2007Kenya

MALARIAIndicator

Survey

Division of Malaria Control,Ministry of Public Healthand Sanitation

Kenya National Bureau ofStatistics

National CoordinatingAgency for Population andDevelopment

Republic of Kenya

KMIS 2007

i

2007 KenyaMALARIA

Indicator SurveyDivision of Malaria Control,

Ministry of Public Health and Sanitation

Kenya National Bureau of Statistics

National Coordinating Agency for Population and Development

March 2009

Republic of Kenya

i i

KMIS 2007

THIS REPORT SUMMARIZES the findings of the 2007 Kenya Malaria Indicator Survey (KMIS) carriedout by the Division of Malaria Control in the Ministry of Public Health and Sanitation in partnershipwith the Kenya National Bureau of Statistics. The Department for International Development (DFID)through the World Health Organization (WHO) provided financial assistance for the survey. Technicalsupport was provided by WHO and Centers for Disease Control and Prevention (CDC). The opinionsexpressed in this report are those of the authors and do not necessarily reflect the views of thedonor organizations.

Any part of this document may be freely reviewed, quoted, reproduced or translated in full or inpart, provided the source is acknowledged. It may not be sold or used in conjunction withcommercial purposes or for profit.

Information about the survey may be obtained from the Division of Malaria Control in the Ministryof Public Health and Sanitation, PO Box 19982 KNH, Nairobi 00202, Kenya (Telephone: +254 202716934/5; Fax: +254 20 2716935; email: head.nmcp@domckenya.or.ke).

Additional information about the KMIS may be obtained from Kenya National Bureau of Statistics(KNBS), PO Box 30266 GPO, Nairobi 00100, Kenya (Telephone: +254 20 340929/317583; Fax: +25420 317559; email: director@knbs.go.ke).

2007 Kenya Malaria Indicator Survey

Published by: Division of Malaria ControlMinistry of Public Health and SanitationPO Box 19982 KNHNairobi 00202, Kenyahead.nmcp@domckenya.or.ke

WORLD HEALTH ORGANIZATION

KMIS 2007

i i i

Contents

List of Tables and Figures vList of Acronyms and Abbreviations viList of Authors viiMap of Kenya – Provincial Boundaries viiiForeword ixExecutive Summary xi

Chapter 1 Introduction 11.1 Kenya Country Profile 1

1.1.1 History 21.1.2 Economy 21.1.3 Population 3

1.2 Health Priorities and Programmes 41.3 The National Malaria Strategy and

Policy Issues 2001–2010 41.4 Malaria Endemicity in Kenya 5

Chapter 2 Survey Organization andMethodology 7

2.1 Survey Organization 72.2 Sample Design 82.3 Training 82.4 Fieldwork 82.5 Questionnaires 102.6 Response Rates 102.7 Data Processing 10

2.8 Ethical Considerations 112.9 Weighting the Sample 11

Chapter 3 Household Population andHousing Characteristics 12

3.1 Household Population 123.2 Household Composition 133.3 Household Headship 133.4 Demographic Characteristics of Women 143.5 Housing Characteristics 14

3.5.1 Electricity 143.5.2 Cooking fuel 143.5.3 Drinking Water 153.5.4 Sanitation 153.5.5 Floor 15

3.6 Household Possessions 16

Chapter 4 Vector Control 174.1 Household Ownership of Mosquito Nets 17

4.1.1 Programme Efforts to IncreaseNet Ownership 18

4.1.2 Prevalence of Net Ownership 184.2 Use of Mosquito Nets by Children

under Five Years of Age 194.3 Use of Nets by Women of Reproductive

Age and Pregnant Women 20

i v

KMIS 2007Chapter 5 Case Management 215.1 Management of Childhood Fevers 21

5.1.1 Prevalence and Prompt Treat-ment of Fever 21

5.1.2 Care Seeking Behaviour forChildren with Fever 23

5.1.3 Management of Children withReported Fevers 23

5.1.4 Treatment Options 245.2 Sources of Malaria Drugs 255.3 Malaria in Pregnancy 25

5.3.1 ANC Attendance 265.3.2 Use of Any Antimalaria Drugs by

Pregnant Women 265.3.3 Use of Intermittent Preventive

Treatment of Malaria in Pregnancy 265.4 Information, Education and

Communication of ACT 27

Chapter 6 Parasitaemia and AnaemiaPrevalence in Children 29

6.1 Parasitaemia Prevalence in Children 296.2 Anaemia Prevalence in Children 30

Chapter 7 Policy Implications,Conclusions andRecommendations 31

7.1 Objectives of the Survey 317.2 Core Malaria Interventions Included

in 2001 NMS 317.2.1 Net Ownership by Households 317.2.2 Use of Mosquito Nets by Children

under Five Years 327.2.3 Use of Nets by Women of Repro-

ductive Age and Pregnant Women 32

7.3 Management of Childhood Fevers 327.3.1 Prevalence and Prompt Treat-

ment of Fever 327.3.2 Management of Children with

Reported Fevers 337.3.3 Sources of Malaria Drugs 337.3.4 Treatment Seeking Areas 34

7.4 Malaria in Pregnancy 347.5 Information, Education and

Communication 357.6 Parasitaemia Prevalence in Children

under Five 357.7 Anaemia Prevalence in Children 357.8 Conclusions and Recommendations 35

References 37

AppendixesA National Malaria Strategy 2006 Targets

and Goals 39B Sample Design for 2007 KMIS 41C Persons Involved in the Survey 45D Survey Questionnaires 47

KMIS 2007

v

Tables and Figures

List of Tables

2.1: Response rates for the householdand individual interviews 10

3.1: Household population by age, sexand residence 13

3.2: Household composition 133.3: Per cent distribution by sex and

education of household head 123.4: Demographic characteristics of

female respondents 143.5: Housing characteristics 153.6: Household possessions 164.1: Ownership of mosquito nets 184.2: Use of mosquito nets by children 194.3: Use of nets by women of reproduc-

tive age and pregnant women 205.1: Prevalence and prompt treatment of

fever 225.2: Standard treatment of children

with fever 235.3: Treatment options 245.4: Sources of malaria drugs 255.5: Use of antimalaria drugs among

pregnant women 26

5.6: Information, education andcommunication on ACT 28

6.1: Prevalence of malaria parasites inchildren 30

6.3: Level of anaemia in children aged6–59 months 30

6.2: Predominant malaria parasites 30B2: Allocation of the sample by

province, district and urban-rural 42B1: Allocation of the sample by province

and urban-rural 42

List of Figures

1.1: Malaria zones in Kenya 63.1: Population pyramid 123.2: Women respondents by age 145.1: Management of children with

reported fevers 225.2: Care seeking behaviour for children

with fever 235.3: Use of intermittent preventive

treatment of malaria in pregnancy 275.4: Use of ANC services by wealth

quintiles 27

v i

KMIS 2007

Acronyms and Abbreviations

ANC Antenatal clinicACT Artemesinin combination treatmentCDC Centers for Disease Control and

PreventionCHW Community health workerDFID Department for International Devel-

opmentDOMC Division of Malaria ControlDRH Division of Reproductive HealthDSO District Statistical OfficerEPSEM Equal probability selection methodERSWEC Economic Recovery Strategy for

Wealth CreationGDP Gross domestic productGIS Geographical information systemGPS Geographic positioning systemIPT Intermittent preventive treatmentIRS Indoor residual sprayingITN Insecticide treated (bed) netIUGR Intra-uterine growth retardationKEMRI Kenya Medical Research InstituteKMIS Kenya Malaria Indicator SurveyKNBS Kenya National Bureau of StatisticsLLITN Long-lasting insecticide treated (bed)

net

MCH Maternal/child healthNMS National Malaria Strategy 2001–2010MOH Ministry of HealthMSH Management Sciences for HealthNASSEP National Sample Survey and EvaluationNCAPD National Coordinating Agency for

Population and DevelopmentNHSSP National Health Sector Strategic PlanNMCP National Malaria Control ProgrammePDA Personal digital assistantPPMOS Probability proportional to measure of

sizePMI Presidential Malaria InitiativePSI Population Services InternationalPPS Probability proportional to sizeRDT Rapid diagnostic testSAS Statistical Analysis SystemSP Sulphadoxine pyrethamineSPSS Statistical Package for Social ScientistsTWG Technical Working GroupVIP Ventilated improved pit latrineWHO World Health OrganizationWRP Walter Reed Project

KMIS 2007

v i i

Authors

Abdulkadir A. Awes - Kenya National Bureau of StatisticsRobert C.B. Buluma - Kenya National Bureau of StatisticsJohn Bore - Kenya National Bureau of StatisticsPaul Waweru - Kenya National Bureau of StatisticsFrancis Kundu - National Coordinating Agency for Population and DevelopmentDr. Rebecca Kiptui – Division of Malaria ControlDr. Elizabeth Juma – Division of Malaria ControlDr. Kiambo Njagi - Division of Malaria ControlDr. Evans Mathenge - Kenya Medical Research InstituteJames Akudian - Division of Malaria ControlDr. Dorothy Memusi - Division of Malaria ControlDr. Andrew Nyandigisi - Division of Malaria ControlJohn Olouch - Division of Malaria ControlPeter Njiru - Division of Malaria ControlPaul Kiptoo - Division of Malaria ControlJacob Kimani - Division of Malaria ControlAllen Hightower – Kenya Medical Research Institute/Centers for Disease Control and PreventionPeter Ouma - Kenya Medical Research Institute/Centers for Disease Control and PreventionMary Hamel - Kenya Medical Research Institute/Centers for Disease Control and PreventionAbdinasir Amin - Management Sciences for HealthDr. Augustine Ngindu - World Health OrganizationManya Andrews - Population Services InternationalAnthony Kanja - Population Services InternationalDaniel Wachira - Presidential Malaria Initiative

viii

KMIS 2007

Map of Kenya – Provincial Boundaries

KMIS 2007

i x

Malaria is among the diseases of majorpublic health significance. Besidesits toll of human suffering anddeath, malaria saps the work forceand drains the economy. Reducing, eveneliminating, its adverse impact is therefore ofgreat humanitarian and economic importance.African Heads of State, meeting in Abuja, Nigeria,in April 2000, issued a declaration committingcountries including Kenya to achieve significantcountry targets for malaria control. At the time,malaria accounted for up to 30 per cent of alloutpatient attendance in Kenya and 19 per centof the admissions to health facilities. TheDivision of Malaria Control (DOMC) is thereforemandated to reduce the level of malaria infec-tion and consequent death in Kenya.

