UNICEF KENYA COUNTRY OFFICE

SUMMARY TABLE ON NUTRITION SURVEYS

13 December, 2000

Please find below a table summarizing information on nutrition surveys that has been made available to UNICEF Health and Nutrition – Emergency Unit. The table indicates the methodology for survey sampling. The Health and Nutrition sub-unit is striving towards standardization of methodology within and across districts. To this end there is a guide for training on nutrition surveys available at UNICEF as well as a health and nutrition questionnaire which can be used by any interested agency.

Please note that this table is complied from data available currently to UNICEF - we would welcome comments and inputs from partners in updating the information herein.

DISTRICT / DATE / ORGANIZA-TION / LOCATIONS / METHOD / NO. SURVEY-ED / INDICATOR / GLOBAL MALNUTRITION RATE / SEVERE MALN. RATE
Turkana / March 2000 / Oxfam / Lokituang sub-district / 2 stage 30x30 cluster sampling / 900 / wt/ht / 21.3 (z- score) / 4.6
June 2000 / ANP/UNICEF / Kerio[1] / 10 clusters of 30 households to yield 300 households per division; criteria for inclusion of a household – at least 1 child 6-59 months in the household / 422 / wt/ht
MUAC / 37.2 (z-score)
27.9 / 8.5
August 2000 / UNICEF and 6 NGOs / Lokichokio
Kakuma & Oropoi / Two-stage cluster sampling; 25% prevalence of malnutrition assumed; eroor risk 5%, precision 5% sample size required 576; 30 clusters with 20 children in each / 599 / wt/ht / 27.13 (Z-score) / 8.32
Lokituang sub-district / As above / 597 / Wt/ht / 31.37 / 6.06
Katilu, Kainuk / As above / 593 / Wt/ht / 30.50 / 9.26
Central, Kalokol / As above / 596 / Wt/ht / 24.88 / 4.37
Lokichar, Lokori & Lomello / As above / 598 / Wt/ht / 34.46 / 8.36
Loima, Turkwell[2] / As above / 498 / Wt/ht / 46.80 / 19.47
Marsabit / June 2000 / UNICEF/ANP / Loyengalani[3] / 10 clusters of 30 households to yield 300 households per division; criteria for inclusion of a household – at least 1 child 6-59 months in the household / 358 / Wt/ht
Muac / 35.2 (z-score)
35.8 / 6.7
June 2000 / Medair / North Horr / 2 stage 30x30 cluster sampling / 924 / wt/ht / 43.5 (z-score) / 6.2
July 2000 / Tearfund / Maikona central and Kalacha / Central division divided into 6 zones + Kalacha children assembled in central place and assessed; 1-day assessment / 836 / wt/ht / 28.2 (z-score) / 3.4
July/Aug / MoH / Laisamis / 310 / wt/ht / 40.0 / ?
Oct. 2000 / Medair / North Horr / 2 stage 30x30 cluster sampling / 924 / Wt/ht / 30.7 (z-score) / 4.9
Oct. 2000 / Medair / Loyengalani / 2 stage 30x30 cluster sampling / 905 / Wt/ht / 29.0 / 3.8
Nov. 2000 /

Maikona

/ Tearfund / 2 stage 30x30 cluster sampling / 900 / Wt/ht (z-scores) / 25.1 / 3.9
Nov. 2000 /

