RESULTS
FOR THE HIGHLAND AREAS
OF BARINGO DISTRICT
Nyandiko Mokaya, MD
October 2000
TABLE OF CONTENTS
1.0) SUMMARY OF RESULTS
2.0) BACKGROUND INFORMATION
2.1) Demographic and Geographical information.
2.2) General situation and the impact of the drought
2.3) Justification for survey
2.4) World Vision in Baringo district
2.5) General Food Distribution
3.0) OBJECTIVES OF THE SURVEY
4.0) METHODOLOGY
4.1) Area surveyed map
4.2) Survey design
4.3) Sample size
4.4) Nutritional indicators
4.5) Data collection and analysis
5.0) RESULTS
4.1) Nutrition
4.2) Health
4.3) Feeding practices
4.4) General food rations
4.5) Supplementary food
4.6) General observations and information from secondary sources
6.0) DISCUSSION
6.1) Nutrition
6.2) Health
7.0) RECOMMENDATIONS
7.1) Recommendations for World Vision
7.2) Recommendations for the Ministry of Health
ACKNOWLEDGEMENTS
World Vision Kenya would like to thank the Baringo District authorities for their support during the nutritional survey; we very much appreciated the loan of four of their vehicles to assist with conducting the nutritional survey. World Vision would also like to extend their thanks to the Ministry of Health who provided staff to carry out the survey. Without such support the survey would not have been possible.
1.0 SUMMARY OF RESULTS
Proportion of malnutrition according to different anthropometric indices
ANTHROPOMETRIC INDICES / NUMBER OF CHILDREN(sample size 755) / PROPORTION
(%) / 95% Confidence
Intervals
Z – SCORES
<-2 z-scores + oedema (global acute) / 61 / 8.1% / (6.2-10.0)
38 / 5% / (3.4-6.6)
<-3 z scores + oedema (severe acute) / 23 / 3.1% / (1.8-4.2)
% OF THE MEDIAN
< 80% + oedema (global acute) / 44 / 5.8% / (4.1-7.5)
>= 70% - < 80% (moderate) / 24 / 3.2% / (1.9-4.5)
< 70% + oedema
(severe acute) / 20 / 2.6% / (1.5-3.7)
MUAC (mm)
< 125 mm (global) / 44 / 7% / ()
110-124 mm (moderate) / 43 / 5.8% / (4.12-7.48)
< 110 mm (severe) / 1 / 0.1% / (4.21-7.49)
Oedema cases / 18 / 2.4%
- There were 18 oedema cases in total.
2.0 BACKGROUND INFORMATION
2.1 Demographic and Geographical information.
The World Food Programme (WFP) working population figure for Baringo district is 265,241. The more mountainous areas are more densely populated.
Baringo is one of the nineteen (19) districts in Rift Valley Province. It borders Turkana, Samburu and Keiyo, Koibater, Marakwet and West Pokot. (See Map) The district is one of the nineteen districts requiring relief intervention (WFP EMOP).
The district can be divided into 2 major zones; the highlands and the lowlands. The highland areas are the areas of the district that receive relatively above average annual rainfall compared to the lowlands and are assumed to be less prone to drought. The highland has some cultivable land, which relies on rain fed agriculture.
There are 2 rainy seasons in this district, the long rains in March-July and the short rain in September-November.
2.2 General situation and the impact of the drought
Over the past 3 years, drought conditions in the arid regions of the Horn of Africa have steadily increased. The impact of this drought and the failure of 3 consecutive harvests in Kenya have been severe and more pronounced in the arid and semi-arid areas of the country, which are in the Eastern, North Eastern and Rift Valley provinces. This has created extreme food insecurity and escalating rates of malnutrition. Another factor that has contributed to food insecurity has been the recent long rains being late, poorly distributed and erratic.
It is generally considered that both the highland and lowland areas within Baringo have been affected by the drought but in different ways. The following are the main consequences of the drought conditions in the lowland and highland areas in Baringo;
- Poor to no harvests during the last 3 years.
- Reduced seed stocks for farmers.
- Purchasing power of many families has been seriously eroded and many are relying mainly on food aid assistance.
- Silting up and drying up of dams as well as a change in river flow direction since the heavy El Nino rains has meant reduced possibilities for irrigation.
- Inadequate natural regeneration of pasture.
- Major livestock losses (cattle, goats and sheep) among the pastoralist communities due to lack of pasture, lack of water and East Coast Fever.