To address the severe burden of malaria inKenya and to be in harmony with internationalRoll Back Malaria targets, the Government ofKenya in collaboration with partners developedKenya’s National Malaria Strategy (NMS) 2001–2010, which was launched in April 2001. TheNMS articulates the key strategic approachesto malaria control and targets in concert with

Foreword

those established by the Roll Back Malariamovement in Abuja.

The NMS is implemented using four keystrategic approaches: to guarantee all peopleaccess to quick and effective treatment, tosignificantly reduce illness and death frommalaria, to provide malaria preventionmeasures and treatment to pregnant women,and to promote the use of insecticide-treatednets by at-risk communities to significantlyreduce rates of infection and mitigate againstthe effects of malaria epidemics.

This document reports the results of the2007 Kenya Malaria Indicator Survey(KMIS), the first such survey to be under-taken by the country. At the time the surveywas done, the NMS had been implemented forsix years. The survey intended to evaluateprogress the country has made towards theAbuja targets across the interventions imple-mented by the programme.

This report details parasitaemia levels inchildren; access to the first line antimalarial;use of insecticide treated nets by vulnerable

x

KMIS 2007groups; and use of intermittent preventivetreatment by pregnant women. The findings ofthe survey are expected to go a long way inassisting the DOMC to review its policies andstrategies for the future. They also highlightsome of the challenges in scaling up of inter-ventions.

Like any major survey, the KMIS requiredthe efforts of a variety of partners andstakeholders from national to communitylevels. We would like to acknowledge the inputof our partners that were involved in this survey:the Department for International Development(DFID) for funding the whole process, as wellas the World Health Organization (WHO) – boththe Kenya country office and the Africa RegionalOffice – Centers for Disease Control andPrevention (CDC), Population ServicesInternational (PSI), Kenya Medical ResearchInstitute (KEMRI)-Wellcome Trust, and theNational Coordinating Agency for Population andDevelopment (NCAPD) for their technical input.

Our gratitude also goes to all members ofthe Malaria Monitoring and Evaluation TechnicalWorking Group. Special thanks go to KenyaNational Bureau of Statistics (KNBS) whoprovided technical support and coordinated thefield data collection activities throughout theprocess. Abercrombie & Kent provided logisticsin the form of transport during the field work.We are also grateful to the field teams whoworked tirelessly and to all the staff of DOMC.

We also would like to appreciate all therespondents and those who gave blood samples,for this is the backbone of this report.

Dr. S.K. Sharif, OGW, MBchB, MMed, DLSHTM,MScActg Director of Public Health and SanitationMinistry of Public Health and Sanitation

KMIS 2007

x i

Up to 30 per cent of outpatient attend-ance and 19 per cent of admissionsto health facilities in Kenya are dueto malaria. To combat this disease theMinistry of Health through the National MalariaControl Programme (NMCP) developed theNational Malaria Strategy (NMS) covering theperiod 2001 to 2010. The main goal of the NMSwas to reduce the level of malaria infectionand consequent death by 30 per cent by theyear 2006 and to sustain that improved level ofcontrol to 2010. During the same period theGovernment also launched the second NationalHealth Sector Strategic Plan (NHSSP II) 2005–2010. The goal of the strategic plan is to reduce

Executive Summary

health inequalities and to reverse the downwardtrend in health-related outcome and impactindicators.

Furthermore, the Government developedKenya Vision 2030, whose goal for the healthsector is to provide equitable and affordablequality health services to all Kenyans. Vision 2030also aims at restructuring the health care deliverysystem to shift the emphasis from curative topromotive and preventive health care.

The Kenya Malaria Indicator Survey (KMIS)was conducted in 2007 with three mainobjectives: to collect up-to-date informationon coverage of the core malaria interventionsincluded in the NMS 2001–2010; to assess malariaparasite prevalence in children aged 1–59months; and to assess the status of anaemiaamong children aged 6–59 months.

Summary of Survey Results

The results of this survey indicate that 63 percent of households own at least one bed netand 34 per cent of households own more thanone net. For every 10 households there are 8

The Kenya Malaria Indicator Survey:w Collected up-to-date information on

coverage of the core malariainterventions included in the NationalMalaria Strategy 2001–2010.

w Assessed malaria parasite prevalence inchildren aged 1–59 months.

w Assessed the status of anaemia amongchildren aged 6–59 months.

x i i

KMIS 2007

insecticide treated nets (ITNs) or 12 nets of anykind.

The results further show that in endemicareas, 74 per cent of households own at leastone net while 39 per cent own more than one.In the same areas 58 per cent of all householdsown at least one ITN and 28 per cent own morethan one ITN. For households in epidemic-proneareas, 59 per cent own at least one net, while33 per cent own more than one. In the sameareas 48 per cent own at least one ITN and 24per cent own more than one ITN.

The data show that 51 per cent of childrenunder five years of age slept under any net thenight preceding the survey, but only 39 per centslept under an ITN. About 61 per cent of childrenunder five in urban areas used a mosquito netthe night preceding the survey, compared with50 per cent in rural areas. Similarly, 43 per centof children under five in urban areas slept underan ITN the night before the survey as comparedwith 39 per cent of rural children.

Endemic areas have the highest use ofmosquito nets (58 per cent) and ITNs (46 percent). Areas in which two mass net distributioncampaigns were conducted show a higher ITNusage among children under five than non-campaign ones. That is, in areas where therewas no campaign, only 27 per cent of childrenunder five had slept under an ITN the nightpreceding the survey whereas 44 per cent ofthose residing in campaign areas had done so.

The survey shows that pregnant women weremore likely than all women within thereproductive age to have slept under a mosquitonet: About 51 per cent of pregnant women sleptunder a net the night preceding the surveycompared with 45 per cent of all women ofreproductive age. Similarly, 40 per cent ofpregnant women had slept under an ITN

compared with 32 per cent of all women ofreproductive age.

One woman in two (50 per cent) in urbanareas slept under a net the night preceding thesurvey, compared with two in five rural women(43 per cent). Similarly, 42 per cent of pregnantwomen in urban areas slept under an ITNcompared with 39 per cent of their ruralcounterparts.

One child in three (32 per cent) had feverduring the fortnight preceding the survey. It wasobserved that fever prevalence in older children(26 per cent of four-year-olds) was lower thanthat in younger children (32 per cent of childrenbelow one year). Prevalence of fever washighest in malaria endemic areas (40 per cent)and lowest in epidemic and highland areas (25per cent).

The results further show that 24 per cent ofthe children with fever took an antimalaria drugand 15 per cent took antimalaria drugs thesame/next day. The results also indicate that70 per cent of children with fever soughttreatment from a facility/health provider thesame or next day. Endemic areas recorded thelowest percentage of patients who soughttreatment within 24 hours following the onsetof fever. But children living in epidemic-proneregions are more likely to seek prompttreatment and take an antimalaria medicine forfever. Only 22 per cent of children living inmalaria-endemic areas sought treatment fortheir fever

Government policy presently recommendsthat all fevers in children under five years ofage be presumptively treated as malaria withartemesinin combination treatment (ACT),which is provided free of charge at governmentand mission facilities. Perhaps surprisingly, theresults show that 35 per cent of the childrentook amodioquine, compared with 29 per centwho took ACT. Eleven per cent of febrile child-ren took SP and 8 per cent took chloroquine,

Of children with fever, 70 per cent ofchildren sought treatment from afacility/health provider the same or nextday. Children living in epidemic-proneregions are more likely to seek prompttreatment and take an antimalariamedicine for fever.

The National Malaria Strategy aimed to:w Guarantee all people access to quick and

effective treatment.w Significantly reduce illness and death

from malaria.w Provide malaria prevention measures and

treatment to pregnant women.w Promote the use of insecticide-treated

nets by at-risk communities.

KMIS 2007

xiii

which has not been recommended as a first linetreatment for malaria for ten years.Unfortunately, chloroquine was mostly used bythe youngest children (13 per cent) and thosein rural settings (9 per cent against 2 per centin urban areas).

About 25 per cent of children under one yeartook ACT drugs, compared with 35 per cent ofthose aged four years. The survey shows thatrural children under five years are moreadvantaged than their urban counterparts intreatment of malaria: Only 23 per cent of urbanchildren under five took ACT, whereas nearly30 per cent of rural children did so. Childrenunder five residing in malaria epidemic regionshad the lowest use of ACT, at 16 per cent,compared with the other zones at over 29 percent.

Most patients obtained their medicines fromeither public or private health facilities, with

47 per cent receiving the medicines from gov-ernment facilities. It was observed that even ingovernment health facilities, non-recommendedtherapies are still in use, including chloroquine.The fact that some patients received ACT andquinine from shops is of concern, since theseare medicines that are supposed to require aprescription from health workers.

The results show that a majority of children(59%) with fever in the two weeks before thesurvey went to public sector facilities fortreatment, compared with only 30 per cent whosought treatment from the private sector. Thoseliving in rural areas (61 per cent) were morelikely than urban residents (50 per cent) to seekcare from the public sector. Only 28 per cent ofrural residents took children with fever toprivate sector facilities, compared with 43 percent of urban residents. Use of retail shops as asource of treatment was low at 12 per cent.

Pregnant women are the second majortarget group for malaria interventions becausemalaria infection is associated with adverseeffects on the outcome of the pregnancy

(leading to miscarriage, stillbirth or low birthweight) and maternal morbidity (anaemia orsevere illness). Low birth weight (under 2,500grams) is associated with neonatal and infantmortality. Since malaria infection may eitherlead to clinical symptoms or be asymptomatic,intermittent preventive treatment for malariain pregnancy (IPTp) has been policy in Kenyasince 1998. The current recommendation is toprovide full treatment dosages of sulphadoxinepyrethamine (SP) beginning from quickening andrepeat them every four weeks until delivery.This policy is particularly important for areas

of high malaria (endemic)transmission in Kenya.

