Selective Feeding Programmes in Wadjir: Some Reasons for Low Coverage and High Defaulter Rate

By Fabienne Vautier

The coverage and default rate in selective feeding programmes are taken as proxies of the accessibility and acceptability of these programmes. This article by Fabienne Vautier describes the problems of low coverage and high default rates in Therapeutic and Supplementary feeding programmes run by Médecins Sans Frontières (MSF) Belgium in Wadjir in Kenya. The programme started in April 1998 and covered a three month period.

Wadjir is the capital of the district of Wadjir (56.500 m2), the district borders Ethiopia and Somalia. The district's population is estimated at between 200,000 and 300,000. Wadjir town has about 60,000 inhabitants, if the 12 peripheral villages (bullas) are included. The bulla population is approximately 52,000 with 12,000 children under five years. Almost 80% of the population are nomadic pastoralists of Somali ethnic origin who depend on livestock for both consumption and income. The semi-arid climate is not conducive to crop production. The community is organised into clans and sub-clans that are governed by elders who administer local customs and laws.
Since 1990, the area has been affected by droughts (1991-92 and 1996) which resulted in significant livestock losses and 'sedentarisation' of destitute pastoralists in the bullas of Wadjir town. The population of the bullas live off petty trading and small livestock.
In November 1997, the El-Niño phenomenon caused severe flooding resulting in further loss of livestock (caused by epidemics) and population movement. Food relief was organised by the Government of Kenya and NGOs. The situation was aggravated at the beginning of 1998 by a major outbreak of malaria killing many people. A nutritional survey carried out in February 1998 showed a high prevalence of malnutrition, 25.3% (<-2 Z Scores or oedema) and of severe malnutrition 3.7% (<-3 Z-scores or oedema). A retrospective mortality survey found an alarming crude mortality rate of 9.3/10,000/day and an under 5 year mortality rate of 28.4/10,000/day. This mortality rate covered a 2 month period from the previous Ramadan (January 1998 to March 1998).

MSF began working in Wadjir in April '97 on a sanitation programme (rehabilitation of water points throughout the district). Following the flooding the programme was extended to include epidemiological surveillance and cholera preparedness measures. When the malaria outbreak occurred, mobile malaria clinics were set-up. In April '98 the programme incorporated a nutritional component. This involved setting up 2 TFCs and 2 SFCs for the population of the 12 peripheral villages of Wadjir town.

Problems encountered in the nutritional programme

Three weeks after the start of the nutritional programme, the coverage^* was only 24.5 % in the TFCs and 40.4 % in the SFCs. Few mothers turned up for the screening of their children and many of those referred by the home visitors did not come to the centre. Of the 116 children admitted to the TFCs, 25 defaulted. Many children left the centres several days before they reached the discharge weight. Tracing was carried out, but the home visitors reported that mothers were reluctant to come for clan-related reasons, i.e. they did not want to attend feeding centres which were employing staff from other bullas/clans. The elders complained about this and that the feeding centres were too far away. In the TFCs, mothers complained about the fact that they were not given tea.

Investigation of the problem

To get a clearer understanding of the underlying reasons for poor coverage and high default rates, MSF conducted focus group discussions with women in the bullas at the end of April. The main points discussed were:

1. the main problems faced by families, and in particular children, in the area
2. the women's perception of the role of health and feeding centres for children
3. the reasons for not coming to the feeding centre.

The main problems in the area were described by most as hunger, poverty and housing, followed by unemployment and lack of latrines. For the children, the main problems were malnutrition and being able to afford school fees. The women seemed to be knowledgeable about signs of illness in their children. Preventive measures against diseases were known (ORS, mosquito avoidance, hygiene). When a child was sick the first strategies were generally praying and use of local traditional medicines (roots, herbs and camel's urine). A visit to health facilities was seen more as a second line strategy. Mothers were aware of the opening of the feeding centres and it seemed that the centres had a good reputation. Nevertheless admission criteria were unclear and some mothers thought that the centres were only for anaemic children.

A number of reasons were given to explain the reluctance to come to the centres. The main ones were:

• mothers were too busy to come - especially for Therapeutic feeding.
• they felt a loss of dignity if they had to go to another bulla (where another clan lived) for help and food.
• they could not accept their children being measured in front of everybody. There was a belief that a child would die if another person saw them being measured.
• they do not like the height measurement being taken with the child lying down as the child looks like he/she is already dead.
• they were afraid about the risk of infections in the feeding centres and also mentioned a lack of hygiene as the trousers (in which the children are placed for the weighing) are not washed from one weighing to the next.
• many children were sent to distant pastoral areas where there is more milk available.

Informal discussions brought out how important the clan-related factors were in preventing attendance at the feeding centres. Mothers were asked about the distance to the centres from their houses, - the large majority of the population lived less than half an hour away.

Discussion

These discussions showed how the overall problem was one of acceptability. Although it was difficult to do anything about the clan related factors, several measures were taken, based on the findings of the focus group discussions, to improve the acceptability of the programmes.

The main strategy to improve programme coverage was to provide better information. Discussions with elders were held to better explain the purposes of the programme, the criteria for employing staff and the admission criteria for children. The logistical, human and financial constraints that would arise by opening one feeding centre in each bulla were also discussed. Elders were asked to help explain all this and convince their communities of the need to attend the feeding centres. A decision was taken to publicise the programme more within the bullas through the use of home visitors, and to explain to mothers that if they could not come themselves then they could send another accompanying person.
A second measure was to improve the screening of children by setting up a mobile screening team going into each bulla to check the weight and height of children at risk ( MUAC < 125 mm). This reduced the likelihood of mothers having to come to the feeding centre for nothing.
The third measure involved improving the services delivered in the centres. Screens were installed for the weight for height measurements, ensuring greater privacy, hygiene was improved in the centres and tea with sugar was provided for accompanying mothers in the TFCs. A fourth measure was to discharge children earlier from the TFC and to refer them to the SFP. At the end of May (the 9th week of the feeding programme), 9253 children had been weighed and measured with 1186 (12.8%) admitted to the SFCs. Of the 1186 admitted, 79% were discharged (cured), 19% defaulted and 2% was transferred.

Even though there was no formal evaluation of the impact of these new measures to improve the acceptability of the programme, mothers were pleased that these measures had been taken. We found that the default rate decreased, while programme coverage remained more or less the same.

The number of weekly admissions (Graph 1) remained below the expected number of beneficiaries based on the results of the nutritional survey conducted in March 98.

But the nutritional survey undertaken in June 98 showed that from March to June, the global malnutrition rate decreased from 25.3% to 12.8% and the severe malnutrition rate from 3.7% to 1.7% . By the time this nutritional programme started in April, the under five mortality rate (Graph 2) and the incidence of malaria were already decreasing. The question arises as to how appropriate it is to use nutritional survey results to estimate feeding programme coverage rates; the previous nutritional survey was conducted several months earlier and the health and nutrition situation is rapidly changing. Even if the impact of the measures taken to improve the acceptability of the programme were difficult to quantify, and may indeed have been limited, there is no question that there is value in getting more information about beneficiary perceptions of a feeding programme and how this fits with traditional practices and beliefs. If nothing else, this allows changes to be made so that programme objectives and beneficiary expectations can be more closely matched.

^*Coverage was estimated from the number of potential beneficiaries, extrapolated from the nutitional survey of March 98 and the number of beneficiaries attending the SFP.