- Animal prices dropped dramatically so farmers received poor financial returns when they were eventually forced to sell their animals due to insufficient pasture and water. This situation was aggravated by the remoteness of some of the areas and their lack of access to trade and fair market prices. (However, the cost of meat at butcher shops has remained the same throughout.)
It is interesting to note that the majority of the camels in Baringo have withstood these drought conditions.
The future also looks bleak as the Meteorological Department has further warned that the rains expected in October 2000 will be poor and below average in the arid and semi-arid regions of the country. There is low expectation that the existing situation will not significantly improve until the middle of next year assuming the long rains are adequate and the harvest yield will be good.
2.3 Justification for survey
In June a small sampling (240) survey was carried out in Baringo district. The global acute malnutrition (<2 z scores and / or oedema) was 33.3% and severe acute malnutrition (<-3 and/or oedema) was 8%. The methodology has obvious limitations necessitating a repeat survey using a more representative sampling frame.
A 30 cluster sampling survey has been carried out in the lowlands in the beginning of October. The lowlands was chosen first before the highlands because of the intensity of the drought there. (The report of lowlands survey is prepared separately.) To complete the picture of the district the highland survey is hereby carried out.
2.4 World Vision in Baringo district
World Vision started work in Baringo in 1984 as a response to the drought. Following the crisis World Vision initiated long-term development programmes. For this reason the organisation has a good understanding of the situation in the area which has proved invaluable during the emergency response.
World Vision is currently co-ordinating the food distributions in Baringo (affected population 129,717). The partner agencies for World Vision in Baringo district are; Full Gospel, AIC, and the Catholic church.
In addition to the general food distributions, World Vision started in September a blanket supplementary food distribution to 25,000 vulnerable population (children <5, lactating mothers and pregnant women).
2.5 General Food Distribution
The WFP and GOK are the main food donors in Baringo, targeting up to 49% of the population. Due to problems of insufficient quantities of food in the WFP pipeline, World Vision has not been able to distribute the full 100% ration during the last few months. The monthly ration distributed provides only up to 70% of the daily caloric requirement.
There are now 231 distribution centres throughout Baringo district. Every distribution centre has a relief committee, which represents the population receiving food aid. It is the responsibility of the relief committee to ensure the food is distributed among the most vulnerable members of their population.
3.0 OBJECTIVES OF THE SURVEY
3.1)To assess the prevalence of malnutrition among children between 65CMs - 115CMs (equivalent to 6-59 months) who are living in the highland areas of Baringo.
3.2)To help to determine if targeted feeding programmes are necessary in Baringo district.
4.0 METHODOLOGY
4.1 Area surveyed map
The district was divided into 2 areas; the lowlands (the arid areas) and the highlands (the wetter areas). This survey was conducted in the highlands. Every sub-location in the district was assigned to either the highland or the lowland group. The rationale for dividing the district into highland and lowland area was so that it would be possible to determine the impact of the drought conditions on nutritional status of the 2 different communities (pastoralist and agriculturists).
4.2 Survey design
The nutrition survey was conducted according to the standard two-stage 30 cluster sampling method recommended by WFP/UNICEF/WHO. All the sub-locations in the highland areas were listed with their relevant under 5-year population. From the sub-locations 30 clusters were selected at random using the appropriate calculated sampling interval. Every sub-location has equal opportunity (chance) of being selected as a cluster for survey. The location of the first cluster was selected from within the first sampling interval using a random number table.
Every survey team went to the geographical centre (or as near as possible) of the selected sub-location. A pen was thrown, the direction of the pen indicates the direction the survey team should select their first household. A random number was selected between 1-5. For example, if 3 was selected the 3rd house in the direction of the pen was chosen as the first house to be surveyed. After the first household had been surveyed the team continued to survey the next nearest household on the right hand side. The team continued until sufficient children for the cluster had been surveyed.
All eligible children in every household were surveyed and the (UNICEF designed) questionnaire was completed for every child. If in the last household there were more eligible children than required to complete the cluster all eligible children would be surveyed.
If a child was absent while visiting a household the survey team tried to follow up the child at the end of the day.
If a team reached the sub-location boundary before completing the cluster if feasible they would return to the centre of the cluster and repeat the exercise as mentioned above. If it was not feasible to return to the centre, the pen was thrown at the boundary to determine the new direction to select the households to survey.
A household was defined as residents who all eat from the same cooking pot.