About 87 per cent ofmothers attended antenatalclinic (ANC) during their lastpregnancy leading to a live birthin the past two years. Moremothers in urban areas (93 per

cent) attended ANC than those in rural areas(85 per cent). The survey further shows thatlow-risk malaria zones reported the highest ANCattendance (92 per cent), while seasonaltransmission areas reported the lowestattendance (74 per cent).

The results indicate that 45 per cent ofmothers took any antimalaria medicines forprevention during pregnancy. Mothers in urbanareas (53 per cent) are more likely than thosein rural areas (43 per cent) to have done so.Similarly, mothers from low malaria transmissionzones were reported to have the highest use ofantimalarials (52 per cent) and mothers fromepidemic-prone zones the lowest (34 per cent).

The percentage of women who took any IPTpwas higher at 25 per cent than of those whotook IPTp2 (13 per cent). One urban mother inthree (33 per cent) took any dose of IPTpcompared with one in four mothers in rural areas(24 per cent). The data further show thatmothers living in low-risk areas had higher useof any dose of IPTp compared with those livingin epidemic-prone areas (33 per cent versus 16.7

More than half of pregnant women sleptunder a net the night preceding thesurvey compared with 45 per cent of allwomen within the reproductive age.

ANC attendance in all zones was much higher than IPTuptake, and women with secondary education were morelikely than those with primary education to attend ANC.

x i v

KMIS 2007

per cent, respectively). Similar trends weremaintained for IPT2, where low malariatransmission zones had the highest rates ofusage at 15 per cent, compared with 7.4 percent in epidemic-prone zones. ANC attendancein all zones was much higher than IPT uptake(87 per cent versus 25 per cent). Women withsecondary education were more likely thanthose with primary education to attend ANC (95per cent and 87 per cent, respectively).

In 2006, when the current treatment policywas launched, an intensive nationwideinformation dissemination campaign wasundertaken to increase awareness about ACT.The campaign included mass media, inter-personal communication and printed materials.Results of interviews with women aged 15–49years show that 39 per cent had heard aboutACT. Of these, 61 per cent received informationfrom radio, 27 per cent from health workersand 11 per cent from television.

Summary of Recommendations

w To meet the Abuja targets, explore morefree delivery options in order to better reachthe poorest wealth quintile more effec-tively, consider a mass net retreatmentcampaign, and invest in qualitative researchto better understand the drivers of net useand behaviour change outcomes.

w Shift the focus of ITN distribution fromvulnerable populations to universalcoverage of 100 per cent (a minimum of twonets per household) and usage to 80 percent.

w Increase access to ACT and SP for IPTp atthe community level by increasing itsemphasis in the NHSSP Community Strategy,ensure the availability of recommendeddrugs in all government facilities, andstrictly enforce the restrictions on non-recommended antimalarials.

w Continue advocacy and dissemination oftargeted messages to promote earlytreatment-seeking behaviour for fever andIPTp uptake

w In light of the current malaria parasiteprevalence, begin diagnosis of all malariacases instead of continued presumptivetreatment for children under five years ofage.

w Emphasize joint planning by stakeholdersto integrate interventions to prevent malariain pregnancy.

w Over the long term, maintain the emphasison girls’ education, as the results consis-tently indicate that better educatedmothers are more likely to attend ANC andto use nets.

w Finally, redouble efforts to reach currentmalaria control targets, not only to reducecurrent morbidity and mortality, but alsobecause all indications are that globalwarming induced climate change willincrease the incidence of the disease as aresult of the expansion of malaria endemiczones into highland regions that arecurrently reasonably free of the disease.

Current malaria parasite prevalenceindicates that all malaria cases shouldbe diagnosed, instead of continuingpresumptive treatment for childrenunder five.

KMIS 2007

1

Throughout sub-Saharan Africa malaria isone of the leading causes of morbidityand mortality. Kenya is no exception,and reducing the toll of this disease,especially on small children and pregnantwomen, is a major public health objective. TheKenya Malaria Indicator Survey (KMIS) was anational sample survey designed to provide up-to-date information on progress towards theachievement of the goals the National MalariaStrategy, which was instituted in 2001. Theresults of the survey are presented in this report.

The report opens with a brief profile ofKenya, summarizes the objectives of the NationalMalaria Strategy, and traces the extent of malariaendemicity in Kenya. Following this introduction,the report describes the survey methodology and

implementation in detail. It then presents theresults of the various aspects of the survey andconcludes with a summary of policy recommen-dations arising from the findings.

1.1 Kenya Country Profile

Lying astride the Equator in Eastern Africa,the Republic of Kenya is bordered byEthiopia to the North, Sudan to the North-west, Somalia to the East, Tanzania to theSouth and Uganda to the West. A 536-kilometrecoastline stretches along the Indian Ocean inthe southeast. Kenya lies across latitude 5°North to 5° South and longitude 34° East to42° East.

Covering an area of 582,646 squarekilometres, the land rises from sea level atthe Indian Ocean to 5,199 metres at the highestpeak of Mount Kenya. About 80 per cent ofthe land area is arid or semi-arid and only 20per cent is arable. The land is mostly arid inthe north and fertile in the Lake Victoria Basinin the southwest of the country. The Great

CHAPTER 1

Introduction

Variations in altitude and terrain createcontrasts in Kenya’s climate, from hot andhumid tropical along the coast to temper-ate in the interior and very dry in theNorth and North East. All have differentimplications for the prevalence of malaria.

2

KMIS 2007Rift Valley bisects the Kenya highlands into eastand west. The highlands are cool andagriculturally rich areas in which both largeand small holder farming are carried out.

The variations in altitude and terrain createcontrasts in the country’s climate, which rangesfrom hot and humid tropical along the coastto temperate in the interior and very dry inthe North and North East. There are two rainyseasons, the long rains and the short rains.The long rainy season occurs from April to Juneand the short rainy season from October toDecember. The temperature remains high duringthese months. The hottest period is from Febru-ary to March and coldest from July to August.

Kenya’s geographic position and highpercentage of arid and semi-arid lands makethe country particularly vulnerable to the impactof global warming and climate change. Theeffects of this phenomenon are, in fact, alreadybeing felt in many areas, through prolongeddrought and more intense flooding than havebeen known in the past. Moreover, over thenext few decades the increasing temperaturesare expected to extend the areas of malariaendemicity to zones that are presently relativelyfree of the disease (UNFPA, 2009).

Administratively, Kenya is divided into eightprovinces, which in turn are subdivided into174 districts, each district into divisions, eachdivision into locations and each location intosub-locations. During the 1999 populationcensus, each sub-location was furthersubdivided into census Enumeration Areas (EAs).

1.1.1 History

Kenya became a British protectorate in 1895and a colony in 1920. The country attained itsindependence on 12 December 1963 after anarmed struggle by the Mau Mau movementagainst the colonial government. Politically, amulti-party system was in place until 1982,when the constitution was amended to providefor a one-party state. In 1991, the amend-

ment was repealed and Kenya reverted tomulti-party. From 1964 to 2002, the party inpower was the Kenya African National Union(KANU). The National Rainbow Coalition(NARC) party governed the country after the2002 General Election until 2007. In early 2008,following extensive civil unrest occasioned bya disputed election, the current Grand Coalitiongovernment of the Party of National Unity(PNU) and the Orange Democratic Movement(ODM) took over the leadership.

The country has great ethnic diversity, withmore than 42 tribes. Christianity and Islamare the major religions, but Hinduism ispractised by many in the minority Asiancommunity and some Kenyans observe Africantraditional worship. The official language isEnglish and Kiswahili is the national language.

1.1.2 Economy

Largely market based, Kenya’s economy hasagriculture as its backbone. The sector provideslivelihood to approximately 80 per cent of thepopulation and is a major share of the grossdomestic product (GDP). In 2007 agricultureand forestry contributed about 22.7 per centto GDP and 17.8 per cent of wage employment.The main agricultural cash crops include tea,coffee and horticultural products (fruits,vegetables and cut flowers). Kenya is theworld’s leading producer of cut roses.

The country also has a strong industrialsector; manufacturing contributed 9.7 percent to GDP in 2007. Tourism, a key socio-economic driver, earned the country Ksh65.4billion and directly employed 10.3 per cent ofwage labour in 2007 (Economic Survey, 2008).

Independent Kenya’s economic growthrecord has been mixed. In the first decadeafter independence (1964–1973), the economygrew at impressive rates, with GDP expandingby 6.6 per cent per annum. The first oil crisisof 1973 brought the rapid growth to an abrupthalt and the growth rate decelerated to below4 per cent for much of the 1970s. Theexception was 1976/77, when the unexpected“coffee boom” saw GDP growth rising to 8.2per cent in 1977. But the collapse of the EastAfrican Community in 1977 and the second oilcrisis of 1979 contributed to furtherdeceleration in economic performance.

Kenya’s eight provinces are home topeoples of diverse cultures - more than 42ethnic groups with as many languages.They have in common a dependence onagriculture, which provides the livelihoodsof 80 per cent of the population.

KMIS 2007

3

Drought experienced in the early 1980s,world recession and the international debt crisisworsened the domestic situation. In addition,misaligned real exchange rates, the prevailinginterest rate regime and poor commoditypricing undermined macroeconomic stability.

To address these economic woes the Govern-ment introduced liberalization and deregulationof trade and exchange rate regimes, as wellas public and financial sector reforms. The imple-mentation of the reforms led to a resurgenceof growth, which averaged 5 per cent duringthe 1986–1990 period. Decline set in again inthe 1990s as a result of inconsistent economicpolicy, poor weather and deteriorating infra-structure, declining donor support, high infla-tion owing to massive rent-seeking, insecurity,depressed investments, declining tourismactivities, and the poor performance of themanufacturing sector. Overall GDP growth slidfurther to 2.5 per cent between 1990 and 1995and to 2 per cent between 1996 and 2000. Atthis rate, growth was negative in real termssince it was not keeping pace with populationgrowth. Poverty levels rose to encompass about60 per cent of the population, with both povertyand economic growth negatively affected bythe impact of HIV and AIDS, which reached anestimated adult prevalence rate of 14 per centin 1999 (UNAIDS, 2000), but then declined toabout 6.7 per cent (CBS et al., 2004). NyanzaProvince, which has the highest HIV prealence,is also a malaria endemic area.

