Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Nutritional profile of the population in a food-for-work project area: A case study from Samburu District, Kenya

A. A. Kielmann, N. S. Kielmann, A. A. J. Jansen, D. N. Njama, G. K. Maritim, R. Mwadime, and K. Saidi



The nutrition status of Samburu tribesmen of northern Kenya and the potential nutritional effects of a food-for-work programme were assessed by anthropometric and dietary surveys over a period of five weeks. The investigations were carried out at four input localities, where the programme had provided a daily average of 173 kcal and 5 g protein per capita over the preceding year, and a fifth control area.

Compared with Kenya as a whole, the Samburu community showed an unexpectedly high prevalence of wasting: 74% of the surveyed individuals were below 90% of the NCHS weight-for-height median, and 34% were below 80%. Relative wasting was especially prevalent among school-age children, of whom 86% were below 90% and 44% were below 80% of the standard, and among elders, with 88% and 64% below these thresholds respectively. Mid-upper arm circumferences were lowest among school-age children (mean 14.6 cm). Haematocrits were low in all age and social groups. Aside from three preschool children with xerophthalmia, a number of both preschool and school-age children had clinical histories suggestive of earlier vitamin A deficiency. Goitre was more prevalent among the older generations than among school-age or preschool children, suggesting a periodic but not uninterrupted availability of iodized salt. In the input locations, both anthropometric and haematocrit values among the most vulnerable members of the communities, preschool children and women of reproductive age, were significantly higher than those for comparable groups in the control location. Overall, the food-for-work programme was judged to be a worthwhile effort to help Samburus to be once again self-sufficient.


Over the preceding decades, the nomadic populations of Africa as a whole, and of Kenya in particular, have experienced a progressive decline in their ability to maintain their food security. The causes are many and include an ever-increasing population, loss of land as a result of encroachment by agriculturists, deterioration of the land through erosion, mainly because of overgrazing, and failure to develop alternative methods to provide for their livelihood. The Samburus, one of the Nilo-Hamitic tribes of Kenya, are not exempt from these changes.

As the closest cousins of the legendary Maa'sai, and because of their still well-preserved cultural heritage, the Samburu people have attracted considerable interest among workers in a wide variety of disciplines. A number of organizations are active in some form of research or development work, whether to study their beliefs and customs, convert them to a religious denomination, or assist them to become self-sufficient, as they were for past millennia.

Several projects fall into the last category, mostly of an agricultural nature, usually sponsored by bilateral agencies and non-governmental organizations. Since October 1982 the government of the Federal Republic of Germany has supported a food-security/ food-for-work programme (WFSP) in Wamba Division of Samburu District, Rift Valley Province.

The project area lies almost in the centre of Kenya, about 400 km due north of Nairobi, on the eastern border of the Rift Valley. Samburu District, with 3.8 persons per kmē (compared with 38.3 for Kenya as a whole in 1987), is one of the less densely populated areas of the country; and Wamba, one of the three divisions of the district, is even more sparsely populated, with only 2.5 persons per kmē. Although the district has an average of more than 600 mm of rainfall per year, a large proportion of the water is lost to run-off and evaporation, and so the lands are arid or semi-arid.

The principal objective of the project was to increase the self-sufficiency of the nomadic Samburu population of Wamba Division in food, mainly of animal origin. This was to be achieved primarily through enhancing animal husbandry and improving the range. Among the measures relating to animals were selecting a mix of livestock that would be less destructive to the environment and hence ecologically more appropriate, improving veterinary services, upgrading and making watering places available, and improving marketing facilities for livestock. With the reduction of grazing lands and a number of protracted dry spells, the number of cattle per individual had declined significantly over the past decade. Goats and, to a lesser extent, sheep have progressively replaced traditional cattle herds. To counteract this trend, camels were introduced into the area, as they give milk and meat and may be used as beasts of burden but are far less destructive to the environment than goats. The major foci for improving the environment included arresting existing soil erosion, regenerating topsoil, and exploring and, where feasible, demonstrating possibilities for range improvement. Soil erosion was effectively arrested through sequestering plots, drawing contour lines in such demarcated areas, and planting indigenous shrubs, grasses, and trees. After the vegetation had successfully recovered, the new mix of livestock was admitted.

