Within one allocation zone there will be a wide range of need. This implies close monitoring by people with detailed local knowledge.

The importance of the settlement in explaining variation in percentage weight for height is probably the result of it being the smallest unit of analysis in the study. Factors affecting even smaller units such as the individual tent or family group may prove even more important in explaining variation in percent weight for height than those which affect settlements, but in rural Red Sea Province at the present time research at the household level is difficult to the point of practical impossibility.

12. Comparison of Oxfam nutritional surveillance results with SERISS results.

The persistently higher nutritional status found in the SERISS surveys than in the Oxfam surveys is most likely to be a result of differences in the sampling frames, the sampling intensity and the distribution of the sampled areas in the two studies. It is not a case of one study being "right" and another "wrong", but rather of them providing different information for different uses. The Oxfam nutritional surveillance had Red Sea Province as the area of study, with the district as the unit of analysis. The SERISS study had Northern Sudan as the area of study, with the province as the unit of analysis. Oxfam sampled exclusively from rural areas, whereas SERISS stratified their sample by urban and rural population.

The urban component of the SERISS study was probably responsible for the higher nutritional status than was found by Oxfam. In addition, SERISS sampled in fewer districts and settlements than Oxfam. As has been discussed previously, this would have important implications for capturing all the variation in nutritional status found in Red Sea Province. It would be valuable for Oxfam to collect or collate more information on the nutritional status of children in Port Sudan. Seasonal migration of families to Port Sudan means that the urban populations in some strongly Beja areas of the city such as Daym al-'arab could be the same as the population being measured by Oxfam nutritional surveillance teams in rural areas. Periods spent living in Port Sudan could be an important factor influencing the nutritional status of such populations.