The faltering economic performance wasapparent in all sectors of the economy. Thedecline in GDP growth partly reflects a seriesof shocks beyond the control of the GovernmentThese included El Nino-induced floods in late1997; weak soft commodity prices followingthe 1997 Asian financial crises; two successiveyears of severe drought (1999 and 2000); andthe worldwide downturn in tourism after the11 September 2001 terrorist attacks in theUnited States. All were exacerbated by theeconomic policy environment. To redress thesituation the new NARC government imple-mented the Economic Recovery Strategy forWealth and Employment Creation (ERSWEC)beginning in 2002. The economy began torecover, registering growth rates that rosefrom 2.8 per cent in 2003 to 7.0 per cent in2007 (Economic Survey, various issues).

1.1.3 Population

According to the 1999 Kenya Population andHousing Census, the population stood at 28.7million. Previous census results indicated anannual population growth rate of 2.9 per centduring the 1989–1999 periods, a reduction from3.4 per cent recorded for both the 1969–1979and 1979–1989 intercensal periods. Decline infertility rates and realization of the effortscontained in the National Population Policy forSustainable Development (NCPD, 2000) werethe major contributors to this decline in popu-lation growth. On the other hand, mortalityrates have increased since the 1980s, largelyas a result of increased AIDS-related deaths,decline in health services and pervasive poverty.The 2007 population was projected to be 37.2million as a result of changing populationdynamics (Statistical Abstract, 2008)

The Kenya Population and Housing Censusof 1979 reported an increase in the crude birthrate of 54 per 1000, but this declined to 48and 41 per 1,000 in 1989 and 1999, respectively.For a long time the crude death rate has beenon the decline, but the period 1989–1999 reportedan increase of 12 per 1,000 from 11 per 1,000of the previous period 1979–1989. Infant mortalityrate, as crude death rate, decreased from 119deaths per 1,000 live births in 1969 to 88 per1,000 in 1979 and to 68 per 1,000 in 1989, butthen increased to 77 per 1,000 in 1999. Kenyais characterized by a youthful population –almost 44 per cent are under 15 years and only4 per cent are aged 65 and older. This is attri-buted to the high fertility and declining mortalityin the past, and will contribute to continuingpopulation momentum for some years to come.

Kenya’s urban population has grown from3.8 million in 1989 to 9.9 million in 1999,constituting 34 per cent of total population.One result has been the proliferation of urbaninformal settlements leading to environmentaldegradation and deteriorating public healthstandards (KBS, 1989, 1999).

Kenya’s estimated population of 37.2million is largely youthful – almost 44 percent are under 15 years and only 4 percent are aged 65 and older – andincreasingly urbanized.

KMIS 2007

3

Drought experienced in the early 1980s,world recession and the international debt crisisworsened the domestic situation. In addition,misaligned real exchange rates, the prevailinginterest rate regime and poor commoditypricing undermined macroeconomic stability.

To address these economic woes the Govern-ment introduced liberalization and deregulationof trade and exchange rate regimes, as wellas public and financial sector reforms. The imple-mentation of the reforms led to a resurgenceof growth, which averaged 5 per cent duringthe 1986–1990 period. Decline set in again inthe 1990s as a result of inconsistent economicpolicy, poor weather and deteriorating infra-structure, declining donor support, high infla-tion owing to massive rent-seeking, insecurity,depressed investments, declining tourismactivities, and the poor performance of themanufacturing sector. Overall GDP growth slidfurther to 2.5 per cent between 1990 and 1995and to 2 per cent between 1996 and 2000. Atthis rate, growth was negative in real termssince it was not keeping pace with populationgrowth. Poverty levels rose to encompass about60 per cent of the population, with both povertyand economic growth negatively affected bythe impact of HIV and AIDS, which reached anestimated adult prevalence rate of 14 per centin 1999 (UNAIDS, 2000), but then declined toabout 6.7 per cent (CBS et al., 2004). NyanzaProvince, which has the highest HIV prealence,is also a malaria endemic area.

The faltering economic performance wasapparent in all sectors of the economy. Thedecline in GDP growth partly reflects a seriesof shocks beyond the control of the GovernmentThese included El Nino-induced floods in late1997; weak soft commodity prices followingthe 1997 Asian financial crises; two successiveyears of severe drought (1999 and 2000); andthe worldwide downturn in tourism after the11 September 2001 terrorist attacks in theUnited States. All were exacerbated by theeconomic policy environment. To redress thesituation the new NARC government imple-mented the Economic Recovery Strategy forWealth and Employment Creation (ERSWEC)beginning in 2002. The economy began torecover, registering growth rates that rosefrom 2.8 per cent in 2003 to 7.0 per cent in2007 (Economic Survey, various issues).

1.1.3 Population

According to the 1999 Kenya Population andHousing Census, the population stood at 28.7million. Previous census results indicated anannual population growth rate of 2.9 per centduring the 1989–1999 periods, a reduction from3.4 per cent recorded for both the 1969–1979and 1979–1989 intercensal periods. Decline infertility rates and realization of the effortscontained in the National Population Policy forSustainable Development (NCPD, 2000) werethe major contributors to this decline in popu-lation growth. On the other hand, mortalityrates have increased since the 1980s, largelyas a result of increased AIDS-related deaths,decline in health services and pervasive poverty.The 2007 population was projected to be 37.2million as a result of changing populationdynamics (Statistical Abstract, 2008)

The Kenya Population and Housing Censusof 1979 reported an increase in the crude birthrate of 54 per 1000, but this declined to 48and 41 per 1,000 in 1989 and 1999, respectively.For a long time the crude death rate has beenon the decline, but the period 1989–1999 reportedan increase of 12 per 1,000 from 11 per 1,000of the previous period 1979–1989. Infant mortalityrate, as crude death rate, decreased from 119deaths per 1,000 live births in 1969 to 88 per1,000 in 1979 and to 68 per 1,000 in 1989, butthen increased to 77 per 1,000 in 1999. Kenyais characterized by a youthful population –almost 44 per cent are under 15 years and only4 per cent are aged 65 and older. This is attri-buted to the high fertility and declining mortalityin the past, and will contribute to continuingpopulation momentum for some years to come.

Kenya’s urban population has grown from3.8 million in 1989 to 9.9 million in 1999,constituting 34 per cent of total population.One result has been the proliferation of urbaninformal settlements leading to environmentaldegradation and deteriorating public healthstandards (CBS, 1989, 1999).

Kenya’s estimated population of 37.2million is largely youthful – almost 44 percent are under 15 years and only 4 percent are aged 65 and older – andincreasingly urbanized.

4

KMIS 20071 .2 Health Priorities and

Programmes

The Kenya Vision 2030 goal for the healthsector is to provide equitable andaffordable quality health services to allKenyans. The Vision also aims at restructuringthe health care delivery system to shift theemphasis from curative to promotive andpreventive health care. In addition, measuresare being taken to control environmentalthreats to health as part of the effort to lowerthe nation’s disease burden. (Kenya Vision 2030First Medium Term Review, 2008).

The Government of Kenya through theMinistry of Health (MOH) launched the secondNational Health Sector Strategic Plan (NHSSPII) covering the period 2005–2010. The goal ofNHSSP II is to reduce health inequalities and toreverse the downward trend in health-relatedoutcome and impact indicators. The missionof the strategic plan is to promote and parti-cipate in the provision of integrated and highquality promotive, preventive, curative andrehabilitative health care services to all Kenyans.

Through the Kenya Essential Package forHealth (KEPH) introduced with the strategy,NHSSP II intended to achieve the followingobjectives:w Increase equitable access to health services.w Improve the quality and responsiveness of

services in the sector.w Improve the efficiency and effectiveness

of service delivery.w Enhance the regulatory capacity of MOH.w Foster partnerships in improving health and

delivering services.w Improve the financing of the health sector.

NHSSP II shifted the focus of health carefrom the treatment of disease to thepromotion of individual health through variousstages in the human life cycle and the provisionof services at six defined levels of the healthcare system – with both life stages and carelevels defined by KEPH. An important intro-duction was the focus on the community level,which has a specific strategy for buildingcommunity participation in the health caresystem.

In the KEPH approach, health programmes(the malaria programme included) centre onthe different phases of human developmentand thus complement each other, so thatsynergy and mutual reinforcement among theprogrammes can be achieved. This is reflectedin the vision of NHSSP II, which is to have anefficient and high quality health care systemthat is accessible, equitable and affordable forevery Kenyan.

The targets of the NHSSP II focusing onmalaria are:w 60 per cent of under-fives and pregnant

women sleep under long-lasting insecticidetreated bed nets or other insecticidetreated nets (LLITNs/ITNs).

w 80 per cent of patients admitted withsevere malaria receive correct treatment.

w 60 per cent of patients with clinical malariaobtain parasitological diagnosis.

w Malaria case management improved, asreflected in there being no stock outs forthe first line antimalaria drug.

1.3 The National MalariaStrategy and Policy Issues2001–2010

Malaria accounts for up to 30 per cent ofoutpatient attendance and 19 per centof admissions to health facilities. Tocombat this disease the Ministry of Healththrough the National Malaria ControlProgramme (NMCP) developed a NationalMalaria Strategy (NMS) covering the period 2001to 2010. The main goal of the NMS was toreduce the level of malaria infection andconsequent death by 30 per cent by the year2006 and to sustain that improved level ofcontrol to 2010. The strategy’s goal can onlybe attained through partnerships andparticipation of all stakeholders. The strategydescribes four strategic interventions:

1. Clinical management incorporatingprompt effective treatment: This involvesawareness of client communities, improvedcase management by service providers andadequate supply of antimalaria drugs by

KMIS 2007

5

monitoring drug efficacy. In the recent pastthe NMCP developed guidelines to guide healthworkers on the new effective forms oftreatment and diagnosis. There was a policychange in 2005 from sulphadoxine pyrethamine(SP) to artemesinin combination treatment(ACT) as the first line antimalarial after itbecame apparent that SP was no longereffective.