Adult Samburus, mainly women, from the manyattas in and around the project locations (a manyatta is a group of dwellings under one chief and within a well-circumscribed boundary) provided the labour for all the physical activities of the programme, especially those dealing with soil and range improvement. Though a family might participate more than once in any given year, there was considerable rotation among households so that, theoretically, one or several members from every household could have participated at least once. The participants usually were selected by the elders of the location concerned, largely, though not exclusively, on the basis of the family's socio-economic status. According to the project director, the majority came from among the poorer groups.

The participants were paid for their work weekly in both cash and food. For a full six-day week's work, each one received 80 Kenyan shillings (approximately US$4.40 at the time of the survey) plus a ration comprising 20 kg of maize meal, 5 kg of beans, and 1 kg of animal fat, providing in all 91,800 kcal, 2,700 g of protein, 7,200 mg of calcium, and 630 mg of iron. Theoretically 91,800 kcal would meet the caloric needs of a family of 5.8 persons for 7.5 days, assuming an average requirement of 2,100 kcal per person per day.

Records were kept on the total number of rations distributed per manyatta as well as on whether the participant was a woman, a young man (a moran— see definition below), or an elder, but not by the name of the individual or family.

At the time of the survey reported here, the project was about to complete its first implementation phase, and its activities as well as its impact were to be evaluated before it was extended. An evaluation of the land-reclamation activities and livestock management showed that arresting range land degradation and, more important, the regeneration of healthy pastures was indeed possible [1].

Research issues

After discussions with the project directors, it was decided that, besides establishing a nutritional profile for the population living in the project's catchment area, an additional objective of the study would be to demonstrate whether the food and money that was being paid in return for work had any demonstrable effect on the nutrition status of the population, specifically its most vulnerable groups. Considering that the work carried out by the participants was entirely physical and rather strenuous, the question arose whether the food ration was sufficient to satisfy the additional energy requirements necessitated by the extra workload.

Methods of investigation

The investigation was carried out at the end of 1987, following the short rains. This period is characterized by a relative food shortage, with the harvest not yet ready.


Sample selection and structure

Five survey sites—four input locations out of a total of ten in which the project was active and one control—were selected jointly by the principal investigator and the director of the WFSP with the aim of including all possible location variables. The input locations included one trading centre (Lodungakwe); two large manyattas, representative of the majority of manyattas, where the project had been operative since 1984, one of them (L'Kisin) close to and the other (Oro Modei) distant from the project headquarters (Wamba); and one rather remote manyatta (Niri Mirimo) in a more arid section of the division, where the project had started in early October 1985.

Ngutuko'ngron, a cluster of manyattas with approximately 400 inhabitants in the south-east of the division, where the food-for-work programme had not yet been implemented but was planned for at a later date, was selected for control purposes to identify the nutritional effects of the programme on the participating individuals and their families. (An earlier suggestion to choose controls from among non-participant families in the input locations was rejected on the grounds that probably most, if not all, of the families had participated at one time or another.) The control manyattas were close to the project headquarters and were comparable to L'Kisin in population density, socio-economic status, topography, and soil conditions. One major difference from the input locations was that the control manyattas were near—less than two kilometres from—the Samburu River, which carries water all year round. All the locations, input and control, had undergone and were recovering from a severe drought in 1983, during which the greater part of their cattle had died.

Before the survey started, the number of people in each of the selected locations was established from project records and census updates, as well as by consultation with the chiefs and sub-chiefs of the areas. The total population of the five study locations (approximately 3,900) represented about 17% of the people of Wamba Division. A sample size of 20%-25% of the population in the selected sites, primarily from among the more vulnerable sectors—preschool children, school-age children, and women—was judged to be adequate for statistical precision.