FHI/MoH

/ Laisamis / Planned survey
Moyale / June 2000 / UNICEF/ANP / Obbu[4] / 10 clusters of 30 households to yield 300 households per division; criteria for inclusion of a household – at least 1 child 6-59 months in the household / 374 / wt/ht
muac / 18.7 (z-score)
14.2 / 5.8
July 2000 / Tearfund / Sololo / 350 / wt/ht / 17.1 / 2.3
Nov.
2000 / UNICEF/MoH/
Worl d Vision / Moyale / Two-stage cluster sampling; 20% prevalence of malnutrition assumed; precision 5%, error risk 5%; sample size required 768; 30 clusters with 32 sampled to account for data rejection during analysis / 958 / Wt/ht (z-scores) / 8.1 / 1.5
Mandera / Dec 1999 / MSF/S / Mandera town and outskirts / 912 / Wt/ht / 27.0 (z-score) / 2.1
June 2000 / MSF/S / Mandera town and outskirts / 918 / Wt/ht / 20.3 (z-score) / 2.2
June 2000 / UNICEF/ANP / Shimbre Fatuma[5] / 10 clusters of 30 households to yield 300 households per division; criteria for inclusion of a household – at least 1 child 6-59 months in the household / 539 / Wt/ht
Muac / 29.3 (z-score)
23.7 / 8.6
August 2000 / MDM/F / All divisions listed below – overall results / 2 stage cluster sampling for the five divisions combined; prevalence of malnutriton assumed 21.7%; precision 4%; however data also analyzed per division / 777 / Wt/ht / 26.1 (Z-score) / 7.1
Rhamu / See above / 251 / 29.1 / 8.8
Rhamu-Dimtu / See above / 49 / 38.8 / 8.2
El Wak / See above / 268 / 22.3 / 5.6
Ashabito / See above / 110 / 23.6 / 7.3
Lafey / See above / 98 / 25.5 / 6.1
Sept.
2000 / UNICEF, MoH and partners / Banisa / Two-stage cluster sampling; 25% prevalence of malnutrition assumed; precision 4%, error risk 5% ; sample size required 576; 30 clusters with 20 children in each / 606 / Wt/ht[6] / 26.8 (z-score) / 4.6[7]
Takaba / As above / 613 / 28.3 (z-score) / 4.7
Malkamari / As above / 597 / 19.93 (z-score) / 10.39
Fino / As above / 557 / 17.2 (z-score) / 3.5
Warankara / As above / 627 / 39.6 (z-score) / 12.1
Wajir / Dec. 1999 / Oxfam / Wajir town / Two stage 30x30 cluster sampling / 917 / wt/ht / 20.0 (z-score) / 3.5
July 1999 / UNICEF/ANP / Griftu[8] / 10 clusters of 30 households to yield 300 households per division; criteria for inclusion of a household – at least 1 child 6-59 months in the household / 459 / Wt/ht
Muac / 32.6 (z-score)
19.7 / 7.6
Sept.
2000 / SCF (UK) / Wajir central / Two stage cluster sampling; malnutrition prevalence assumed at 24%; 3% precision; 5% error risk; required sample 1560 [9] / 1,553 / Wt/ht / 22.5 (z-score) / 5.0
Oct.
2000 / North/West / Two-stage 30X30 cluster sampling / 746[10] / Wt/ht / 21.2 (z-score) / 5.8
Garissa / July 2000 / UNICEF/ANP / Mogadashe[11] / 10 clusters of 30 households to yield 300 households per division; criteria for inclusion of a household – at least 1 child 6-59 months in the household / 410 / Wt/ht / 27.2 (z-score) / 6.6
Sept.
2000 / CARE/
MoH and partners / All Garissa –combined results / 2 stage cluster sampling;expected prevalence 27%; precision 5%; 602 children required from each zone / 1675[12] / Wt/ht (z-scores) / 21.1 / 8.4
Northern / As above / 465 / Wt/ht / 22.7 (z-score) / 9.3
Central / As above / 427 / 20.7 (z-score) / 10.0
Southern (Ijara) / As above / 420 / 20.8 (z-score) / 7.5
Baringo / July 2000 / World Vision / 6 divisions chosen to represent all agro-ecological zones / 240 / wt/ht / 33.3 / 8.0
Sept.
2000 / WV / Lowland dry areas / 1011 / wt/ht / 23.4 (z-score) / 5.4
Oct.
2000 / WV / Highland areas / Two stage cluster sampling; expected prevalence 20%, error risk 5%, precision 4%; 30clusters with 26 children – sample size 768 - needed / 755 / Wt/ht (z-score) / 8.1 / 3.1
Samburu / August 2000 / MSF/F / Baragoi / Rapid MUAC assessment / 4,097 / MUAC / 21.2 / 1.7
El Nyiro / As above / 1,768 / MUAC / 9.8 / 0.7
Oct/Nov 2000 / IMC / MoH and partners / Baragoi and El Nyiro / Two stage 30x30 cluster sampling / 909 / Wt/ht (z-scores) / 18.3 / 1.9
Lokorri and Kirissia / As above / 909 / Wt/ht (z-scores) / 27.6 / 3.7
Wamba and Waso / As above / 913 / Wt/ht (z-scores) / 24.2 / 3
Laikipia / Nov. 2000 / World Vision / Mukodogo and Central divisions[13] / Cluster sampling method; expected prevalence 25%, error risk 5%, precision 4%; 900 children required; 27 clusters with 24 children in each surveyed / 903 / Wt/ht (z-scores) / 20.3 / 6.8
Makueni / July 2000 / AMREF / Greater Kibwezi area / ? / ? / wt/age / 21.1
Kajiado / October 2000 / OCI and partners, AMREF (UNICEF) / District-wide / Two stage 30x30 cluster sampling method / 911 / Wt/ht (z-scores) / 19.6 / 7.9

Msz – kenyasurveys 12/13/00 6:08 PM

1

[1] This division was chosen for the survey as it was considered, by the District Steering Group, to be the worst-affected.

[2] The malnutrition levels are extremely high. It is to be noted that the survey was carried out soon after a measles outbreak and there were an unusual number of sick and malnourished children there at that point.

[3] This division was chosen as it was considered, by the District Steering Group, to be the worst affected

[4] This division was chosen as it was considered, by the district steering group, to be the worst affected.

[5] This division was chosen as it was considered, by the District Steering Group, to be the worst affected.

[6] In general the malnutrition rates data from Mandera are very high. It is possible that genetically the Somali populations are leaner and thus comparing them to the NCHS reference standards may not give the ‘true’ picture of malnutrition in these areas. There may an over-estimation of malnutrition. However, the data from the CBS/UNICEF 5th nutrition survey in 1994 found the GAM (<-2 z-scores) in Mandera to be 27.7%

[7] Oedema data was left out of the analysis as it appears that this was not collected correctly according to the criteria explained in the training session. High numbers of oedema were reported. However in a re-assessment in October by MSF/Spain in Malkamari area it was also seen that although the global malnutrition levels appeared not that high there were relatively high numbers of oedema cases. This has prompted MSF to open a therapeutic feeding centre in Rhamu to serve the Malkamari area as well.

[8] This division was chosen as it was considered, by the District Steering Group, to be the worst affected.

[9] A total of 25 clusters were completed. Due to insecurity the team could not collect additional data.

[10] The survey could not be completed due to insecurity. Data from 25 clusters were collected.

[11] This division was chosen as it was considered, by the District Steering Group, to be the worst affected.

[12] Data collection in some parts of Dadaab, Hulugho and Jarajilla was not possible due to inaccessibility, insecurity and movement of nomads. As a result data for 104 children could not be collected. Data was collected on 1702 as opposed to be required 1806 children; analysis included 1675 children.

[13] These divisions were chosen as they are the areas where World Vision has developmental programmes and are distributing general food ration (EMOP).