2. Management of malaria and anaemia inpregnancy: This involves provision of SP toantenatal clinics (ANC) so that pregnant womenin endemic areas can access intermittentpreventive treatment (IPT) during their clinicvisits. According to the guidelines, pregnantwomen are supposed to receive at least twodoses of IPT during their gestation afterquickening. The strategy also involves theprovision of ITNs to pregnant women throughthe ANCs in endemic areas. This interventionis done in consultation with the Division ofReproductive Health (DRH).

3. Vector control using ITNs and othermethods: This is achieved by creating demandfor ITNs, reducing taxation on materials usedto manufacture ITNs and protecting economical-ly vulnerable groups. Provision of subsidized netsto vulnerable groups is also part of the policy;this is done through maternal/child healthservices (MCH). Other vector control methodsare larviciding and fogging.

4. Epidemic preparedness and response(including indoor residual spraying): Thisstrategy includes empowering the districtsdefined as epidemic prone to recognize whetherthere is an epidemic or not. In cooperationwith the Division of Disease Surveillance andResponse (DDSR), districts are trained andencouraged to set their thresholds. Ways andmeans of forecasting epidemics are also beinglooked into. The epidemic districts benefit fromannual indoor residual spraying (IRS) campaigns.Buffer stocks of medicines and supplies arealso procured and distributed to districts onceit has been established that there is anepidemic.

The strategy also describes two supportstructures that cut across the four inter-ventions:

w Information, education and communica-tion, which involves the development anddissemination of messages about thevarious interventions. The messages aredisseminated through the mass media,print materials, school health programmes,and the NMCP website. This supportstructure is also responsible for planningKenya’s observance of World Malaria Day,celebrated annually on 25 April, incollaboration with the Division of HealthPromotion.

w Monitoring and evaluation, which is donethrough research measuring targetindicators and assessment of the overallimpact of the NMS, e.g., KMIS.

The targets of the NMS according to theinterventions are included Appendix A.

1 .4 Malaria Endemicity inKenya

F ive rather distinct areas or zones ofmalaria prevalence can be defined inKenya (see Figure 1.1). The zonesillustrated in Figure 1.1 were refined in theKMIS as the following:

Zone 1 depicts the arid/seasonal regionsof the country. These are the North Eastern,North Western and the Southern lowland areas.In these areas malaria is experienced incommunities near bodies of water. There istransmission for a few months of the year.

Zone 2 indicates the coastal endemicregion. In this region, transmission and diseaserisks are similar to that of the lakesideendemic area (zone 4) but exhibit strongerseasonality. The transmission is lower towardsthe Somali border.

Zone 3 depicts the highland epidemicregion: This region has a low disease risk inan average year, but variations in rainfall andtemperatures between the years can lead toepidemics affecting all populations in the

National Malaria Control Programmewww.nmco.or.ke

6

KMIS 2007region. The population in this region does nothave immunity against malaria. In light of therisk of epidemics, IRS is usually done annuallyto prevent transmission in this region. In thesurvey these regions were called Epidemic.

Zone 4 near Lake Victoria indicates thelakeside endemic zone. Malaria transmissionin this zone is throughout the year. The adultpopulation living here acquires partial immunity,so that the risk of and death from malaria are

highly concentrated among children under fiveyears and pregnant women. In the KMIS, thelakeside and coastal regions have been mergedbecause of their similar characteristics andcalled endemic regions.

Zone 5 depicts the low risk regions: Theseregions cover the central area of the country.Several parts of these regions experiencealmost no malaria risk, e.g., Nairobi, Nyeriand Nakuru.

Figure 1.1: Malaria zones in Kenya

KMIS 2007

7

The 2007 Kenya Malaria Indicator Survey(KMIS) was the first survey of its kind tobe carried out in Kenya. The survey wasdesigned to assess the Abuja Declarationtargets, which were adopted in NMS.

The main objective of the KMIS was tomeasure progress toward achieving the goalsand targets set in the NMS 2001–2010. Thespecific objectives of the KMIS 2007 were:1. To collect up-to-date information on

coverage of the core malaria interventionsincluded in the NMS 2001–2010.

2. To assess malaria parasite prevalence inchildren 1–59 months.

3. To assess the status of anaemia amongchildren 6–59 months.

4. To build capacity of the DOMC and itspartners in the implementation of the KMIS.

CHAPTER 2

Survey Organization and Methodology

2.1 Survey Organization

A number of institutions took part inconducting the 2007 KMIS. The Divisionof Malaria control (DOMC) in the Ministryof Health coordinated the overall exercise. Thesample design, training, data collection andanalysis, and report writing were carried outby the Kenya National Bureau of Statistics(KNBS). The Centers for Disease Control andPrevention (CDC) provided technical assistancein the provision and programming of thepersonal digital assistants (PDAs) used for datacollection, training of research assistants, anddata analysis. The Kenya Medical ResearchInstitute/Walter Reed Project Centre ofExcellence for Diagnostic Microscopy (KEMRI/WRP) assisted in training and the collectionand analysis of blood slides. The National Coor-dinating Agency for Population and Development(NCAPD) assisted in training, fieldwork andreport writing. The World Health Organization(WHO) provided technical assistance, whilefunds were provided by the UK’s Departmentfor International Development (DFID).

The population sample surveyed wasexpected to provide estimates for mostof the indicators in malaria zones at theprovincial and urban-rural levels.

8

KMIS 20072.2 Sample Design

Specific targets of the 2007 KMIS werewomen of reproductive age (15–49 years)and children under the age of five yearsliving in malaria endemic or epidemic-proneareas. The geographical coverage included atotal of 63 (out of 69) administrative districts(as at the 1999 Kenya Population and HousingCensus). The excluded areas were NairobiProvince; Kiambu, Nyandarua and Nyeridistricts in Central Province; Meru Centraldistrict in Eastern Province; and Laikipia districtin Rift Valley Province.

The KMIS utilized the National SampleSurvey and Evaluation Programme (NASSEP IV)sampling frame. The master sample is anational two-stage cluster sampling framedeveloped by the KNBS after the 1999 Censusspecifically for household-based sample surveys.The frame was developed using the districtsas the first level stratification. The first stagesampling process of the frame involvedselection of enumeration areas (EAs) andcreation of clusters using the probabilityproportional to measure of size (PPMOS)method. A total of 1,800 clusters was created,of which 1,260 are rural and 560 urban.

A representative sample of 7,200households from a total of 200 clusters wasselected for KMIS and a uniform sample of 36households allocated to each cluster. Theintention was to provide comparable estimatesof key malaria indicators including theprevalence of anaemia in children aged 6 to59 months. The sample was expected toprovide estimates for most of the indicatorsin malaria zones at the provincial and urban-rural levels.

The allocation of the sample into domainswas done to ensure that the lowest domain ofthe study (North Eastern Province) had aminimum of 21 clusters. The sample wasfurther distributed proportionately into districtswithin a province. The first stage of samplingof the clusters was done by KNBS prior tocommencement of fieldwork and the detailsloaded into PDAs fitted with a geographicpositioning system (GPS). All the selectedclusters were mapped using PDAs in a processthat involved collecting the basic descriptionsof all the households and their geographic

coordinates. (See box on facing page.) A simplerandom sampling of 36 households per clusterwas selected with the aid of the PDAs.

Further details on the sample design areprovided in Appendix B.

2.3 Training

An initial 122 persons identified to conductthe field work took part in a seven-daytraining programme prior to commence-ment of the survey. They comprised 80 researchassistants/supervisors and 42 health workers(22 lab technologists and 20 nurses/clinicalofficers). All were trained on the objectives ofKMIS, malaria transmission and the lifecycleof the malaria parasite, interviewing skills, andthe use of PDAs. The research assistants/supervisors were further instructed on theidentification and mapping of the households.

The health workers were trained on thenew treatment policy, the use of hemocues(for measurement of anaemia) and collectionof blood smears for malaria parasites (formicroscopy examination). This was done incollaboration with KEMRI/WRP.

All the personnel participated in a one-dayfield test (dry run) to ensure they understoodthe practical application and to test theinstruments.

In addition, all the supervisors were takenthrough a special session on trouble shootingof PDAs and backing up of the data. They werealso briefed on managing field procedures andpromoting teamwork.

2.4 Fieldwork

From the 122 trained research assistants/supervisors and health workers, 101 wereselected for final data collection. Theremaining 21 were retained as a pool of trainedKMIS personnel to act as buffers in case of anyshortage thereafter. The final personnel con-sisted of 20 teams, each with one supervisor,two research assistants, one lab technologistand one nurse/clinical officer. Each team wasallocated one driver and a vehicle and givenspecific clusters to cover within a given period

KMIS 2007

9

of time. Appendix C contains a list of thesupervisors, enumerators and others involvedin conducting the survey.

A few weeks to the fieldwork, the commu-nities residing in the clusters were mobilizedand sensitized on the KMIS. This was necessarybecause the KMIS involved taking blood samplesfrom under-five children, which is a sensitiveissue to many households.

The fieldwork ran from 10 July to 15 August2007. The teams spent an average of threedays in each cluster with the first day dedicatedto mapping of the households. Once all thehouseholds in each cluster were mapped, the

data were consolidated into the supervisor’sPDA and used to sample 36 households using asimple random sampling method.

The entire fieldwork was closely supervisedby a team of national supervisors who visitedthe teams in the field and ensured that problemsencountered were sorted out. The teams werefacilitated in the field by KNBS district staff –District Statistical Officers and Enumerators –who made sure that the sampled areas wereaccurately identified. Village elders were alsoinstrumental in guiding the teams and mobiliz-ing communities in their respective clusters.