Obtaining population participation

Soon after the survey was decided on, the WFSP project officer informed village and civil authorities of the plans for the investigation. Two to three days before starting the surveys, the principal investigator together with members of his team and a representative from the WFSP visited the sites to explain once again the reasons, objectives, and procedures for the survey to assembled elders. They requested the elders to call all the people living in the defined area who were physically present on that day to gather at a specified location where all the surveys and examinations would take place.

This method of sampling clearly carries the risk of introducing a bias in favour of those who come, that is, the self-selected, and predominantly those remaining at the manyatta (women, preschoolers, elders), and excluding those active in tasks that take them away from it (morani, school-age children). In addition, since more people would be idle in the control manyattas, more would be available to be surveyed. However, it was hoped that the bias introduced by this method of self-selection would run against our hypothesis, since it was assumed that people staying at home in the input villages would be the weaker ones, with the stronger ones being active in the WFSP. Differences identified in nutrition status, as well as in the sources from which food was procured, would hence result mainly from differences in inputs from participation in the programme. As we will see later, this assumption was most likely wrong.

Socio-economic and age/social group categorization

To isolate the effects of socio-economic status on nutrition and health, if any, all the households were classified into one of two locally used and well-known groups, the "able" and the "unable." This classification is well established in the Samburu culture and language and hence does not require definition for its determination, nor does it seem to be a sensitive issue. By definition, the able are those who are entirely self-sufficient with respect to food and shelter. The unable are households that depend either entirely or partly on the support of the community for their livelihood. Usually the able have large herds of cattle, goats, or camels, or, in the trading centres, own a shop or hold a salaried job. At the end of a typical survey day, the principal investigator or one of the other two physicians conferred with the chief and sub-chief of each location to classify the individual households according to these two groups. Of the 894 people seen, 38 (4.3%) could not be thus classified, 27 from the input locations (4.4%) and 11 (3.9%) from the control location.

To determine specific age effects, the population was divided along traditional cultural lines into five age and social groups: small children (preschoolers), school-age children, women (> 15 years old and circumcised), morani (circumcised men 15-25 years old, traditionally responsible for looking after the cattle herds and defending the manyatta against raiders), and elders (men >25 years old). In the course of a pilot survey it became evident that precise ages could not be obtained even for preschool children. The use of a local calendar of events was extremely time-consuming and was done only to separate preschoolers from school-age children.

Determining participation in the food-for-work programme

Using WFSP records, a participation index was determined for each eligible age group (women, morani, and elders) in the four input locations, as well as for each location as a whole, by dividing the number of rations earned by a specific age or social group by the total number of individuals in the group. This provided a ready indicator of the average annual frequency of participation of each eligible member of the community in the work programme.


Specific investigations

It was necessary to determine both the total amount of nutrients provided to the input manyattas and the nutrition status of the discrete age and social groups in addition to carrying out clinical examinations for specific nutritional deficiency states.


Weight was taken to the nearest 0.1 kg using a Seca digital scale. For infants and small children, a Salter spring scale with 100-g gradations was used. Up to approximately three years of age, length was determined in the supine position; above that age it was taken in the standing position using a height-measuring stick with a baseboard. Mid-upper arm circumference (MUAC) was measured on all individuals using a Zerfass tape.

In all locations anthropometric measurements were performed by the same persons, who had been trained as anthropometrists for the UCLA-University of Nairobi collaborative investigation on determinants of child growth in Embu, and two specifically trained investigators as well as either of the two physicians. Both intra- and inter-observer errors were measured before beginning the surveys and were found to be within acceptable limits.

Since, as noted, precise ages could not be obtained, it was not possible to use age-dependent reference charts to determine specific nutrition status. The US National Center for Health Statistics (NCHS) weight-for-height medians were therefore used as an age-independent measure, with 90% and 80% of the median taken as thresholds for assessing the extent of borderline (mild) and frank (moderate or severe) wasting respectively for the total population as well as for the specific population subgroups.


Blood samples were drawn by finger prick, using heparinized capillary tubes, to determine haematocrit. These were analysed using a special haematocrit reader.

Clinical examination

All individuals were subjected to a clinical examination by one of the three physicians for skin and eye pathology, goitre, and status of teeth.

Contents - Previous - Next