Children between 65cm and 115 cm (equivalent of 6-59 months) were included in the survey. All eligible children (except those with oedema) had their weight, height and MUAC measured. Weight was measured to the nearest 100g using Salter hanging scales. The scale was calibrated beginning of day and every weight measurement using a known weight (1 kilogram of rice). Children were weighed without clothes. Height was measured to the nearest 0.1 cm using the standard height/length board (children less than 2 years old were measured lying down while those above 2 years was measured standing). Age was recorded using data from the road to health card, and/or mother’s information with reference to local calendar of events, e.g., General election in December 1999, Cattle raid in April 1999 and El Nino rain in June 1997 till February 1998. MUAC was recorded to the nearest 2mm using the MUAC tape. Oedema was assessed by applying normal thumb pressure for 15-30 seconds on both feet (near the ankle). If an indentation or pit remained on both feet when the thumbs were lifted then the child was assumed to have oedema. Information on age, sex, immunisation, health, general food rations and supplementary food was collected. (See questionnaire)
4.3 Sample size
To calculate the sample size the team assumed the prevalence rate of malnutrition to be 20%. This assumption was based on the results of the survey in June and the fact that populations in the highlands are assumed to have better nutritional status compared with those from the lowlands. It was assumed 1.96 for an error of risk of 5% and the absolute precision required was 4% (0.04). The number of children required to be surveyed was therefore 384 and to take into consideration the design effect for cluster surveying the sample was multiplied by 2. Therefore 768 children was the required sample size, this meant 26 children per cluster.
4.4 Nutritional indicators
Individual nutritional status was assessed using the weight for height index (W/H) and results were expressed in Z scores and in per centage of the median of the reference population (NCHS/CDC/WHO reference table).
Global acute malnutrition was defined as W/H below 2 Z scores and/or the presence of oedema, and severe acute malnutrition as a W/H below – 3 Z scores and/or oedema. In % of the median, global acute malnutrition was defined to be W/H less than 80% and/or oedema, and severe acute malnutrition was a W/H below 70% and/or oedema.
MUAC’s of less than 125mm were considered as global acute malnutrition and less than 110mm as severe acute malnutrition.
4.5 Data collection and analysis
The survey was conducted during the third week of October by 7 survey teams with 3-4 people per team plus a guide/translator from the area. Data was collected using a standardised questionnaire (see annex). The data was entered using EPI info software (v 6.0). Anthropometric indices were calculated with Epinut and analysis was completed using Epi info.
5.0 RESULTS
5.1 Nutrition
Table 1 Proportion of malnutrition according to different anthropometric indices
ANTHROPOMETRIC INDICES / NUMBER OF CHILDREN(sample size 755) / PROPORTION
(%) / 95% Confidence
Intervals
Z – SCORES
<-2 z-scores + oedema (global acute) / 61 / 8.1% / (6.2-10)
>=-3 -<-2 (moderate) / 38 / 5.0% / (3.4-6.6)
<-3 z scores + oedema (severe acute) / 23 / 3.1% / (1.8-4.2)
% OF THE MEDIAN
< 80% + oedema (global acute) / 44 / 5.8% / (4.1-7.5)
>= 70% - < 80% (moderate) / 24 / 3.2% / (1.9-4.5)
< 70% + oedema
(severe acute) / 20 / 2.6% / (1.5-3.7)
MUAC (mm)
< 125 mm (global) / 44 / 0.1% / ()
110-124 mm (moderate) / 43 / 5.7 / (4.12-7.48)
< 110 mm (severe) / 1 / 5.8% / (4.21-7.49)
Oedema cases / 18 / 2.4%
Twenty seven % of the oedema cases were identified in 1 sub-location namely Kaimogul in Kabartonjo division. The results seem to indicate that the classification of the severe acute was generous particularly the definition of oedema giving rise to potential bias and narrowing of the gap between severe acute and global acute. There was no way of correcting the numbers as households were not numbered.
5.2 Health
Ownership of health cards
The survey teams commented that several households had health cards for their children, but the health cards were blank except for the name of the child. Obviously, such a health card provides no information concerning the health and vaccination status of the child.
Table 2 Presence of health cards
HEALTH CARDS
/ No. of children / Proportion of childrena) Yes / 703 / 93.2%
b) No / 51 / 6.8%
TOTAL / 755 / 100%
Measles vaccination coverage
3 weeks prior to the nutrition survey a measles vaccination campaign took place throughout Baringo district. During this campaign it was not indicated on children's health cards whether they had been vaccinated. However, the general perception based on anecdotal information the measles campaign had been successful.