Personal Digital Assistant (PDA)

Apersonal digital assistant (PDA) is a small handheld computer. It generally has asmall screen (approximately 3–4 inches), a long-life battery (allowing more than onefull day of field use), a secure data storage card for data backup (like the ones used indigital cameras) and a touch-sensitive screen. A stylus is used with the touch screen forentering data.

PDAs were used for field data collection by the KMIS interviewers. Each of the PDAshad a global positioning system (GPS) installed to allow mapping of households in thesurveyed clusters. The system allowed the interviewers to pinpoint the location of everydwelling in a chosen cluster and then choose a random sample,thereby creating an up-to-date sampling frame.

A second software programme guided the interviewersthrough the data collection process. The programme checkedinformation as it was being entered in the field, not allowinginvalid entries (e.g., pregnant males) and verifying suspiciousvalues (are you sure there were 11 bed nets in this house?).The software programme showed the interviewers onlyquestions that were appropriate to the respondent beinginterviewed at the moment. For example, if an interviewerwas speaking to a man, questions about a woman’sreproductive history would be automatically skipped. Picturesof various brands of bed nets were included to assistinterviewers in proper identification of the bed nets in eachhousehold. Also included were training videos on the proper useof the hemocue machines for determining haemoglobin levelsand how to properly prepare thin and thick smear slides for assessing parasitaemia. At the end of the survey, the PDAs were returned to Nairobi, where the data on eachwere aggregated onto a single database on a laptop computer. This process took only afew hours. After this, report programmes were run to produce preliminary and finalanalyses. Since the data had already been screened by the data entry programme on thePDA, little data editing was required. The KMIS was the first national survey in Kenya to use the PDA technology.

1 0

KMIS 2007

2.5 Questionnaires

The KMIS adopted the model question-naires developed by the MEASUREDHS+ programme and adopted andrecommended for use by the Roll Back MalariaMonitoring and Evaluation Reference Group(MERG) Task Force on Household Surveys. Thequestionnaires were first reviewed by the KMISTechnical Working Group (TWG) and adaptedto the Kenyan situation. They were laterprogrammed into PDAs by a team from CDC.The PDA English version of the questionnairewas translated into the Swahili language. AllKMIS interviews were done using PDAs; therewere no paper forms.

Appendix D contains samples of thequestionnaires that guided the survey. Theywere split into two categories covering house-hold and individual women components:

w The household questionnaire: Thehousehold questionnaire captured informa-tion on household membership and socio-economic characteristics and was used asa basis for identifying the eligible women(15–49 years). It also contained informationon possession and use of mosquito nets.Haemoglobin measurements for assessinganaemia status were recorded in the house-hold module, along with the results of anyrapid diagnostic tests (RDTs) that wereneeded.

w The woman’s questionnaire: The woman’squestionnaire recorded a recent birthhistory, pregnancy status, use of IPT duringpregnancy and access to prompt treatmentwith antimalarials for children under fivewith reported fever.

2.6 Response Rates

H ousehold and individual interviewresponse rates are presented in Table2.1. About 95 per cent of the targetedhouseholds were interviewed. The surveyyielded response rates of 89 per cent and 87per cent of eligible women and eligible children,respectively. The response rate for householdswas higher in rural areas (96 per cent) thanurban areas (92 per cent).

2 . 7 Data Processing

A s noted, the KMIS data were fullycaptured using PDAs fitted with GPS. Thestandard MIS questionnaires wereprogrammed into the PDAs and tested beforethe actual field work. The data werecontinuously saved in a central PDA in eachteam and later transferred into a personalcomputer for merging and analysis.

The data underwent various cleaningprocesses before analysis. First, the data were

The benefits of participation in thesurvey included identification and on-the-spot treatment of malaria and anaemia:Any child found to be ill because of lowhaemoglobin and/or a positive RDT wasimmediately treated for malaria andreferred to the nearest health facility.

Table 2.1: Response rates for the householdand individual interviews

Number of households, number of interviews and responserates, according to residence

Sample Eligible Com- Responsepleted rate (per

cent)

National Households 7,200 7,200 6,854 95.2*Women(females 15–49) - 6,893 6,111 88.7Children (under5 years) - 5,840 5,105 87.1

Rural Households 5,868 5,868 5,629 95.9* Women(females 15–49) - 5,528 4,892 88.5Children (under5 years) - 4,975 4,406 88.6

UrbanHouseholds 1,332 1,332 1,225 92.0*Women(females 15–49) - 1,365 1,219 89.3Children (under5 years) - 865 725 83.8

* The household response rate is computed as the number ofcompleted household interviews divided by the number of eligiblehouseholds (i.e., sampled households minus households thatwere vacant, destroyed or where all members were absent).

KMIS 2007

1 1

corrected for any mismatch and wrong coding.The process involved using ArcGIS software toplot the coordinates and identify misplacedinformation. The data were further split andmerged into various data sets to ease analysis.The Statistical Package for Social Scientists(SPSS) and the Statistical Analysis System (SAS)software were used for analysis of the data.

2.8 Ethical Considerations

The KMIS protocol was submitted to theKenyatta National Hospital/University ofNairobi Scientific and Ethical Committeeand the KEMRI Scientific and Ethical Commit-tees for approval. During data collection in thefield, informed consent was sought from thehead of the household to administer thequestionnaires. Consent was also sought fromthe parents/guardians of children before anypricking for haemoglobin and blood slide wasdone. Where this was not obtained the interviewand/or procedures were not carried out. Strictconfidentiality was maintained and all personalidentifiers were removed from the data duringanalysis.

The risks and benefits of participation inthe survey were explained to each participantduring the process of informed consent. Therisk of participation for children under five wasminimal, since it was limited mostly totemporary discomfort associated with finger-prick blood collection. The benefits ofparticipation in the survey included identifica-tion and on-the-spot treatment of malaria andanaemia. There was minimal risk to womenwho participated in the interview, other thanthe possible temporary discomfort during thediscussion of sensitive information aroundreproductive history and child survival.

Any child found to be ill because of lowhaemoglobin and/or a positive RDT was imme-diately treated for malaria and referred to ahealth facility nearest to the place of residenceas per the national guidelines.

2.9 Weighting the Sample

The resulting sample was not self-weightedbecause of the imbalance in allocationamong the strata and non-response athousehold or individual levels. A final weightingadjustment procedure was done to providecomparable estimates for the domains of study.

The weighting procedure used designweights derived during the creation of theNASSEP IV sampling frame together with theresponse levels of the survey. Weighting wasfirst done using the design selection proba-bilities and later adjusted to cater for householdand individual non-response. Further, post-stratification adjustment was done on the basisof projected provincial estimates of the studyareas. Finally, the final aggregate weights werenormalized. Details on weighting proceduresare presented in Appendix B.

All the results presented in this report,except those on response rates, are based onthe weighted data.

Nearly all of the targeted householdswere interviewed, as well as 89 per centof eligible women and 87 per cent ofeligible children.

1 2

KMIS 2007

CHAPTER 3

Household Population and HousingCharacteristics

Basic demographic and socio-economiccharacteristics of the surveyedhousehold population are summarizedin this chapter. The householdbackground characteristics include age, placeof residence, sex, educational attainment,household socio-economic status and housingcharacteristics. The background informationassisted in linking the household demographicand socio-economic characteristics to ANCattendance for women aged 15–49 years, healthcare seeking behaviour of children under five,ownership and use of mosquito nets by childrenand pregnant women, and the use ofantimalaria drugs during pregnancy.

3.1 Household Population

R esults of the household populationdistribution by five-year age groups byplace of residence are shown in Table3.1. The results show that the surveyedpopulation consists of 52 per cent females and

48 per cent males. The results further showthat children under five years of age constituteabout 20 per cent of the population.

Figure 3.1 presents the population pyramidof the surveyed population. The age-sexstructure shows a wide base, indicating thatthe population is mostly youthful.

Figure 3.1: Population pyramid

10 8 6 4 2 0 2 4 6 8 10

0 - 45 - 9

10 - 14

15 - 1920 - 2425 - 2930 - 3435 - 39

40 - 4445 - 4950 - 5455 - 5960 - 6465 - 69

70 - 7475 - 79

80+

Age

Percent

Male Female

KMIS 2007

1 3

3.2 Household Composition

Mean household size has implications forthe pattern of household expenditure.It affects the health care status of thehousehold in terms of nutrition, as well as theaffordability and accessibility of health carenecessities such as drugs and mosquito nets.Table 3.2 shows the percentage distribution ofhousehold size, according to residence andmean size of household. The mean size of aKenyan household is 4.4 persons. On average,urban households are smaller (3.7 persons) thanrural households (4.5).

3.3 Household Headship

Nearly 30 per cent of the households arefemale headed. Three in four urbanhouseholds and more than 70 per centof rural households are male headed. Over 60per cent of the household heads have not beeneducated beyond primary level. Heads ofhousehold in urban areas are twice as likely astheir rural counterparts to have at leastsecondary school education. Table 3.3 presentsthe percentage distribution of household headsby sex and education.

Table 3.2: Household composition

Percentage distribution of household size, according toresidence and mean size of household

Number of usual members Urban Rural Total

1 14.0 9.4 10.22 15.9 11.3 12.13 21.1 15.5 16.54 18.9 16.4 16.85 12.2 15.7 15.16 9.9 12.2 11.87 2.9 8.1 7.28 2.8 4.9 4.59+ 2.3 6.4 5.7

Mean size of household 3.7 4.5 4.4Total number of households 1,225 5,629 6,854

NOTE: Table is based on de jure household members, i.e., usualmembers. The data are not for Kenya as a whole but for malaria-prone areas only and exclude Nairobi Province, Central (Kiambu,Nyandarua, Nyeri), Eastern (Meru Central), Rift Valley (Laikipia).

Table 3.1: Household population by age, sex and residence

Percentage distribution of the household population by five-year age groups, according to sex and residence

Urban Rural Total

Age Male Female Total Male Female Total Male Female Total

0–4 20.6 18.4 19.4 20.4 18.8 19.6 20.4 18.7 19.55–9 12.9 11.8 12.3 16.9 15.4 16.1 16.3 14.9 15.610–14 7.1 7.8 7.5 13.0 12.3 12.6 12.2 11.6 11.915–19 6.4 10.8 8.8 9.2 8.9 9.0 8.8 9.1 9.020–24 9.5 14.6 12.3 6.9 9.1 8.0 7.2 10.0 8.725–29 12.6 12.8 12.7 6.3 7.5 6.9 7.2 8.3 7.730–34 8.5 7.4 7.9 5.8 6.1 5.9 6.1 6.3 6.235–39 8.0 6.2 7.0 4.6 4.9 4.8 5.1 5.1 5.140–44 5.3 2.8 4.0 3.7 3.5 3.6 4.0 3.4 3.745–49 3.1 2.8 2.9 3.0 2.7 2.9 3.0 2.7 2.950–54 2.5 2.0 2.2 2.5 3.0 2.7 2.5 2.8 2.755–59 1.6 0.7 1.1 2.6 2.3 2.4 2.4 2.1 2.260–64 0.7 0.5 0.6 1.7 1.8 1.7 1.5 1.6 1.665–69 0.5 0.5 0.5 1.2 1.5 1.3 1.1 1.3 1.270–74 0.2 0.3 0.3 1.1 0.9 1.0 0.9 0.8 0.975–79 0.1 0.1 0.1 0.6 0.7 0.6 0.5 0.6 0.680 + 0.2 0.4 0.3 0.7 0.7 0.7 0.6 0.7 0.6Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0Number 2,233 2,492 4,725 12,323 13,001 25,324 14,556 15,493 30,049

NOTE: Table is based on de jure household members, i.e., usual members. The data are not for Kenya as a whole, but for malaria-proneareas only and exclude Nairobi Province, Central (Kiambu, Nyandarua, Nyeri), Eastern (Meru Central) and Rift Valley (Laikipia).

Table 3.3: Sex and education of householdhead

Percentage distribution by sex and education of householdhead

Characteristic Residence

Urban Rural Total

GenderMale 74.9 70.1 70.9Female 25.1 29.9 29.1

Household educationNursery/Kindergarten 8 23.6 21Primary 31.6 51.5 48.1Post-primary/Vocational 1.7 1.8 1.8Secondary/ ‘A’ level 40.5 18.5 22.3College 12.1 3.8 5.3University 6.1 0.8 1.7

Total 100 100 100Total number of households 1,225 5,629 6,854

NOTE: The data are not for Kenya as a whole but for malaria-prone areas only and exclude Nairobi Province, Central (Kiambu,Nyandarua, Nyeri), Eastern (Meru Central), Rift Valley (Laikipia).

1 4

KMIS 20073.4 Demographic

Characteristics of Women

F igure 3.2 and Table 3.4 present thedistribution of women age 15–49 yearsby background characteristics. About 40per cent of the respondents were aged 15–24years and 32 per cent aged 25–34 years.

Figure 3.2: Women respondents by age

The survey further shows that about halfof the of women of reproductive age haveattained primary level school of education, but24 per cent have no education at all. Overall,three in four women (75 per cent) have notgone beyond the primary level of education.Most of the women respondents (85 per cent)are Christians of various denominations; only8 per cent are Muslims.

3.5 Housing Characteristics

Information on housing characteristic suchas access to electricity, source of drinkingwater, sanitation and flooring materials iskey to explaining interrelationships betweensocial and economic conditions of the householdand its likely exposure to and prevalence ofdiseases (2003 Kenya Demographic and HealthSurvey, CBS et al., 2004). Table 3.5 presentsthe distribution of households by housing char-acteristics.

3.5.1 Electricity

Only 10.3 per cent of the households inter-viewed have access to electricity. There is asignificant difference in the supply of electricitybetween urban and rural residents. Half ofurban households in malaria endemic areas areconnected to electricity supply, but only 2.2per cent of rural households.

3.5.2 Cooking fuel

About 78 per cent of households depend onfirewood or other biomass as cooking fuel.Whereas urban households mostly use charcoal(50 per cent) or kerosene (26 per cent), rural

Table 3.4: Demographic characteristics offemale respondents

Percentage distribution of women aged 15-49 by backgroundcharacteristics

Background characteristic Percentage

ResidenceUrban 19.2Rural 80.8

ProvinceCentral 7.5Coast 10.5Eastern 17.0North Eastern 3.3Nyanza 18.2Rift Valley 30.1Western 13.4

School levelNone 23.7Nursery/Kindergarten 1.0Primary 50.5Post-primary/Vocational 1.1Secondary/ ‘A’ level 19.5College (Middle level) 3.3University 0.8

ReligionRoman Catholic 22.4Protestant/Other Christian 63.3Muslim 8.3No religion 2.5Other 3.3

Total 100Total women 6,111

NOTE: The data are not for Kenya as a whole but for malaria-prone areas only and exclude Nairobi Province, Central (Kiambu,Nyandarua, Nyeri), Eastern (Meru Central), Rift Valley (Laikipia).

0

5

10

15

20

25

15- 19 20 - 24 25- 29 30 - 34 35-39 40 - 44 45-49

The mean size of a Kenyan household is4.4 persons. On average, urbanhouseholds are smaller (3.7 persons) thanrural households (4.5). Over 60 per centof household heads have not beeneducated beyond primary level. Heads ofhouseholds in urban areas are twice aslikely as their rural counterparts to haveat least secondary school level ofeducation.

KMIS 2007

1 5

Biomass is the leading cooking fuel forKenyans in malaria-prone areas. Nine outof ten rural households in these areasdepend on firewood or straw and nearlyhalf of urban households use charcoal.Another 26 per cent of urban householdsuse kerosene. Less than 3 per cent ofrural folk have access to electricity,against nearly half of urbanites.

households use mostly firewood/straw (91 percent).

3.5.3 Drinking Water

Almost 28 per cent of households obtaindrinking water directly from rivers or streams;nearly 27 per cent use either water piped intothe dwelling or plot or a public tap. About threein four urban households (75 per cent) and 17per cent of those in rural areas have waterpiped into the dwelling, plot or public tap.

For households in rural areas, the mainsource of drinking water is a river or stream(33 per cent), followed by a spring (14 percent) and a covered public well (10 per cent).

3.5.4 Sanitation

Nearly two-thirds of the households intervieweduse traditional pit latrines, about 10 per centuse ventilated improved pit latrines (VIPs) and7 per cent use flush toilets. But 19 per centhave no toilet facilities.

In urban areas, 35 per cent, 44 per centand 19 per cent of the households have flushtoilets, traditional pit latrines and VIPs,respectively. The most commonly used toiletsin rural areas are traditional pit latrines (68.6per cent).

3.5.5 Floor

Over half of the households use earth as theirflooring material. The next predominantly usedflooring materials are cement (30 per cent)and dung (19 per cent). Among urban house-holds, 84 per cent use cement, while 59 percent of rural households use earth as theirflooring material.

Table 3.5: Housing characteristics

Percentage distribution of households by housing charac-teristic, according to residence

Housing characteristic Residence

Urban Rural Total

ElectricityYes 49.8 2.2 10.3No 50.2 97.8 89.7

Cooking fuelElectricity 1.4 0.1 0.3LPG/natural gas 9.8 0.4 2.0Biogas 0.1 0.0 0.0Kerosene 26.3 1.2 5.5Coal/lignite 0.0 0.0Charcoal 49.7 6.7 14.1Firewood/straw 11.6 91.4 77.7Other 1.1 0.2 0.3

Drinking waterPiped into dwelling 18.1 1.5 4.3Piped into yard/plot 17.1 5.3 7.3Public tap/standpipe 40.1 10.2 15.3Open well in compound 1.0 2.0 1.8Open public well 1.9 6.0 5.3Covered well in compound/plot 2.9 4.2 3.9Rainwater 2.3 2.6 2.5Bottled water 2.2 0.1 0.4Covered public well 7.4 10.4 9.8Spring 1.6 14.4 12.2River/stream 2.1 33.2 27.9Lake/pond 0.0 4.6 3.8Other 3.2 1.7 2.0Dam 3.8 3.2

ToiletFlush toilet 35.0 0.8 6.6Traditional pit latrine 44.3 68.6 64.4Ventilated improved pit latrine (VIP) 18.6 8.2 9.9No facility/ bush/field 2.1 22.4 18.9Other 0.1 0.2 0.1

FloorEarth/sand 9.8 58.8 50.4Dung 3.9 21.7 18.7Wood planks 0.5 0.4 0.4Palm/bamboo 0.0 0.0Parquet or polished wood 0.1 0.0 0.0Ceramic tiles 1.4 0.0 0.2Cement 84.3 18.9 30.1Other 0.2 0.1

Total 100.0 100.0 100.0Total number of households 1,225 5,629 6,854

NOTE: The data are not for Kenya as a whole but for malaria-prone areas only and exclude Nairobi Province, Central(Kiambu, Nyandarua, Nyeri), Eastern (Meru Central), Rift Valley(Laikipia).

Over a quarter of rural households gettheir water directly from rivers, streamsor springs. The most commonly usedtoilets in rural areas are traditional pitlatrines, but nearly a fifth of ruralhouseholds have no toilet facilities.

1 6

KMIS 20073.6 Household Possessions

Household possessions indicate the socio-economic status of households. Apartfrom possession of a refrigerator, theKMIS collected information on means oftransportation and communication. As shownin Table 3.6, over 70 per cent of the householdsown a radio, 37 per cent have a cell phone and33 per cent have a bicycle. Radios are foundin 83 per cent of urban households and 68 percent of rural ones. The survey further showsthat 73 per cent of urban households have acell phone, compared with 30 per cent of ruralhouseholds. A notable difference in possessionsis evidenced by 46 per cent of urban householdsowning a television against 10 per cent of ruralhouseholds. More rural households (36 per cent)than urban ones (21 per cent) possess bicycles.