The table below indicates the measles vaccination coverage for all children interviewed.
Table 3 Measles vaccination status
MEASLES / No of children / % of children / Year / No of children / % of children2000 / 569 / 82.8%
a) Has received measles vaccination / 684 / 97.3% / 1999 / 38 / 5.5%
b) Has not received measles vaccination / 19 / 2.7% / Earlier / 80 / 11.6%
TOTAL
/ 703 / 100% / TOTAL / 687 / 100%Note: there were only 687 questionnaires that had indicated the year which the measles vaccination was received.
Polio
Regular polio vaccination campaigns are conducted throughout Baringo district as part of the Kenya polio eradication programme. During the polio vaccination campaigns it is not indicated on children's health cards that they have received the polio vaccination. The table below indicates the polio vaccination coverage for all children irrespective of whether they had a card or not.
A polio campaign is planned throughout the country, including Baringo district on the third week of October.
Table 4 Polio vaccination status.
POLIO / No of children / % of children / Year / No of children / % of children2000 / 111 / 15.9%
a) Has received polio vaccination / 657 / 93.5% / 1999 / 418 / 60%
b) Has not received polio vaccination / 46 / 6.5% / Earlier / 168 / 24.1%
TOTAL
/ 703 / 100% / TOTAL / 697 / 100%Note: there were 697 questionnaires that had indicated the year which the polio vaccination was received.
Vitamin A
During the recent measles campaign vitamin A was also administered to all children who received the measles vaccination. This explains why a large proportion (94.4%) had received vitamin A during the previous 3 months. Vitamin A data was collected from cards and mothers’ recall. The table below indicates that a high proportion (94.7%) had received vitamin A.
Table 5 Vitamin A coverage
VITAMIN A / No of children / % of children / months / No of children / % of children> 6 m / 30 / 4.2%
a) Has received vitamin A / 711 / 94.7% / 4–6 m / 10 / 1.4%
b) Has not received vitamin A / 40 / 5.3% / 1-3m / 678 / 94.4%
TOTAL
/ 751 / 100% / TOTAL / 718 / 100%Note: there were 718 questionnaires that had indicated the period which the vitamin A was received.
Morbidity
The most common illness reported during the last 2 weeks was cough/cold 41.5%, followed by a fever–like illness, 10.9. The most common illnesses in the "other" category were skin rashes, ringworms, scabies, stomach ache, otitis, and conjunctivitis. The morbidity patterns appeared to vary throughout the area; i.e. some areas had higher morbidity levels compared to other areas.
Table 6 Morbidity patterns during the previous two weeks
MORBIDITY /YES
/NO
/ Total no of respondentsNumber
/ % /Number
/ %a) Diarrhoea / 65 / 8.7% / 686 / 91.3% / 751
b) Cough/cold / 310 / 41.5% / 437 / 58.5% / 747
c) Fever-like illness / 82 / 10.9% / 672 / 89.1% / 754
d) Measles / 6 / 0.8% / 747 / 99.2% / 753
e) Other / 57 / 7.5% / 698 / 92.5% / 755
5.3 Feeding practices
The proportion of children currently breast-feeding was low (28.1%) see table 8, however this is not surprising considering the youngest children to be included in the sample were 6 months. The introduction of foods other than breast milk among the surveyed sample occurs at a young age, for example 80.8% of children are introduced to additional food sources less than 3 months of age. (See table 9)
Table 7: Proportion of children surveyed who were currently breast-feeding
BREAST FEEDING
/ No of children / Proportion of childrenCurrently breastfed
/ 210 / 28.1%Not being breast fed / 537 / 71.9%
TOTAL / 747 / 100%
Table 8: Period of exclusive breast-feeding
Period of exclusive breast feeding /No. of children
/ Proportion of childrena) Up to 1 month / 373 / 50.9%
b) 2 – 3 months / 219 / 29.9%
c) 4 – 6 months / 129 / 17.6%
d) 7 – 9 months / 6 / 0.8%
e) over 9 months / 6 / 0.8%
TOTAL / 733 / 100%
5.4 General food rations
In Baringo district 49% of the population are targeted to receive food aid. However, from the survey sample 69.4% (see table 10) of the population were accessing the food aid. This is possible because there is internal redistribution of food at the household and community level. The malnutrition rate in the highlands is below 10%, at this rate it does not require a food aid intervention.