Table 3.6: Household possessions

Percentage distribution of various household possessions byresidence

Residence

Household possessions Urban Rural Total

Radio 83.3 68.4 71.0TV 45.9 9.7 15.9Phone 1.3 0.4 0.5Cell phone 73.1 29.6 37.1Refrigerator 10.4 0.4 2.1Bicycle 20.6 35.7 33.1Motorcycle 1.2 0.6 0.7Car 8.0 1.7 2.7Boat 0.3 0.3 0.3Donkey 0.4 10.3 8.6None of the above 7 22.9 20.2Total number of households 1,225 5,629 6,854

NOTE: The data are not for Kenya as a whole but for malaria-prone areas only and exclude Nairobi Province, Central(Kiambu, Nyandarua, Nyeri), Eastern (Meru Central), Rift Valley(Laikipia.)

KMIS 2007

1 7

CHAPTER 4

Vector Control

Malaria distribution in Kenya is notuniform because of wide geograph-ical differences in altitude, rainfall,temperature and humidity. Thesefactors determine the intensity of malariatransmission as they directly influence malariavector distribution, vector density, bitingintensity and the development of malariaparasites in the vectors. Malaria control effortsare thus targeted to maximize the benefits ofthe various interventions.

The primary vector control method in Kenyais the use of insecticide treated mosquito nets(ITNs), which have been proven to be effectivein preventing malaria. Kenya’s NMS set twomajor targets for ITN use by 2006:

w 60 per cent of children under five sleepingunder an ITN.

w 60 per cent of pregnant women sleepingunder an ITN.

4.1 Household Ownership ofMosquito Nets

Several kinds of mosquito nets are foundin households in Kenya, but two categoriesof nets are discussed in this survey:Conventional untreated nets and insecticidetreated nets (ITNs). ITNs offer far greaterprotection against malaria than untreated nets.They include both conventional nets with up-to-date insecticide treatment (treated withinthe last six months) and long-lasting insecticidetreated nets (LLITNs). LLITNs can last 3–5 years(the assumed lifetime of the net) and thusnever require retreatment.

LLITNs are mosquito nets that are pre-treated in the factory with an insecticidethat can withstand many washes withoutdecreasing the efficacy of the insecticide.Treated this way, the nets are veryeffective against mosquitoes.

1 8

KMIS 20074.1.1 Programme Efforts to

Increase Net Ownership

In late 2001, a social marketing programmebegan distributing subsidized conventional netsbundled with an insecticide treatment(deltamethrin) through rural retail shops andother private sector commercial outlets at aprice of Ksh100. The number of nets distributedthrough this programme up to 2007 totals11,162,204 (PSI/Kenya, 2008).

Starting in 2004, public sector ANCs inmalaria endemic and epidemic districts begandistributing subsidized ITNs to pregnant womenand to children under five years at a price ofKsh50. Then, in 2006, the Kenya Governmentconducted a mass net distribution campaignthat resulted in the distribution of 3.4 millionLLITNs in malaria-prone districts. The campaignwas conducted in two phases. The first phasewas integrated with measles and polioimmunization and vitamin A supplementation.The second phase was a stand-alone campaigncarried out by the Division of Malaria Control.The two campaigns yielded similar results interms of net uptake and usage.

All of these initiatives were supported by arange of multi-media messages:w A behaviour change campaign that attempted

to establish ITN use as a social norm, andhighlighted vulnerable populations.

w A branded campaign that attempted toremind people that you sleep better undera net because you won’t be bothered bymosquitoes.

w A branded campaign promoting a longer-lasting retreatment kit.

These campaigns are considered to havesignificantly influenced the culture of netownership and use.

4.1.2 Prevalence of NetOwnership

Overall, 63 per cent of households own at leastone net and 34 per cent of households ownmore than one net (Table 4.1). As mentionedabove, ITNs are much more effective thanconventional nets, so it’s important to isolateITN use. About 48 per cent of households haveat least one ITN and 23 per cent have more

Table 4.1: Ownership of mosquito nets

Percentage distribution of households with at least one and more than one mosquito net (treated or untreated), ever-treated net, ITNand long-lasting insecticide-treated net (LLITN), and average number of nets of each type by household, according to backgroundcharacteristics

Background Percentage Percentage Average Percentage Percentage Average Numbercharacteristics of house- of house- number of of house- of house- number of

holds that holds that nets per holds that holds that of ITNs per house-have at have more house have at have more houehold holds

least one net than one net hold least one net than one ITN

ResidenceUrban 68.7 35.3 1.3 47.7 21.3 0.8 1,225Rural 61.2 33.4 1.2 47.7 22.8 0.8 5,629

Wealth IndexLowest 46.8 21.4 0.8 35.3 14.6 0.6 1,371Second 61.3 30.4 1.1 49.7 21.5 0.8 1,369Third 58.1 27.5 1.0 44.4 18.9 0.7 1,372Fourth 70.3 43.5 1.4 55.3 30.0 1.0 1,372Highest 72.3 42.6 1.5 50.9 25.6 1.0 1,370

Level of endemicityEndemic 73.4 39.3 1.4 57.9 27.6 1.0 2,770Epidemic 59.0 33.4 1.2 47.9 24.4 0.8 876Low risk 48.0 22.8 0.8 32.1 12 0.5 1,318Seasonal Transmission 58.9 33.6 1.2 43.0 20.8 0.8 1,890

Net campaign districtNo campaign 49.0 23.9 0.9 33.4 13.3 0.5 2,226First campaign 74.0 39.5 1.4 59.2 28.9 1.0 2,148Second campaign 63.3 36.5 1.3 49.2 24.3 0.9 2,480

Total 62.5 33.7 1.2 47.7 22.5 0.8 6,854

NOTE: The data are not for Kenya as a whole but for malaria-prone areas only and exclude Nairobi Province, Central (Kiambu,Nyandarua, Nyeri), Eastern (Meru Central), Rift Valley (Laikipia).

KMIS 2007

1 9

than one. The average number of ITNs perhousehold is 0.8, compared with an averageof 1.2 nets of any kind per household.

Even though KMIS did not cover all theregions in Kenya, the national averages areimportant to understand, as they contributeto Abuja targets. They represent a total of allregions, including those areas that have lowrisk of malaria, and those that have not beenbeneficiaries of net campaigns and public netinitiatives. Far more revealing are the figuresfrom malaria endemic and epidemic areas.

In endemic areas, 73 per cent of householdsown at least one net and 39 per cent ownmore than one. Among those owning nets, theaverage number of nets per household stoodat 1.4. In the same areas 58 per cent of allhouseholds own at least one ITN and 28 percent own more than one. On average,households in these areas owned one net.

In epidemic-prone areas, 59 per cent ofhouseholds own at least one net and 33 percent own more than one. The average numberof nets per household in these areas is 1.2nets. In the same areas, 48 per cent of allhouseholds own at least one ITN and 24 percent own more than one. The average numberof ITNs per household in these areas is 0.8.

Ownership of at least one ITN is consistentacross the top four wealth quintiles. It is consis-tently lowest for the lowest wealth quintile,however, suggesting a need for free distributionof nets to increase coverage for the poor.

4.2 Use of Mosquito Nets byChildren under Five Yearsof Age

Responses revealed that 51 per cent ofchildren under five years of age sleptunder any net the night preceding thesurvey, but only 39 per cent slept under an ITN(Table 4.2). The data show that the youngerthe child the more likely to have slept underany net.

The place of residence plays a major partin use of mosquito nets by children under five.About 61 per cent of children under five inurban areas used a mosquito net the nightpreceding the survey, compared with 50 per

cent in rural areas. For ITNs, the rates were43 per cent of urban children and 39 per centof their counterparts in the rural areas.

The survey further shows that the wealthierthe household, the more likely it is that childrenunder five will sleep under nets.

The results also show that endemic areashave the highest use of mosquito nets (58 percent) and ITNs (46 per cent). In epidemic, lowrisk and seasonal transmission areas, bothchildren and adults are at high risk of malariabecause adults may not have developed immu-nity, unlike those in endemic areas. In all areas,however, ITN use by children under five is belowthe 60 per cent 2006 Abuja target.

Examination of the potential effect of themass net distribution campaigns shows thatboth campaigns yielded higher ITN usage

Table 4.2: Use of mosquito nets by children

Percentage distribution of children under five years of agewho slept under a mosquito net or an ITN the night precedingthe survey, by background characteristics

Background Percentage Percentage Num-characteristics of children of children ber of

under five under five childrenwho slept who slept under

under a under an fivemosquito ITN last years

net last night of agenight

Age in years0 54.1 40.6 1,2781 56.9 42.8 1,0762 49.5 38.5 1,1593 47.3 36.5 1,1414 47.6 37.4 1,186

SexMale 50.7 37.9 2,959Female 51.6 40.5 2,881

ResidenceUrban 60.6 42.5 865Rural 49.5 38.6 4,975

Wealth indexLowest 39.2 29.1 1,374Second 47.4 37.3 1,409Third 50.3 38.6 1,025Fourth 58.7 46.5 1,199Highest 61.6 44.5 833

Level of endemicityEndemic 58.2 45.8 2,468Epidemic 45.9 37.9 842Low risk 42.7 29.7 802Seasonal transmission 48.7 34.1 1,728

Net campaign districtNo campaign 40.1 26.6 1,854First campaign 55.3 43.9 2,044Second campaign 55.4 44.1 1,942

Total 51.1 39.2 5,840

NOTE: The data are not for Kenya as a whole but for malaria-prone areas only and exclude Nairobi Province, Central (Kiambu,Nyandarua, Nyeri), Eastern (Meru Central), Rift Valley (Laikipia).

2 0

KMIS 2007among children under five compared with thenon-campaign districts: In areas where therewas no campaign, only 27 per cent of childrenunder five had slept under an ITN the nightpreceding the survey, compared with 44 percent of those residing in areas where thesecond campaign was carried out.

4.3 Use of Nets by Women ofReproductive Age andPregnant Women

Survey results show that pregnant womenare more likely to sleep under a mosquitonet than all women within the reproduc-tive age (Table 4.3). About 51 per cent ofpregnant women slept under a net the nightpreceding the survey, compared with 45 percent of all women of reproductive age.

Similarly, 40 per cent of pregnant women sleptunder an ITN on the night preceding the survey,compared with 32 per cent of all women withinthe reproductive years.

Women of reproductive age in urban ar